diff options
Diffstat (limited to 'pages/docs/tech')
| -rw-r--r-- | pages/docs/tech/en.haml | 5 | ||||
| -rw-r--r-- | pages/docs/tech/hard-problems/en.md | 169 | ||||
| -rw-r--r-- | pages/docs/tech/hard-problems/pt.md | 133 | ||||
| -rw-r--r-- | pages/docs/tech/infosec/_table-style.haml | 59 | ||||
| -rw-r--r-- | pages/docs/tech/infosec/_table.haml | 233 | ||||
| -rw-r--r-- | pages/docs/tech/infosec/en.haml | 105 | ||||
| -rw-r--r-- | pages/docs/tech/limitations.md | 123 | ||||
| -rw-r--r-- | pages/docs/tech/routing.md | 65 | ||||
| -rw-r--r-- | pages/docs/tech/secure-email/en.md | 578 |
9 files changed, 1470 insertions, 0 deletions
diff --git a/pages/docs/tech/en.haml b/pages/docs/tech/en.haml new file mode 100644 index 0000000..c03b89b --- /dev/null +++ b/pages/docs/tech/en.haml @@ -0,0 +1,5 @@ +- @title = "LEAP Technology Notes" +- @nav_title = "Technology" +- @summary = "Musings, notes, and background information on various technology issues that relate to LEAP." + += child_summaries
\ No newline at end of file diff --git a/pages/docs/tech/hard-problems/en.md b/pages/docs/tech/hard-problems/en.md new file mode 100644 index 0000000..4d9598a --- /dev/null +++ b/pages/docs/tech/hard-problems/en.md @@ -0,0 +1,169 @@ +@title = 'Hard problems in secure communication' +@nav_title = 'Hard problems' +@summary = "How LEAP addresses the difficult problems in secure communication" + +## The big seven + +If you take a survey of interesting initiatives to create more secure communication, a pattern starts to emerge: it seems that any serious attempt to build a system for secure message communication eventually comes up against the following list of seven hard problems. + +1. **Public key problem**: Public key validation is very difficult for users to manage, but without it you cannot have confidentiality. +2. **Availability problem**: People want to smoothly switch devices, and restore their data if they lose a device, but this is very difficult to do securely. +3. **Update problem**: Almost universally, software updates are done in ways that invite attacks and device compromises. +4. **Meta-data problem**: Existing protocols are vulnerable to meta-data analysis, even though meta-data is often much more sensitive than content. +5. **Asynchronous problem**: For encrypted communication, you must currently choose between forward secrecy or the ability to communicate asynchronously. +6. **Group problem**: In practice, people work in groups, but public key cryptography doesn't. +7. **Resource problem**: There are no open protocols to allow users to securely share a resource. + +These problems appear to be present regardless of which architectural approach you take (centralized authority, distributed peer-to-peer, or federated servers). + +It is possible to safely ignore many of these problems if you don't particularly care about usability or matching the features that users have grown accustomed to with contemporary methods of online communication. But if you do care about usability and features, then you are stuck with finding solutions to these problems. + +## Our solutions + +In our work, LEAP has tried to directly face down these seven problems. In some cases, we have come up with solid solutions. In other cases, we are moving forward with temporary stop-gap measures and investigating long term solutions. In two cases, we have no current plan for addressing the problems. + +### Public key problem + +The problem: + +> Public keys is very difficult for users to manage, but without it you cannot have confidentiality. + +If proper key management is a precondition for secure communication, but it is too difficult for most users, what hope do we have? + +The problem of public keys breaks down into five discrete issues: + +* **Key discovery** is the process of obtaining the public key for a particular user identifier. Currently, there is no commonly accepted standard for mapping an identifier to a public key. For OpenPGP, many people use keyservers for this (although the keyserver infrastructure was not designed to be used in this way). A related problem is how a client can discover public key information for all the contacts in their addressbook for phonebook without revealing this information to a third party. +* **Key validation** is the process ensuring that a public key really does map to a particular user identifier. This is also called the "binding problem" in computer science. Traditional methods of key validation have recently become discredited. +* **Key availability** is the assurance that the user will have access, whenever needed, to their keys and the keys of other users. Almost every attempt to solve the key validation problem turns into a key availability problem, because once you have validated a public key, you need to make sure that this validation is available to the user on all the possible devices they might want to send or receive messages on. +* **Key revocation** is the process of ensuring that people do not use an old public key that has been superseded by a new one. + +Of these problems, key validation is the most difficult and most central to proper key management. The two traditional methods of key validation are either the X.509 Certificate Authority (CA) system or the decentralized "Web of Trust" (WoT). Recently, these schemes have come under intense criticism. Repeated security lapses at many of the Certificate Authorities have revealed serious flaws in the CA system. On the other hand, in an age where we better understand the power of social network analysis and the sensitivity of the social graph, the exposure of metadata by a "Web of Trust" is no longer acceptable from a security standpoint. + +An alternative method of key validation is called TOFU for Trust On First Use. With TOFU, a public key is assumed to be the right key the first time it is used. TOFU can work well for long term associations and for people who are not being targeted for attack, but its security relies on the security of the discovery transport and the application's ability to retain a memory of discovered keys. + +TOFU can break down in many real-world situations where a user might need to generate new keys or securely communicate with a new contact. TOFU is widely used for protocols like SSH, where the user receives confirmation of key continuity each time they connect to a server. There is no such confirmation with asynchronous messaging protocols, making TOFU much less appropriate in these situations. + +Other strategies for addressing parts of the key management problem include: + +1. Inline Keys: Many projects plan to facilitate discovery by simply including the user's public key in every outgoing message (as an attachment, in a footer, or in a header). +1. DNS: Key distributed via DNSSEC, where a service provider adds a DNS entry for each user containing the user's public key or fingerprint. This places all the trust in the DNS owner. +1. Network perspective: Validation by key endorsement (third party signatures), with audits performed via network perspective. +1. Introductions: Discovery and validation of keys through acquaintance introduction. +1. Mobile: Although too lengthy to manually transcribe, an app on a mobile device can be used to easily exchange keys in person (for example, via a QR code or bluetooth connection). +1. Append-only log: There is a proposal to modify Certificate Transparency to handle user accounts, where audits are performed against append-only logs. +1. Biometric feedback: In the one case of voice communication, you can use recognition of the other person's voice as a means to validate the key (when used in combination with a Short Authentication String). This is how ZRTP works. + +For LEAP, we have developed a unique federated system called [Nicknym](/nicknym) that automatically discovers and validates public keys allowing the user to take advantage of public key cryptography without knowing anything about keys or signatures. Nicknym uses a combination of TOFU, provider endorsement, and network perspective. There is also a new proposal very similar to Nicknym called [Nyms](https://nymsio.github.io/) which we hope to also support. Nyms adds the ability of users from non-participating service providers to register their keys. + +### Availability problem + +The problem: + +> People want to smoothly switch devices, and restore their data if they lose a device, but this very difficult to do securely. + +Users today demand the ability to access their data on multiple devices and to have piece of mind that their data will not be lost forever if they lose a device. + +At LEAP, we have worked to solve the availability problem with a system we call [Soledad](/soledad) (for Synchronization of Locally Encrypted Documents Among Devices). Soledad gives the client application an encrypted, synchronized, searchable document database. All data is client encrypted, both when it is stored on the local device and synced with the cloud. This is very powerful, as it allow the client developer to take advantage of a rich document database without needing to worry about how it is backed up or synchronized. + +As far as we know, there is nothing else like it, either in the free software or commercial world. However, there are several projects in a similar problem space: + +* [Mylar](http://css.csail.mit.edu/mylar/), for client-encrypting web application data. +* [Crypton](https://crypton.io/), for client-encrypting web application data. +* [Firefox Sync](https://wiki.mozilla.org/Services/Sync) + +Soledad tries to solve the problem of general data availability, but other initiatives have tried to tackle the more narrow problem of availability of private keys and discovered public keys. These initiatives include: + +* [Whiteout key sync](https://blog.whiteout.io/2014/07/07/secure-pgp-key-sync-a-proposal/) +* Nilcat, experimental [code for cloud storage of keys](https://github.com/mettle/nilcat) +* Ben Laurie's [proposed protocol for storing secrets in the cloud](http://www.links.org/files/nigori/nigori-protocol-01.html) +* Phillip Hallam-Baker's [thoughts along similar lines](http://tools.ietf.org/html/draft-hallambaker-prismproof-key-00) + +### Update problem + +The problem: + +> Almost universally, software updates are done in ways that invite attacks and device compromises. + +The sad state of update security is especially troublesome because update attacks can now be purchased off the shelf by repressive regimes. The problem of software update is particular bad on desktop platforms. In the case of mobile and HTML5 apps, the vulnerabilities are not as dire, but the issues are also harder to fix. + +To address the update problem, LEAP is adopting a unique update system called Thandy from the Tor project. Thandy is complex to manage, but is very effective at preventing known update attacks. + +Thandy, and the related [TUF](https://updateframework.com), are designed to address the many [security vulnerabilities in existing software update systems](https://github.com/theupdateframework/tuf/blob/develop/SECURITY.md). In one example, other update systems suffer from an inability of the client to confirm that they have the most up-to-date copy, thus opening a huge vulnerability where the attacker simply waits for a security upgrade, prevents the upgrade, and launches an attack exploiting the vulnerability that should have just been fixed. Thandy/TUF provides a unique mechanism for distributing and verifying updates so that no client device will install the wrong update or miss an update without knowing it. + +Related to the update problem is the backdoor problem: how do you know that an update does not have a backdoor added by the software developers themselves? Probably the best approach is that taken by [Gitian](https://gitian.org/), which provides a "deterministic build process to allow multiple builders to create identical binaries". We hope to adopt Gitian in the future. + +### Meta-data problem + +The problem: + +> Existing protocols are vulnerable to meta-data analysis, even though meta-data is often much more sensitive than content. + +As a short term measure, we are integrating opportunistic encrypted transport (TLS) for email and chat messages when relayed among servers. There are two important aspects to this: + +* Relaying servers need a solid way to discover and validate the keys of one another. For this, we are initially using DNSSEC/DANE. +* An attacker must not be able to downgrade the encrypted transport back to cleartext. For this, we are modifying software to ensure that encrypted transport cannot later be downgraded. + +This approach is potentially effective against external network observers, but does not protect the meta-data from the service providers themselves. Also, it does not, by itself, protect against more advanced attacks involving timing and traffic analysis. + +In the medium term, LEAP plans to support direct delivery from client to server via Tor for service providers that support this. These anonymously delivered messages would be kept in a separate folder and only displayed to the user if the message signatures are valid. There is a lot of open debate as to the efficacy of using a low-latency onion routing network like Tor for something that might be better suited to a high latency mixing network. For now, Tor is useful because it exists and has a lot of traffic already. For a great discussion comparing mix networks and onion routing, see [Tom Ritter's blog post on the topic](https://ritter.vg/blog-mix_and_onion_networks.html). + +In the long term, we plan to adopt one of the proposed schemes for securely routing meta-data. These include: + +* Auto-alias-pairs: Each party auto-negotiates aliases for communicating with each other. Behind the scenes, the client then invisibly uses these aliases for subsequent communication. The advantage is that this is backward compatible with existing routing. The disadvantage is that the user's server stores a list of their aliases. As an improvement, you could add the possibility of a third party service to maintain the alias map. +* Onion-routing-headers: A message from user A to user B is encoded so that the "to" routing information only contains the name of B's server. When B's server receives the message, it unwraps (unencrypts) a supplementary header that contains the actual user "B". Like aliases, this provides no benefit if both users are on the same server. As an improvement, the message could be routed through intermediary servers. +* Third-party-dropbox: To exchange messages, user A and user B negotiate a unique "dropbox" URL for depositing messages, potentially using a third party. To send a message, user A would post the message to the "dropbox". To receive a message, user B would regularly polls this URL to see if there are new messages. +* Mixmaster-with-signatures: Messages are bounced through a mixmaster-like set of anonymization relays and then finally delivered to the recipient's server. The user's client only displays the message if it is encrypted, has a valid signature, and the user has previously added the sender to a 'allow list' (perhaps automatically generated from the list of validated public keys). +* Tor: One scheme employed by Pond is to simply allow for direct delivery over Tor from the sender's device to the recipient's server. This is fairly simple, and places all the work on the existing Tor network. + +In all of these cases, meta-data protected routing can make abuse prevention more difficult. For this reason, it probably makes sense to only allow once of these options once both parties have already exchanged key material, in order to prevent the user being flooded with anonymous Spam. + +### Asynchronous problem + +The problem: + +> For encrypted communication, you must currently choose between forward secrecy or the ability to communicate asynchronously. + +With the pace of growth in digital storage and decryption, forward secrecy is increasingly important. Otherwise, any encrypted communication you engage in today is likely to become cleartext communication in the near future. + +In the example of email and chat, we have OpenPGP with email and OTR with chat: the former provides asynchronous capabilities, and the latter forward secrecy, but neither one supports both abilities. We need both better security for email and the ability to send/receive offline chat messages. + +In the short term, we are layering forward secret transport for email and chat relay on top of traditional object encryption (OpenPGP). This approach is identical to our stop-gap approach for the meta-data problem, with the one addition that relaying servers need the ability to not simply negotiate TLS transport, but to also negotiate forward secret ciphers and to prevent a cipher downgrade. + +This approach is potentially effective against external network observers, but does not achieve forward secrecy from the service providers themselves. + +In the long term, we plan to work with other groups to create new encryption protocol standards that can be both asynchronous and forward secret: + +* [Triple elliptical curve Diffie-Hellman handshake](https://whispersystems.org/blog/simplifying-otr-deniability/) +* [Forward Secrecy Extensions for OpenPGP](http://tools.ietf.org/html/draft-brown-pgp-pfs-03) + +The [Axolotl protocol](https://github.com/trevp/axolotl/wiki) used by both Pond and TextSecure currently has the most mature approach to asynchronous forward secrecy. This could be added as an invisible upgrade to normal email encryption when the client detects that both parties support it. + +### Group problem + +The problem: + +> In practice, people work in groups, but public key cryptography doesn't. + +We have a lot of ideas, but we don't have any solutions yet to fix this. Essentially, the question is how to use existing public key primitives to create strong cryptographic groups, where membership and permissions are based on keys and not arbitrary server-maintained access control lists. + +Most of the interesting work in this area has been done by companies working on secure file backup/sync/sharing, such as Wuala and Spideroak. Unfortunately, there are not yet any good open protocols or free software packages that can handle group cryptography. + +At the moment, probably the best approach is the simple approach: a protocol where the client encrypts each message to each recipient individually, and has some mechanism to verify the transcript to ensure that all parties received the same messages. + +There is some free software work on some of interesting building blocks that could be useful in building group cryptography. For example: + +* [Proxy re-encryption](https://en.wikipedia.org/wiki/Proxy_re-encryption): This allows the server to re-encrypt to new recipients without gaining access to the cleartext. The [SELS mailing list manager](http://sels.ncsa.illinois.edu/) uses OpenPGP to implement a [clever scheme for proxy re-encryption](http://spar.isi.jhu.edu/~mgreen/proxy.pdf). +* [Ring signatures](https://en.wikipedia.org/wiki/Ring_signature): This allows any member of a group to sign, withing anyone knowing which member. + +### Resource problem + +The problem: + +> There are no open protocols to allow users to securely share a resource. + +For example, when using secure chat or secure federated social networking, you need some way to link to external media, such as an image, video or file, that has the same security guarantees as the message itself. Embedding this type of resource in the messages themselves is prohibitively inefficient. + +We don't have a proposal for how to address this problem. There are a lot of great initiatives working under the banner of read-write-web, but these do not take encryption into account. In many ways, solutions to the resource problem are dependent on solutions to the the group problem. + +As with the group problem, most of the progress in this area has been by people working on encrypted file sync (e.g. strategies like Lazy Revocation and Key Regression). + diff --git a/pages/docs/tech/hard-problems/pt.md b/pages/docs/tech/hard-problems/pt.md new file mode 100644 index 0000000..50c0541 --- /dev/null +++ b/pages/docs/tech/hard-problems/pt.md @@ -0,0 +1,133 @@ +@title = 'Problemas difíceis na comunicação segura' +@nav_title = 'Problemas difíceis' +@summary = "Como o LEAP aborda os problemas difíceis na comunicação segura" + +## Os sete grandes + +Se você pesquisar iniciativas interessantes para a criação de formas mais seguras de comunicação, verá que surge um padrão: aparentemente toda tentativa séria de construir um sistema para transmissão de mensagens seguras eventualmente se depara com a seguinte lista de sete problemas difíceis: + +1. **Problema da autenticidade**: a validação de chaves públicas é muito difícil para ser gerenciada por usuários, mas sem isso não é possível obter confidencialidade. +2. **Problema dos metadados**: os protocolos existentes são vulneráveis à análise de metadados, mesmo que os metadados muitas vezes sejam mais sensíveis do que o conteúdo da comunicação. +3. **Problema da assincronicidade**: para estabelecer comunicação criptografada, atualmente é necessário escolher entre sigilo futuro (forward secrecy) e a habilidade de se comunicar de forma assíncrona. +4. **Problema do grupo**: na prática, pessoas trabalham em grupos, mas a criptografia de chave pública não. +5. **Problema dos recursos**: não existem protocolos abertos que permitam aos usuários compartilharem um recurso de forma segura. +6. **Problema da disponibilidade**: as pessoas querem alternar suavemente entre dispositivos e restaurar seus dados se perderem um dispositivo, mas isso é bem difícil de se fazer com segurança. +7. **Problema da atualização**: quase que universalmente, atualizações de software são feitas de maneiras que são convidativas a ataques e comprometimento de dispositivos. + +Tais problemas parecem estar presentes independentemente da abordagem arquitetônica escolhida (autoridade centralizada, peer-to-peer distribuído ou servidores federados). + +É possível ignorar muitos desses problemas se você não se importar especificamente com a usabilidade ou com o conjunto de funcionalidades com as quais os/as usuários/as se acostumaram nos métodos contemporâneos de comunicação online. Mas se você se importa com a usabilidade e recursos, então você terá que encontrar soluções para esses problemas. + +## Nossas soluções + +Em nosso trabalho, o LEAP tentou enfrentar diretamente esses sete problemas. Em alguns casos, chegamos a soluções sólidas. Noutros, estamos avançando com medidas paliativas temporárias e investigando soluções de longo prazo. Em dois casos não temos nenhum plano atual para lidar com os problemas. + +### O problema da autenticidade + +O problema: + +> A validação de chaves públicas é muito difícil para ser gerenciada por usuários, mas sem isso não é possível obter confidencialidade. + +Se a validação de chaves adequada é um pressuposto para uma comunicação segura, mas é muito difícil para a maioria dos usuários/as, que esperança temos? Desenvolvemos um sistema federado único chamado [Nicknym](/nicknym) que descobre e valida automaticamente as chaves públicas, permitindo ao usuário tirar partido de criptografia de chave pública sem saber nada sobre chaves ou assinaturas. + +O protocolo padrão que existe hoje para solucionar este problema chama-se [DANE](https://en.wikipedia.org/wiki/DNS-based_Authentication_of_Named_Entities). O DANE pode ser a melhor opção no longo prazo, mas atualmente é difícil de ser configurado, difícil de ser utilizado por clientes, vaza informações sobre associação para um observador da rede, e depende da confiança na zona raíz do DNS e nas zonas TLD. + +### Problema dos metadados + +O problema: + +> Os protocolos existentes são vulneráveis à análise de metadados, mesmo que os metadados muitas vezes sejam mais sensíveis do que o conteúdo da comunicação. + +Como medida de curto prazo, estamos integrando transporte criptografado oportunístico (TLS) para email e mensagens de chat ao serem retransmitidas entre servidores. Há dois aspectos importantes nisso: + +* Servidores repetidores (relaying servers) precisam de uma maneira sólida para descobrir e validar as chaves uns dos outros. Para isso, estamos utilizando inicialmente DNSSEC/DANE. +* Um atacante não deve ser capaz de fazer o downgrade do transporte criptografado para texto não cifrado. Para isso, estamos modificando o software para assegurar que o transporte criptografado não possa sofrer downgrade. + +Tal abordagem é potencialmente eficaz contra observadores externos na rede, mas não protege os metadados dos próprios provedores de serviços. Além disso, ela não protege, por si só, contra ataques mais avançados que envolvam análise de tráfego e de tempo. + +No longo prazo, pretendemos adotar um dos vários esquemas distintos para roteamento seguro de metadados. Estes incluem: + +* Pareamento automático de pseudônimos (auto-alias-pairs): cada uma das partes autonegocia pseudônimos para se comunicarem umas com as outras. Nos bastidores, o cliente utiliza de forma invisível esses pseudônimos para a comunicação subsequente. A vantagem é que isso é compatível com o roteamento existente. A desvantagem é que o servidor do usuário/a armazena uma lista de seus pseudônimos. Como uma melhoria, pode-se adicionar a possibilidade de usar um serviço de terceiros para manter o mapa dos pseudônimos. +* Cabeçalhos de roteamento do tipo "cebola" (onion-routing-headers): uma mensagem de um/a usuário/a para o/a usuário/a B é codificada de forma que as informações de roteamento do destinatário/a contenham apenas o nome do servidor usado por B. Quando o servidor de B recebe a mensagem, decodifica um cabeçalho adicional que contém o utilizador real "B". Como o uso de pseudônimos, isso não proporciona benefícios se os usuários estão no mesmo servidor. Como uma melhoria, a mensagem pode ser encaminhada por meio de servidores intermediários. +* Caixa de depósito de terceiros (third-party dropbox): para trocar mensagens, o/a usuário/a A e o/a usuário/a B negociam uma URL única de uma "caixa de depósito" (dropbox) para depositar mensagens, potencialmente usando um agente intermediário. Para enviar uma mensagem, o usuário A depositaria a mensagem na caixa. Para receber uma mensagem, o usuário B acessaria regularmente esta URL para ver se há novas mensagens. +* Misturador com assinaturas (mixmaster-with-signatures): as mensagens são enviadas através de um conjunto de repetirores anonimizadores do tipo mixmaster e ao final são entregues ao servidor do destinatário. O programa cliente do usuário apenas exibe a mensagem se ela for criptografada, tiver uma assinatura válida, e se o usuário tiver adicionado anteriormente o remetente a uma 'lista de permissões' (talvez gerada automaticamente a partir da lista de chaves públicas validadas). + +Para uma boa discussão comparando redes misturadoras com roteamento cebola, veja a [postagem no blog de Tom Ritter](https://ritter.vg/blog-mix_and_onion_networks.html) sobre o tema. + +### Problema da assincronicidade + +O problema: + +> Para estabelecer comunicação criptografada, atualmente é necessário escolher entre sigilo futuro (forward secrecy) e a habilidade de se comunicar de forma assíncrona. + +Com o ritmo de crescimento do armazenamento digital e da criptanálise, o sigilo futuro é cada vez mais importante. Caso contrário, qualquer comunicação criptografada que você fizer hoje possivelmente se tornará uma comunicação em texto não cifrado num futuro próximo. + +No caso do email e do bate-papo, existem o OpenPGP para email e o OTR para bate-papo: o primeiro fornece recursos assíncronos e o segundo fornece sigilo futuro, mas nenhum deles possuem ambas as habilidades. Precisamos tanto de uma melhor segurança para email quanto da capacidade de enviar e receber mensagens de bate-papo em modo offline. + +No curto prazo, estamos empilhando transporte de email com sigilo futuro e relay de chat em cima de criptografia tradicional de objetos (OpenPGP). Esta abordagem é idêntica à nossa abordagem paliativa para o problema dos metadados, com o acréscimo de que os servidores repetidores precisam ter a capacidade de não apenas negociar transporte TLS mas também de negociar cifras que suportem sigilo futuro e que evitem uma precarização (downgrade) da cifra utilizada. + +Esta abordagem é potencialmente eficaz contra os observadores externos na rede, mas não obtém sigilo futuro dos próprios prestadores de serviço. + +No longo prazo, pretendemos trabalhar com outros grupos para criar novos padrões de protocolo de criptografia que podem ser tanto assíncronos quanto permitir o sigilo futuro: + + * [Extensões para sigilo futuro para o OpenPGP](http://tools.ietf.org/html/draft-brown-pgp-pfs-03). + * [Handshake Diffie-Hellman triplo com curvas elípticas](https://whispersystems.org/blog/simplifying-otr-deniability/). + +### Problema do grupo + +O problema: + +> Na prática, as pessoas trabalham em grupos, mas a criptografia de chave pública não. + +Temos um monte de ideias, mas não temos ainda uma solução para corrigir este problema. Essencialmente, a questão é como usar primitivas de chaves públicas existentes para criar grupos criptográficos fortes, onde a adesão e as permissões são baseadas em chaves e em listas de controle de acesso mantidas no lado do servidor. + +A maioria dos trabalhos interessantes nesta área tem sido feitos por empresas que trabalham com backup/sincronização/compartilhamento seguro de arquivos, como Wuala e Spideroak. Infelizmente, ainda não há quaisquer protocolos abertos bons ou pacotes de software livre que possam lidar com criptografia para grupos. + +Neste momento, é provável que a melhor abordagem seja a abordagem simples: um protocolo no qual o cliente criptografa cada mensagem para cada destinatário individualmente, e que tenha algum mecanismo para verificação da transcrição de forma a garantir que todas as partes tenham recebido a mesma mensagem. + +Existem alguns trabalhos em software livre com blocos construtivos interessantes que podem ser úteis na construção da criptografia para grupos. Por exemplo: + + * [Re-criptografia de proxy (proxy re-encryption)](https://en.wikipedia.org/wiki/Proxy_re-encryption): permite que o servidor cifre o conteúdo para novos beneficiários sem que tenha acesso ao texto não cifrado. O [gerenciador de lista de discussão SELS](http://sels.ncsa.illinois.edu/) usa OpenPGP para implementar um [sistema inteligente para o proxy de re-encriptação](http://spar.isi.jhu.edu/~mgreen/proxy.pdf). + * [Assinaturas em anel (ring signatures)](https://en.wikipedia.org/wiki/Ring_signature): permite que qualquer membro do grupo assine, sem que se possa saber qual membro fez a assinatura. + +### Problema dos recursos + +O problema: + +> Não existem protocolos abertos que permitam aos usuários compartilharem um recurso de forma segura. + +Por exemplo, ao usar um bate-papo seguro ou rede social segura federada, você precisa de alguma forma de criar links para uma mídia externa, como uma imagem, vídeo ou arquivo, que tenha as mesmas garantias de segurança que a própria mensagem. A incorporação deste tipo de recurso nas mensagens em si é proibitivamente ineficiente. + +Nós não temos uma proposta de como resolver este problema. Há um monte de grandes iniciativas que trabalham sob a bandeira da read-write-web, mas que não levam em conta a criptografia. De muitas maneiras, as soluções para o problema dos recursos são dependentes de soluções para o problema do grupo. + +Tal como acontece com o problema do grupo, a maior parte do progresso nesta área tem sido por pessoas que trabalham em sincronização de arquivos criptografados (por exemplo, estratégias como a Revogação Preguiçosa -- Lazy Revocation -- e Regressão de Chave -- Key Regression). + +### Problema da disponibilidade + +O problema: + +> As pessoas querem alternar suavemente entre dispositivos e restaurar seus dados se perderem um dispositivo, mas isso é bem difícil de se fazer com segurança. + +Os/as usuários/as atuais exigem a capacidade de acessar seus dados em múltiplos dispositivos e de terem em mente que dados não serão perdidos para sempre se perderem um dispositivo. No mundo do software livre, só o Firefox abordou este problema adequadamente e de forma segura (com o Firefox Sync). + +No LEAP, temos trabalhado para resolver o problema de disponibilidade com um sistema que chamamos de [Soledad](/soledad) (um acrônimo em inglês para "sincronização, entre dispositivos, de documentos criptografados localmente"). Soledad dá ao aplicativo cliente um banco de dados de documentos sincronizado, pesquisável e criptografado. Todos os dados são criptografados no lado do cliente, tanto quando ele é armazenado no dispositivo local quanto quando é sincronizado com a nuvem. Até onde sabemos, não há nada parecido com isso, seja no mundo do software livre ou comercial. + +Soledad tenta resolver o problema genérico da disponibilidade de dados, mas outras iniciativas tentaram abordar o problema mais específico das chaves privadas e da descoberta de chaves públicas. Estas iniciativas incluem: + +* [O protocolo proposto por Ben Laurie para armazenamento de segredos na nuvem](http://www.links.org/files/nigori/nigori-protocol-01.html). +* [Código para armazenamento de chaves na nuvem](https://github.com/mettle/nilcat), experimental e similar ao anterior. +* [Comentários de Phillip Hallam-Baker sobre questões similares](http://tools.ietf.org/html/draft-hallambaker-prismproof-key-00). + +### O problema da atualização + +O problema: + +> Quase que universalmente, atualizações de software são feitas de maneiras que são convidativas a ataques e comprometimento de dispositivos. + +O triste estado das atualizações de segurança é especialmente problemático porque os ataques de atualização já podem ser comprados prontos por regimes repressores. O problema de atualização de software é especialmente ruim em plataformas desktop. No caso aplicativos em HTML5 ou para dispositivos móveis, as vulnerabilidades não são tão terríveis, mas os problemas também são mais difíceis de corrigir. + +Para resolver o problema da atualização, o LEAP está adotando um sistema de atualização exclusivo chamado Thandy do projeto Tor. Thandy é complexo para administrar, mas é muito eficaz na prevenção de ataques de atualização conhecidos. + +Thandy, e o projeto relacionado [TUF](https://updateframework.com/), são projetados para dar conta das muitas [vulnerabilidades de segurança em sistemas de atualização de software](https://updateframework.com/projects/project/wiki/Docs/Security) existentes. Num exemplo, outros sistemas de atualização sofrem de uma incapacidade do cliente de confirmar que possuem a cópia mais recente, abrindo assim uma enorme vulnerabilidade onde o atacante simplesmente espera por uma atualização de segurança, evita que o upgrade ocorra e lança um ataque para a exploração da vulnerabilidade que deveria ter acabado de ser corrigida. Thandy/TUF fornecem um mecanismo único para a distribuição e verificação de atualizações de modo que nenhum dispositivo cliente irá instalar a atualização errada ou perder uma atualização sem saber. + +Um problema relacionado com o problema da atualização é o problema do backdoor: como você sabe que uma atualização não tem um backdoor adicionado pelos próprios desenvolvedores do software? Provavelmente, a melhor abordagem é aquela tomada pelo [Gitian](https://gitian.org/), que fornece um "processo de construção determinística para permitir que vários construtores criem binários idênticos". Nós pretendemos adotar o Gitian no futuro. diff --git a/pages/docs/tech/infosec/_table-style.haml b/pages/docs/tech/infosec/_table-style.haml new file mode 100644 index 0000000..c9a3495 --- /dev/null +++ b/pages/docs/tech/infosec/_table-style.haml @@ -0,0 +1,59 @@ +%style + :sass + table.properties + td + vertical-align: top + padding-bottom: 0.75em + th + vertical-align: top + padding-right: 1em + text-align: right + font-weight: normal + font-style: italic + + table.infosec + width: 100% + border-collapse: collapse + tbody + border-top: 1px solid black + border-bottom: 1px solid black + th span.normal + font-weight: normal + th.first, th.second, th.cell + width: 14.285714286% + th.spacer, td.spacer + width: 1px !important + padding: 0 !important + background: black + // border: 1px dotted black + //border: 0 !important + //th.second + // width: 0% + //th.cell + // width: 20% + td.cell + border-top: 1px solid black + border-bottom: 1px solid black + //border-right: 1px dotted rgba(0,0,0,.1) + border-left: 1px dotted rgba(0,0,0,.1) + text-align: center + padding: 4px + &.none + //background-color: #ccc + background: #888 + &.low, &.lower, &.worse + //background-color: #FFCCCC + background: #aaa + &.medium, &.higher + //background-color: #FFFFCC + background: #ccc + &.high, &.better + //background-color: #CCFFCC + background: #e6e6e6 + &.better, &.worse + font-weight: bold + tr.footer td + border-left: 1px dotted rgba(0,0,0,.1) + text-align: center + font-size: 0.8em + diff --git a/pages/docs/tech/infosec/_table.haml b/pages/docs/tech/infosec/_table.haml new file mode 100644 index 0000000..6981dac --- /dev/null +++ b/pages/docs/tech/infosec/_table.haml @@ -0,0 +1,233 @@ +:ruby + table_type = locals[:table_type] + if table_type == :small + ## + ## SMALL TABLE + ## + columns = [:p2p, :ssilo, :sfed] + column_data = { + :ssilo => [:silo, :encrypted], + :sfed => [:federation, :encrypted], + :p2p => [:peer_to_peer, :encrypted] + } + rows = [:availability, :usability, :compatibility, :authenticity, :control, :anonymity] + row_groups = [] + footer = false + cells = { + :ssilo => { + :control => [:lower], + :compatibility => [:lower], + :usability => [:higher], + :authenticity => [:lower], + :availability => [:higher], + :anonymity => [:lower] + }, + :sfed => { + :control => [:higher], + :compatibility => [:higher], + :usability => [:lower], + :authenticity => [:higher], + :availability => [:lower], + :anonymity => [:lower] + }, + :p2p => { + :control => [:higher], + :compatibility => [:lower], + :usability => [:lower], + :authenticity => [:higher], + :availability => [:lower], + :anonymity => [:higher] + } + } + elsif table_type == :big + ## + ## BIG TABLE + ## + columns = [:silo, :fed, :ssilo, :sfed, :p2p] + column_data = { + :silo => [:silo, :cleartext, :silo_example], + :fed => [:federation, :cleartext, :fed_example], + :ssilo => [:silo, :encrypted, :ssilo_example], + :sfed => [:federation, :encrypted, :sfed_example], + :p2p => [:peer_to_peer, :encrypted, :p2p_example], + :spacer => [:spacer, :spacer, :spacer] + } + rows = [ + :control, :compatibility, :usability, + :anonymity, :unmappability, :authenticity, + :availability, :confidentiality, :integrity + ] + row_groups = [:message_security, :identity_security, :user_freedom] + row_groups_data = { + :user_freedom => [:control, :compatibility, :usability], + :identity_security => [:authenticity, :anonymity, :unmappability], + :message_security => [:confidentiality, :integrity, :availability] + } + footer = true + cells = { + :silo => { + :control => [:none], + :compatibility => [:none], + :usability => [:high], + :anonymity => [:none], + :unmappability => [:none], + :authenticity => [:none], + :availability => [:high], + :confidentiality => [:none], + :integrity => [:none] + }, + :fed => { + :control => [:medium], + :compatibility => [:high], + :usability => [:medium], + :anonymity => [:none], + :unmappability => [:none], + :authenticity => [:none], + :availability => [:medium], + :confidentiality => [:none], + :integrity => [:none] + }, + :ssilo => { + :control => [:none], + :compatibility => [:none], + :usability => [:high], + :anonymity => [:low], + :unmappability => [:none], + :authenticity => [:none], + :availability => [:high], + :confidentiality => [:high], + :integrity => [:high] + }, + :sfed => { + :control => [:medium], + :compatibility => [:medium], + :usability => [:low], + :anonymity => [:low], + :unmappability => [:none], + :authenticity => [:low], + :availability => [:medium], + :confidentiality => [:high], + :integrity => [:high] + }, + :p2p => { + :control => [:high], + :compatibility => [:none], + :usability => [:low], + :anonymity => [:medium], + :unmappability => [:medium], + :authenticity => [:low], + :availability => [:low], + :confidentiality => [:high], + :integrity => [:high] + }, + :spacer => { + :control => [:spacer], + :compatibility => [:spacer], + :usability => [:spacer], + :anonymity => [:spacer], + :unmappability => [:spacer], + :authenticity => [:spacer], + :availability => [:spacer], + :confidentiality => [:spacer], + :integrity => [:spacer] + } + } + elsif table_type == :leap + ## + ## LEAP TABLE + ## + columns = [:fed, :sfed, :leap] + column_data = { + :ssilo => [:silo, :encrypted], + :sfed => [:federation, :encrypted], + :p2p => [:peer_to_peer, :encrypted], + :fed => [:federation, :cleartext], + :leap => [:leap, :encrypted] + } + rows = [ + :control, :compatibility, :usability, + :anonymity, :unmappability, :authenticity, + :availability, :confidentiality, :integrity + ] + row_groups = [:message_security, :identity_security, :user_freedom] + row_groups_data = { + :user_freedom => [:control, :compatibility, :usability], + :identity_security => [:authenticity, :anonymity, :unmappability], + :message_security => [:confidentiality, :integrity, :availability] + } + footer = false + cells = { + :fed => { + :control => [:medium], + :compatibility => [:high], + :usability => [:medium], + :anonymity => [:none], + :unmappability => [:none], + :authenticity => [:none], + :availability => [:medium], + :confidentiality => [:none], + :integrity => [:none] + }, + :sfed => { + :control => [:medium], + :compatibility => [:medium], + :usability => [:low], + :anonymity => [:low], + :unmappability => [:none], + :authenticity => [:low], + :availability => [:medium], + :confidentiality => [:high], + :integrity => [:high] + }, + :leap => { + :control => [:medium], + :compatibility => [:worse], + :usability => [:better], + :anonymity => [:low], + :unmappability => [:better], + :authenticity => [:better], + :availability => [:medium], + :confidentiality => [:high], + :integrity => [:high] + } + } + end + +%table.infosec + %tr + %th.first + - if row_groups.any? + %th.second + - columns.each do |column| + - if column == :spacer + %th.spacer + - else + %th.cell + = I18n.t(column_data[column][0], :scope => 'infosec') + %br<> + %span.normal + = I18n.t(column_data[column][1], :scope => 'infosec') + - if row_groups.any? + - row_groups.each do |row_group| + %tbody + - rows = row_groups_data[row_group] + - rows.each do |row| + %tr + - if rows.first == row + %td{:rowspan=>3}= I18n.t(row_group, :scope => 'infosec').sub(' ', '<br/>') + %td= I18n.t(row, :scope => 'infosec') + - columns.each do |column| + %td.cell{:class => cells[column][row]}= I18n.t(cells[column][row].first, :scope => 'infosec') + - else + - rows.each do |row| + %tbody + %tr + %td= I18n.t(row, :scope => 'infosec') + - columns.each do |column| + %td.cell{:class => cells[column][row]}= I18n.t(cells[column][row].first, :scope => 'infosec') + - if footer + %tr.footer + %td{:colspan=>2}= I18n.t(:for_example, :scope => 'infosec') + - columns.each do |column| + %td= I18n.t(column_data[column][2], :scope => 'infosec') + diff --git a/pages/docs/tech/infosec/en.haml b/pages/docs/tech/infosec/en.haml new file mode 100644 index 0000000..6b042ec --- /dev/null +++ b/pages/docs/tech/infosec/en.haml @@ -0,0 +1,105 @@ +- @title = "Architecture comparison" +- @nav_title = "Architecture comparison" +- @summary = "A comparison of the trade-offs made by different communication archectures" + += render :partial => 'table-style' + +%h1.first You can't have it all + +%p Every messaging architecture makes certain design choices that privilege one property of information security over another. Although there is no intrinsically necessary trade off among different information security properties, when we examine the technical limitations of actual implementations we see clearly that existing architectures are structurally biased toward certain properties and against others. + +%h1 A fancy table + +%p This table provides a rough comparison of the choices made by common messaging architectures. See #{link 'below for details' => '#table-notes'} regarding the column and row headings. + +.p + %b Table 1. Information security of common messaging architectures + = render partial: 'table', locals: {table_type: :big} + +%p Reasonable people may disagree: this table represents one defensible assessment of the various architecture categories. Many people would adjust one or two cells, but on the whole we believe this table is a fair and accurate comparison. Some squares get low marks because of user error. For example, peer-to-peer systems have a hard time with user friendly keys, leading to high user error and low effective authenticity. + +%p In table 2 we see a simplified representation that highlights the relative differences between the encrypted architectures: + +.p + %b Table 2. Relative trade-offs of encrypted messaging architectures + = render partial: 'table', locals: {table_type: :small} + +%p Relatively better is not necessarily good. For example, federated and peer-to-peer models have better authenticity than silo models, but still in practice have many authenticity problems. + +%h1 The LEAP strategy + +%p In a nutshell, the LEAP strategy is this: take a federated architecture and improve the authenticity, unmappability, and usability. In table form, that looks like this: + +.p + %b Table 3. The LEAP strategy for improving federated architectures + = render partial: 'table', locals: {table_type: :leap} + +%p Why this focus on authenticity, unmappability, and usability? + +%p First, there is a lot of room for improvement. We believe that there is actually no particular structural reason why these properties are so low in existing federated encrypted architectures. + +%p Second, these property are extremely important and yet are typically given low priority or are ignored completely. + +%ul + %li + %b Authenticity: + Message security rests entirely on a foundation of authenticity. Without proper validation of encryption keys, you cannot be assured of confidentiality or integrity. Unfortunately, current system of establishing message authenticity are so difficult to use that most users simply ignore this step. LEAP will address these problems with a system of #{link 'strong and automatic identity validation' => 'nicknym'}. + %li + %b Usability: + There are plenty of high security tools that are nearly impossible for the common user to use correctly. If the tool is too difficult, it will not be widely adopted and will likely be used incorrectly by those who do adopt it. LEAP with address these problems with the #{link 'LEAP client' => 'client'} that is tightly coupled with the server-side software and is autoconfiguring. + %li + %b Unmappability: + Recent advances in social network analysis and the greatly expanded of ability state and corporate actors to gather social graph information have made unmappability an urgent requirement for any architecture that seeks to address the surveillance situation we face today. LEAP will address these problems with our proposal for #{link 'graph resistant routing' => 'routing'}. + +%p Improvement in these areas will come at a price. Although LEAP communication tools will be backward compatible with existing federated standards, a user of the LEAP platform will not have the same degree of choice in client software and provider as does a user of a traditional federated system. Our goal is to actively help providers adopt the LEAP platform, in order to give users more options in the long run. + +%h1#table-notes Decoding the table + +%h2 Communication architectures (columns) + +(to be written) + +%h2 Aspects of information security (rows) + +%p Classical information security consists of a trio of properties: confidentiality, integrity, availability. To this list, others have added authenticity, control, and anonymity (among many others). + +%p For our purposes here, we also add usability, compatibility, and unmappability. What do all these mean? Let's use the example of a single message, and group these nine properties in three categories: + +%h3 Message Security + +%table.properties + %tr + %th Confidentiality + %td A message has highly confidentiality if only the intended recipients are able to read the message. + %tr + %th Integrity + %td A message has high integrity if the recipient is assured the message has not been altered. + %tr + %th Availability + %td A message has high availability if the user is able to get to the message when they so desire. + +%h3 Identity Security + +%table.properties + %tr + %th Authenticity + %td A message has high authenticity if the recipient is certain who sent the message and the sender is certain who received it. + %tr + %th Anonymity + %td A message has high anonymity if the identity of the sender cannot be established by examining the message or the pattern of message delivery. + %tr + %th Unmappability + %td A message has high unmappability if the social network that you communicate with cannot be easily discovered. Unmappability is often collapsed under anonymity. This is unfortunate. It is true the anonymity is one of the issue at stake with social network mapping, but it is just one of many. Because of recent advances in social network analysis and the ability to gather social graph information, we feel that unmappability deserves to be highlighted on its own. + +%h3 User Freedom + +%table.properties + %tr + %th Control + %td If a user is in possession of their own data and can do with it exactly what they want (and others cannot use the data in ways contrary to the wishes of the user), then we say that they have high control. + %tr + %th Usability + %td For a communication system to have high usability, the common user must be able to operate the system in a way that does not compromise their security. + %tr + %th Compatibility + %td For a system to have high compatibility, the user must not be locked into a particular provider or technology, but should have competing and compatible options available to them. In other words, a user's data should be portable and they should have a choice of clients. diff --git a/pages/docs/tech/limitations.md b/pages/docs/tech/limitations.md new file mode 100644 index 0000000..77592f3 --- /dev/null +++ b/pages/docs/tech/limitations.md @@ -0,0 +1,123 @@ +@title = 'Known Limitations' +@toc = true +@summary = 'Known limitations, issues, and security problems with the LEAP platform' + +Herein lie the know limitations, issues, and security problems of the LEAP platform and Bitmask client application. + +Provider problems +========================================== + +Meta-data can be recorded by the provider +------------------------------------------------- + +Currently, the service provider is able to observe the meta-data routing information of messages in transit of their own users (email and chat). This information is not stored, but a nefarious provider could observe this information in transit and record it. + +We have several [plans to eliminate this](/routing), but these are not part of the initial release. + +Incoming cleartext email can be recorded by the provider +--------------------------------------------------------------------- + +Currently, if an incoming email is not already encrypted, the provider encrypts the email to the recipient's public key. + +Potentially, a compromised or nefarious service provider could alter the LEAP software to keep a copy of these cleartext emails. Over time, as more people send encrypted email, this will become less of an issue. Providers will simply see fewer and fewer incoming cleartext emails. + +The provider can undermine the security of the web application +------------------------------------------------------------------------- + +Both the client and the web application use something called SRP (Secure Remote Password) to prevent the server from ever seeing a cleartext copy of the password. This is in contrast to normal password systems, where the password is hashed on the server, so the server could record a copy of the password when it is initially set. + +However, because all the javascript cryptographic libraries used by the user's web browser to perform the SRP negotiation are loaded from the provider's server, a nefarious or compromised provider could give the user's browser bad libraries that secretly sent a cleartext copy of the user's password. + +There are three methods that can be used to prevent this: + +* We could offer the option of first visiting a third party website that loads the authentication libraries before redirecting to the provider's website. Unfortunately, this user experience is a bit awkward. +* We could allow providers the option of not allowing authentication or signup through the webapp. Instead, the client could authenticate with the provider's session API, receiving a session token, and then pass this token to the web browser. +* Currently, the web application is needed for email settings, help tickets and billing, but potentially these functions could be rolled into the Bitmask client. + +It is not either/or, we could support a combination of options. + +The details for both of these are bit more tricky than in these simple descriptions, because of the need to work around the single origin policy, but it is still entirely possible to do this securely (using either CORS or PostMessage). + +Still possible to brute force the password verifier +----------------------------------------------------------------- + +SRP (Secure Remote Password) does not store a hash of the password in the provider's user database, but instead a password verifier. However, if an attacker gains access to the database of password verifiers (plus salts) they can perform a brute force attack similar to a normal attack on a database of password hashes. The attack in the case of SRP is more difficult, since there is much more cryptographic work involved, but it is still possible. + +To mitigate exposure of the password verifiers, we plan to separate them out into a separate database with only a separate single-purpose authentication API daemon granted access. Additionally, we hope to offer a password-less option that uses OpenPGP smart cards. + +The provider can observe VPN traffic +-------------------------------------------------- + +The "Encrypted Internet" feature of LEAP currently works using OpenVPN secure tunnel to proxy all network traffic to a gateway operated by the provider. Once the traffic exits this gateway, it is cleartext (unless otherwise encrypted on the client, e.g. HTTPS) and could be observed and record by the provider, or any network observer able to monitor all traffic into and out of the gateway. + +This limitation is mitigated by having the LEAP client authenticate with the VPN gateway using semi-anonymous or anonymous client certificates. A nefarious or compromised provider could attempt to record the moment that a user fetches a new client certificate, and record the IP address or authentication credentials of the user at that time. + +In the future, we plan to remove this vulnerability in two ways: + +* Allow the client to fetch new client certificates using blind signatures, so that there is no way for the provider to associate user with certificate, but we can also ensure that only valid users get client certificates. +* Use Tor as an alternate and optional transport for "Encrypted Internet". From the stand point of the user, it would work the same (using perhaps tun socks proxy and dnscrypt-proxy). This option would be slower and not support UDP traffic, but it would be much more secure. The Tor project prefers that every application that uses Tor be specifically designed for Tor so that it does not leak information in other ways. Using Tor as a default route like we do with OpenVPN would violate this, but would be more user friendly. + +Device problems +================================== + +A compromised device is a sad device +---------------------------------------------- + +The LEAP client tries to minimize attacks related to physical access to the device. In particular, we try to be very resistant to offline attacks, where an attacker has captured the user's device while the LEAP client does not have an open session. For example, locally stored data is kept in an encrypted database that is only unlocked when the user authenticates with the application. + +However, if an attacker gains access to the device, and then the device is returned to the user, they can do all kinds of nasty things, like install a keylogger that captures every keystroke. + +This vulnerability is true of all software, not just LEAP, but it is worth noting. + +Mail clients cache data in cleartext +-------------------------------------------------- + +Currently, LEAP relies on the use of a standard email client like Thunderbird, Apple Mail, or Outlook. Although all LEAP data is stored encrypted on the user's device, these mail clients cache and index email data in the clear on their own. + +To fix this problem, we have two plans: + +* Write plugins for Thunderbird, Apple Mail, and Outlook to make the integration with the Bitmask client more easy and to automatically configure the email client to not-cache or index. +* Distribute a custom email client with the Bitmask application, perhaps based on mailpile. + +The Bitmask application provides a client-encrypted searchable database, so it should be possible to get the same functionality provided by the indexing done by the existing mail clients. + +User problems +================================= + +Passwords are never going to be very good +--------------------------------------------------- + +LEAP relies on the user's password to unlock access to the user's client encrypted data storage. It does this the right way, using a solid KDF, but many users choose passwords that are weak, offering marginal security if an attacker gains offline access to the user's encrypted storage (for example, if they obtain the device). + +In the future, we hope to add support for OpenPGP smart cards in order to overcome many of the problems associated with passwords. + +Design problems +============================================ + +Enumeration of usernames +----------------------------- + +The system LEAP uses to validate the public keys of users is inherently vulnerable to an attacker enumerating usernames. Because requests for public keys may be proxy'ed through other providers, there is no good method of preventing an attacker from launching many queries for public keys and eventually mapping most of the usernames. + +This is unfortunate, but this is also a problem with all other such systems of key discovery and validation (i.e. DANE). For now, we consider this to be an acceptable compromise. + +Much trust is placed in LEAP +------------------------------------------- + +In order to shield the service provider from being pressured by a host government or criminal organization to add a backdoor into the client, the model with the LEAP platform is that the client is normally downloaded from the leap.se website and subsequent updates are signed by LEAP developers. + +This is good for the provider, but not so good for LEAP, since this system could potentially place pressure on LEAP. Because LEAP does not have a provider-customer relationship with any user, LEAP cannot target compromised applications for particular users. LEAP could, however, introduce a backdoor in the client used by all users. + +To prevent this, we plan to adopt [Gitian](https://gitian.org/) or something equivalent. Gitian allows for a way to standardize the entire build environment and build process in order for third parties to be able to verify that the released binary application does indeed match the correct source code. + +External authority problems +================================================= + +Certificate authorities considered dangerous +--------------------------------------------------- + +The long term goal with LEAP is to entirely rid ourselves of reliance on the x.509 certificate authority system. However, there are a few places where the platform still relies on it: + +* When the client first validates a new provider, it will assume the provider's TLS connection is valid if presented with a server certificate signed by a commercial CA recognized by the operating system. Subsequent connections to the provider's API use pinned certificates. +* When a nicknym agent discovers new public keys for users, it uses a TLS connection validated by a commercial CA recognized by the operating system. In the future, nicknym responses will also be signed, eliminating some of the vulnerability. +* Currently, the web application does not get deployed with any other TLS validation than the standard commercial CA method. Eventually, we plan to support [DNS-based Authentication of Named Entities (DANE)](https://datatracker.ietf.org/wg/dane/), [Trust Assertions for Certificate Keys (TACK)](http://tack.io/), [Public Key Pinning Extension for HTTP](https://datatracker.ietf.org/doc/draft-ietf-websec-key-pinning/), or [Certificate Transparency](http://www.certificate-transparency.org/) (whatever gets the most traction). diff --git a/pages/docs/tech/routing.md b/pages/docs/tech/routing.md new file mode 100644 index 0000000..46af93f --- /dev/null +++ b/pages/docs/tech/routing.md @@ -0,0 +1,65 @@ +@title = "Graph Resistant Routing" +@summary = "LEAP's plans for protecting routing meta-data" + +# A social graph is highly sensitive data + +As messages are sent and delivered, they contain "meta-data" describing where these messages should be routed. With existing protocols for email and chat, this meta-data is sent in the clear and can be used to build a social graph of how people interact. + +As the field of network analysis has advanced in recent years, the social graph has become highly sensitive and critical information. Knowledge of the social graph can give an attacker a blueprint into the inner workings of an organization or reveal surprisingly intimate personal details, such as sexual orientation or even health status. + +In the short term, LEAP is opportunistically encrypting the message transport whenever possible. This protects the meta-data routing information from an external observer, but does not protect against a nefarious or compromised service provider. This page is about our plans for a better way. + +# Possible solutions + +There are four strategies we might employ to add protection of routing information to email and chat: + +* Auto-alias-pairs: Each party auto-negotiates aliases for communicating with each other. Behind the scenes, the client then invisibly uses these aliases for subsequent communication. The advantage is that this is backward compatible with existing routing. The disadvantage is that the user's server stores a list of their aliases. As an improvement, you could add the possibility of a third party service to maintain the alias map. +* Onion-routing-headers: A message from user A to user B is encoded so that the "to" routing information only contains the name of B's server. When B's server receives the message, it unwraps (unencrypts) a supplementary header that contains the actual user "B". Like aliases, this provides no benefit if both users are on the same server. As an improvement, the message could be routed through intermediary servers. +* Third-party-dropbox: To exchange messages, user A and user B negotiate a unique "dropbox" URL for depositing messages, potentially using a third party. To send a message, user A would post the message to the "dropbox". To receive a message, user B would regularly polls this URL to see if there are new messages. +* Mixmaster-with-signatures: Messages are bounced through a mixmaster-like set of anonymization relays and then finally delivered to the recipient's server. The user's client only displays the message if it is encrypted, has a valid signature, and the user has previously added the sender to a 'allow list' (perhaps automatically generated from the list of validated public keys). +* Direct-delivery: Instead of relaying messages from client to server to server to client, in this model the sender's client delivers the message directly to the recipient's server. This delivery would need to happen over an anonymization network akin to Tor, or through proxies set up for this purpose. In order to prevent spam, the recipient server would only accept messages delivered in this manner if the message was signed using a group signature (this ensures that the server doesn't know who the sender is, but can confirm that they are allowed to deliver to a particular user). This would require advanced confirmation on the part of both users that they may send messages to one another. This is how Pond works. + + +None of these are currently used for email or chat. + +# Auto alias pairs + +How would auto alias pairs work? Imagine users Alice and Bob. Alice wants to correspond with Bob but doesn't want either her service provider or Bob's service provider to be able to record the association. She also doesn't want a network observer to be able to map the association. + +When Alice first sends a message to Bob, Alice's client will initiate the following chain of events on her behalf, automatically and behind the scenes. + +* Alice's client requests a new alias from her service provider. If her normal address is alice@domain.org, she receives an alias hjj3fpwv84fn@domain.org that will forward messages to her account. +* Alice's client uses automatic key discovery and validation to find Bob's public key and discover if Bob's service provider supports map resistant routing. +* If Bob does support it, Alice's client will then send a special message, encrypted to Bob's key, that contains Alice's public address and her private alias. +* When Bob's client encounters this special message, it records a mapping between Alice's public address (alice@domain.org) and the private alias she has created for Bob's use (hjj3fpwv84fn@domain.org). +* Bob's client then creates an alias for Bob and sends it to Alice. +* Alice's client receives this alias, and records the mapping between Bob's public address and his private alias. +* Alice's client then relays the original message to Bob's alias. + +Subsequently, whenever Alice or Bob want to communicate, they use the normal public addresses for one another, but behind the scenes their clients will rewrite the source and recipient of the messages to use the private aliases. + +This scheme is backwards compatible with existing messaging protocols, such as SMTP and XMPP. + +## Limitations + +There are five major limitations to this scheme: + +1. **Alias unmasking attacks:** because each service provider maintains an alias map for their own users, an attacker who has gained access to this alias map can de-anonymize all the associations into and out of that particular provider, for the past and future. +2. **Timing attacks:** a statistical analysis of the time that messages of a particular size are sent, received, and relayed by both service providers could reveal the map of associations. +3. **Log correlation problem:** a powerful attacker could gain access to the logs of both service providers, and thereby reconstruct the associations between the two providers. +4. **Single provider problem:** this scheme does not protect associations between people on the same service provider. +5. **Client problem:** if the user's device is compromised, the record of their actual associations can be discovered. + +At this point, we feel that it is OK to live with these limitations for the time being in order to simply the logic of the user's client application and to ensure backward compatibility with existing messaging protocols. + +Possible future enhancements could greatly mitigate these attacks: + +* We could use temporarily aliases that rotate daily, perhaps based on an HMAC counter, based on a shared secret between two users. +* The 'alias service' could be run by a third party, so that providers don't have access to the alias maps (thus migitating problems 1, 3, and 4). +* A service provider with sufficient traffic could be in a very good position to be able to aggregate and time-shift the messages it relays in order to disrupt timing attacks. + +Ultimately, however, most of these attacks are a problem when faced with an extremely powerful adversary. This scheme is not designed for these situations. Instead, it is designed to prevent the casual, mass surveillance of all communication that currently takes place in repressive and democratic countries alike, by both governments and corporations. It greatly reduces the capacity for association mapping of both traffic over the wire and in stored communication. It is not designed to make a particular user anonymous when specifically targeted by a powerful adversary. + +# How would X work? + +How would onion-routing-headers, third-party-dropbox, or mixmaster-with-signatures work? To be written. diff --git a/pages/docs/tech/secure-email/en.md b/pages/docs/tech/secure-email/en.md new file mode 100644 index 0000000..fe1f51f --- /dev/null +++ b/pages/docs/tech/secure-email/en.md @@ -0,0 +1,578 @@ +@title = "Secure Email Report" +@summary = "A report on the state of the art in secure email projects" +@toc = false + +There are an increasing number of projects working on next generation secure email or email-like communication. This is an initial draft report highlighting the projects and comparing the approaches. Please help us fill in the missing details and correct any inaccuracies. To contribute to this document, fork repository found at https://github.com/OpenTechFund/secure-email and issue a pull request. + +Contents: + +1. [Common Problems](#common-problems) + 1. [Key Management](#key-management) + 1. [Metadata Protection](#metadata-protection) + 1. [Forward Secrecy](#forward-secrecy) + 1. [Data Availability](#data-availability) + 1. [Secure Authentication](#secure-authentication) +1. [Web Mail](#web-mail) + 1. [Lavaboom](#lavaboom) + 1. [Mega](#mega) + 1. [PrivateSky](#privatesky) + 1. [Scramble](#scramble) + 1. [Startmail](#startmail) + 1. [Whiteout](#whiteout) +1. [Browser Extensions](#browser-extensions) + 1. [Mailvelope](#mailvelope) +1. [Mail Clients](#mail-clients) + 1. [Bitmail](#bitmail) + 1. [Mailpile](#mailpile) + 1. [Parley](#parley) +1. [Self-Hosted Email](#self-hosted-email) + 1. [Dark Mail Alliance](#self-hosted-dark-mail) + 1. [FreedomBox](#freedombox) + 1. [Mailpile](#self-hosted-mailpile) + 1. [Mail-in-a-box](#mail-in-a-box) + 1. [kinko](#kinko) +1. [Email Infrastructure](#email-infrastructure) + 1. [Dark Mail Alliance](#dark-mail-alliance) + 1. [LEAP Encryption Access Project](#leap) +1. [Post-email alternatives](#post-email-alternatives) + 1. [Bitmessage](#bitmessage) + 1. [Bote mail](#bote-mail) + 1. [Cables](#cables) + 1. [Dark Mail Alliance](#p2p-dark-mail-alliance) + 1. [Enigmabox](#enigmabox) + 1. [FlowingMail](#flowingmail) + 1. [Goldbug](#goldbug) + 1. [Pond](#pond) +1. [Related Works](#related-works) + +<a name="common-problems"></a>Common Problems +=========================================================== + +All of the technologies listed here face a common set of problems when trying to make email (or email-like communication) secure and easy to use. These problems are hard, and have defied easy solutions, because there are no quick technological fixes: at issue is the complex interaction between user experience, real world infrastructure, and security. Although no consensus has yet emerged on how best to tackle any of these problems, the diversity of projects listed in this report reflect an surge of interest in this area and an encouraging spirit of experimentation. + +<a name="key-management"></a>Key Management +----------------------------------------------------------- + +All the projects in this report use public-key encryption to allow a user to send a confidential message to the intendant recipient, and for the recipient to verify the authorship of the message. Unfortunately, public-key encryption is notoriously difficult to use properly, even for advanced users. The very concepts are confusing for most users: public key versus private key, key signing, key revocation, signing keys versus encryption keys, bit length, and so on. + +Traditionally, public key cryptography for email has relied on either the X.509 Certificate Authority (CA) system or a decentralized "Web of Trust" (WoT) for key validation (authenticating that a particular person owns a particular key). Recently, both schemes have come under intense criticism. Repeated security lapses at many of the Certificate Authorities have revealed serious flaws in the CA system. On the other hand, in an age where we better understand the power of social network analysis and the sensitivity of the social graph, the exposure of metadata by a "Web of Trust" is no longer acceptable from a security standpoint. + +This is where we are now: we have public key technology that is excessively difficult for the common user, and our only methods of key validation have fallen into disrepute. The projects listed here have plunged into this void, attempting to simplify the usage of public-key cryptography. These efforts have four elements: + +* Key discovery: There is no commonly used standard for discovering the public key attached to a particular email address. All the projects here that use OpenPGP intend to initially use, as a stop-gap measure, the OpenPGP keyservers for key discovery, although the keyserver infrastructure was not designed to be used in this way. +* Key validation: If not Certificate Authorities or Web of Trust, what then? Nearly every project here uses Trust On First Use (TOFU) in one way or another. With TOFU, a key is assumed to be the right key the first time it is used. TOFU can work well for long term associations and for people who are not being targeted for attack, but its security relies on the security of the discovery transport and the application's ability to retain a memory of discovered keys. TOFU can break down in many real-world situations where a user might need to generate new keys or securely communicate with a new contact. The projects here are experimenting with TOFU in different ways, and these problems can likely be mitigated by combining TOFU with other measures. +* Key availability: Almost every attempt to solve the key validation problem turns into a key availability problem, because once you have validated a public key, you need to make sure that this validation is available to the user on all the possible devices they might want to send or receive messages on. +* Key revocation: What happens when a private key is lost, and a user want to issue a new public key? None of the projects in this report have an answer for how to deal with this in a post-CA and post-WoT world. + +The projects that use a public key as a unique identifier do not have the key validation problem, because they do no need to try to bind a human memorable identifier to a long non-memorable public key: they simply enforce the use of the public key as the user's address. For example, rather than `alice@example.org` as the identifier, these systems might use `8b3b2213ff00e5fb684b003d005ed2fb`. In place of the key validation problem, this approach raises the key exchange problem: how do two parties initially exchange long public keys with one another? This approach is taken by all the P2P projects listed here (although there do exist some P2P application that don't use public key identifiers). + +Some of the major experimental approaches to solving the problem of public key discovery and validation include: + +1. Inline: Many of the projects here plan to simply include the user's public key as an attachment to every outgoing email (or in a footer or SMTP header). +1. DNS: Key distributed via DNSSEC, where a service provider adds a DNS entry for each user containing the user's public key or fingerprint. +1. Append-only log: Proposal to modify Certificate Transparency to handle user accounts, where audits are performed against append-only logs. +1. Network perspective: Validation by key endorsement (third party signatures), with audits performed via network perspective. +1. Introductions: Discovery and validation of keys through acquaintance introduction. +1. Mobile: Although too lengthy to manually transcribe, an app on a mobile device can be used to easily exchange keys in person (for example, via a QR code or bluetooth connection). + +<a name="metadata-protection"></a>Metadata Protection +----------------------------------------------------------- + +Traditional schemes for secure email have left metadata exposed. We now know that metadata is often more sensitive than message content: metadata is structured data, easily stored forever, and subject to powerful techniques of social network analysis that can can be incredibly revealing. + +Metadata protection, however, is **hard**. In order to protect metadata, the message routing protocol must hide the sender and recipient from all the intermediaries responsible for relaying the message. This is not possible with the traditional protocol for email transport, although it will probably be possible to piggyback additional (non-backward compatible) protocols on top of traditional email transport in order to achieve metadata protection. + +Alternately, some projects reject traditional email transport entirely. These decentralized peer-to-peer approaches to metadata protection generally fall into four camps: (1) directly relay the message from sender's device to recipients device; (2) relay messages through a network of friends; (3) broadcast messages to everyone; (4) relay messages through an anonymization network such as Tor. The first two approaches protect metadata, but at the expense of increasing vulnerability to traffic analysis that could reveal the same metadata. The third solution faces serious problems of scalability. Pond uses the fourth method, discussed below. + +All schemes for metadata protection face the prospect of increasing Spam (since one of the primary methods used to prevent Spam is analysis of metadata). This is why some schemes with strong metadata protection make it impossible to send or receive messages to anyone you are not already in contact with. This works brilliantly for reducing Spam, but is unlikely to be a viable long term strategy for entirely replacing the utility of email. + +<a name="forward-secrecy"></a>Forward Secrecy +----------------------------------------------------------- + +Forward secrecy is a security property that prevents an attacker from saving messages today and then later decrypting these messages once they have captured the user's private key. Without forward secrecy, an attacker is more likely to be able to capture messages today and simply wait for computers to become powerful enough to crack the encryption by brute force. Traditional email encryption offers no forward secrecy. + +All methods for forward secrecy involve a process where two parties negotiate an ephemeral key that is used for a short period of time to secure their communication. In many cases, the ephemeral key is generated anew for every single message. Traditional schemes for forward secrecy are incompatible with the asynchronous nature of email communication, since with email you still need to be able to send someone a message even if they are not online and ephemeral key generation requires a back and forth exchange between both parties. + +There are several new experimental (and tricky) protocols that attempt to achieve both forward secrecy and support for asynchronous communication, but none have yet emerged as a standard. These protocols either (1) require an initial bootstrap message that is not forward secret, (2) require an initial synchronous exchange to start the process, or (3) rely on a pool of pre-generated ephemeral key pairs that can be used on first contact. When the continually changing ephemeral key for a conversation is lost by either party, then the initialization stage is performed again. + +Another possible approach is to use traditional encryption with no support for forward secrecy but instead rely on a scheme for automatic key discovery and validation in order to frequently rotate keys. This way, a user could throw away their private key every few days, achieving a very crude form of forward secrecy. + +<a name="data-availability"></a>Data Availability +----------------------------------------------------------- + +Users today demand data availability: they want to be able to access their messages and send messages from any device they choose, wherever they choose, and whenever they choose. Most importantly, they don't want the loss of any particular device to result in a loss of all their data. For insecure communication, achieving data availability is dead simple: simply store everything in the cloud. For secure communication, however, we have no proven solutions to this problem. As noted above, the key management problem is also really a data availability problem. + +Most of the email projects here have postponed dealing with the data availability problem. A few have used IMAP to synchronize data or developed their own secure synchronization protocol. Several of the email-like P2P approaches rely on a P2P network for data availability. + +<a name="secure-authentication"></a>Secure Authentication +----------------------------------------------------------- + +For those projects that make use of a service provider, one of the key problems is how to authenticate securely with the service provider without revealing the password (since the password is probably also used to encrypt the private key and other secure storage, so it is important that the service provider does not have cleartext access as with typical password authentication schemes). The possible schemes include: + +* Separate passwords. The application can use one password for authentication and a separate password for securing secrets. +* Pre-hash the password on the client before sending it to the server. This method can work, although it does not also authenticate the server (an impostor server can always reply with a success message), and is still vulnerable to brute force dictionary attacks. +* Use Secure Remote Password (SRP), a type of cryptographic zero-knowledge proof designed for password authentication in which the client and server mutually authenticate. SRP has been around a while, and is fairly well analyzed, but it is still vulnerable to brute force dictionary attacks (albeit much less than traditionally password schemes). +* Sign a challenge from the server with the user's private key. This has the advantage of being nearly impossible to brute force attack, but is vulnerable to impostor server providers and requires that the user's device has the private key. + +No consensus or standard has yet emerged, although SRP has been around a while. + +<a name="web-mail"></a>Web Mail +=========================================================== + +Most users are familiar with web-based email and the incredible convenience it offers: you can access your email from any device, and you don't need to worry about data synchronization. Developers of web-based email faces several difficult challenges when attempting to make a truly secure application. These challenges can be overcome, but not easily. + +First, because the web application is loaded from the web server each time you use it, the service provider could be targeting a version of the client to you that includes a backdoor. To overcome this vulnerability, it is possible to load the code for the web application from a third party. There are two ways of doing this: + +1. App Store: Most web browsers support special extensions in the form of "Browser Applications". These are loaded from some kind of app store and installed on the user's device. In this case, the third party that provides the application is the app store. Therefore, the user is then relying on the app store to furnish them with a secure version of the app. For example, this is the approach taken by [cryptocat](https://crypto.cat). +2. Third Party: There are two advanced mechanisms to allow a web application to be loaded from one website and allow it to access data from another website. One is called CORS ([Cross-origin resource sharing](https://en.wikipedia.org/wiki/Cross-origin_resource_sharing)) and the other is HTML5's [window.postMessage command](https://developer.mozilla.org/en-US/docs/Web/API/window.postMessage). With either method, anyone can be the third party furnishing the application, or it can be self hosted. For example, this is the approach taken by [Unhosted](https://unhosted.org). + +Second, even if the application is loaded from a trusted third party, web browsers are not an ideal environment for sensitive data: there are many ways for an in-browser application to leak data and web browsers are notoriously prone to security holes (it is a very difficult problem to be able to run untrusted code locally in a secure sandbox). To their credit, the browser developers are often vigilant about fixing these holes (depending on who you ask), but the browser environment is far from a secure computing environment. It continues to be, however, the most convenient environment. + +Third, developers of web-based secure email face an additional challenge when dealing with offline data or data caching. Modern HTML5 apps typically store a lot of data locally on the user's device using the localStorage facility. Currently, however, no browser stores this encrypted. A secure web-based email application must either choose to not support any local storage, or develop a scheme for individually encrypting each object put in localStorage, a process which is very inefficient. Even storing keys temporarily in short lived session storage is problematic, since these can be easily read from disk later. + +These challenges to do not apply to downloaded mail clients that happen to use HTML5 as their interface (Mailpile, for example). + +<a name="lavaboom"></a>Lavaboom +----------------------------------------------------------- + +[lavaboom.com](https://www.lavaboom.com) + +Lavaboom is a new web-based mail provider from Germany using client-side encryption in the browser. No further details are available at this time. + +Lavaboom's name is a tribute to the shuttered Lavabit service, although Lavaboom has no affiliation or people in common with Lavabit. + +<a name="mega"></a>Mega +----------------------------------------------------------- + +[mega.co.nz](https://mega.co.nz) + +The relaunch of Mega has featured client-side encryption via a javascript application running in the browser. Mega has announced plans to extend their offerings to include email service with a similar design. No details are yet forthcoming. In interviews, Mega has said the javascript running in the browser will be open source, but the server component will be proprietary. + +<a name="privatesky"></a>PrivateSky +----------------------------------------------------------- + +PrivateSky was a secure web-based email service that chose to shut down because their design was not compatible with UK law. Many in the press have [said GCHQ forced the closure](http://www.ibtimes.co.uk/articles/529392/20131211/gchq-forced-privatesky-secure-email-service-offline.htm), which the [company refutes](http://www.certivox.com/blog/bid/359788/The-real-story-on-the-PrivateSky-takedown). + +<a name="startmail"></a>StartMail +----------------------------------------------------------- + +[startmail.com](http://startmail.com) + +The makers of the secure search engine [startpage.com](https://startpage.com) have announced they will be providing secure email service. + +Despite the tag line as the "world's most private email," StartMail is remarkably insecure. It offers regular IMAP service and a webmail interface that supports OpenPGP, but the user still must trust StartMail entirely. For example when you authenticate, your password string is sent to StartMail, and new OpenPGP keypairs are generated on the server, not the client. The website also makes some dubious statements, such as claiming to be more secure because their TLS server certificate supports extended validation. + +Verdict: oil of snake + +<a name="scramble"></a>Scramble +----------------------------------------------------------- + +[scramble.io](https://scramble.io) + +Scramble is an OpenPGP email application that can be loaded from a website (with plans to add app store support). Additionally, you can sign up for email service from scramble.io. + +**Keys:** Private keys are generated in the browser app, encrypted with the user's passphrase, and then stored on the server. The server never sees the user's passphrase (password is hashed using scrypt before sent to the server during account creation and authentication). The master storage secret (symmetric key) used to encrypt keys is stored in the browser's sessionStorage, which is erased when the user logs out. Keys are validated using notaries. + +**Infrastructure:** Scramble uses a system of network perspectives to discover and validate public keys. The client will come with a list of pre-blessed notaries that can be used to query for public keys. If the notaries agree, the client will consider the key to be validated. + +**Application:** Currently, Scramble is a traditional HTML5 javascript application loaded from the website. In the future, Scramble will also be an installable browser app. + +* Written in: Go, Javascript +* Source code: https://github.com/dcposch/scramble +* Design documentation: https://github.com/dcposch/scramble/wiki/Scramble-Protocol +* License: LGPL +* Platforms: Windows, Mac, Linux (with Android planned). + +<a name="whiteout"></a>Whiteout +----------------------------------------------------------- + +[whiteout.io](https://whiteout.io) + +Whiteout is a commercial service featuring an HTML5-based OpenPGP email client that is loaded from the web. + +* Written in: Javascript +* Source code: https://github.com/whiteout-io/mail-html5 +* License: proprietary, but the code is available for inspection. + +<a name="browser-extensions"></a>Browser Extensions +=========================================================== + +A browser extension modifies the behavior of the web browser (not to be confused with a browser application, which has far fewer permissions and consists of a self-contained application). Browser extensions are able to modify how the user interacts with a variety of websites. Browser extensions share many of the same advantages and disadvantages of [web mail approaches](#webmail). + +<a name="mailvelope"></a>Mailvelope +----------------------------------------------------------- + +[mailvelope.com](http://mailvelope.com) + +Mailvelope is a browser extension that allows you to use OpenPGP email with traditional web-mail providers like Gmail, Yahoo, and Outlook.com. + +**Keys:** The private key is generated for you, password protected, and stored in the browser's local storage (along with public keys). In the future, the plan is to support automatic discovery and validation of public keys using OpenPGP keyservers and message footers. + +**Application:** When the extension detects you have opened a web page from a supported web-mail provider such as Gmail, it offers the user the opportunity to encrypt what you type in the compose window and decrypt messages you receive. + +**Limitations:** Because of an inherent limitation in the way Mailvelope can interface with web-mail, it is not able to send OpenPGP/MIME (although it can read it fine). As mentioned elsewhere, browser storage is not a particular ideal place to be storing keys. When a web-mail provider changes their UI (or API if they happen to have one), the extension must be updated to handle the new format. + +* Contact: info@mailvelope.com +* Written in: Javascript +* Source code: https://github.com/toberndo/mailvelope +* Design documentation: http://www.mailvelope.com/help +* License: AGPL +* Platforms: Windows, Mac, Linux (with Android planned). + +<a name="mail-clients"></a>Mail Clients +=========================================================== + +An email client, or MUA (Mail User Agent), provides a user interface to access email from any service provider. Traditional examples of email clients include Thunderbird or Microsoft Outlook (although both these application include a lot of other functionality as well). Nearly all email clients communicate with the email service provider using IMAP or POP and SMTP, although some also support local mailboxes in mbox or Maildir format. + +There are two primary advantages to the mail client approach: + +1. Existing accounts: By using a custom secure mail client, a user can continue to use their existing email accounts. +1. Tailored UI: A custom client has the potential to rethink the email user experience in order to better convey security related details to the user. + +The mail client approach, however, also has several disadvantages: + +1. Insecure service providers: A mail client cannot address many of the core problems with email security when used with a traditional email provider. For example, metadata will not be protected in storage or transit, and the provider cannot aid in key discovery or validation. Most importantly, many existing mail providers are highly vulnerable, since few rely on DNSSEC for their MX records or validate their StartTLS connections for mail relay (when they even bother to enable StartTLS). A traditional email provider also requires authentication via password that is seen by the provider in clear text, and might be recorded by them. Finally, most service providers retain significant personally identifiable information, such as IP address of clients. +1. Install a new app: As with many of the other approaches, the custom mail client approach requires that users download and install a specialized application on their device before they can use it. + +Ultimately, the level of email security that is possible with the custom mail client approach will always be limited. However, custom email clients may be an excellent strategy for gradually weening users away from email and to a different and more secure protocol. Most of the projects in this section see email support as a gateway to ease the transition to something that can replace email. + +<a name="bitmail"></a>Bitmail +----------------------------------------------------------- + +[bitmail.sf.net](http://bitmail.sf.net) + +Bitmail is a desktop application that provides a user interface for traditional IMAP-based mail, but also supports a custom peer-to-peer protocol for relaying email through a network of friends. Bitmail will support both OpenPGP and S/MIME. + +**Keys:** Keys are validated using a shared secret or fingerprint validation. Public keys are discovered over the P2P network. Keys are stored locally in an encrypted database. + +**Routing:** Bitmail uses an opportunistic message distribution model where every message is sent to every neighbor. It is called "Echo" and it is very similar to the protocols used by Retroshare and Briar. + +**Application:** Bitmail uses the Qt library for cross platform UI. + +There are also plans to include a Bitmail MUA extension. + +* Written in: C +* Source code: http://sourceforge.net/projects/bitmail/files +* Design documentation: http://sourceforge.net/p/bitmail/code/HEAD/tree/branches/BitMail.06.2088_2013-11-03/BitMail/branches/BitMail/Documentation/ +* License: GPL v2 +* Platforms: Windows (with Mac and Linux planned). + +Note: I am unclear which of the previous features are planned and which are currently working. + +<a name="mailpile"></a>Mailpile +----------------------------------------------------------- + +[mailpile.is](http://mailpile.is) + +Mailpile is an email client designed to quickly handle large amounts of email and also support user-friendly encryption. The initial focus is on email, with plans to eventually support post-email protocols like bitmessage, flowingmail, or darkmail. Also, the developers hope to add support for XMPP-based chat in the future. Since the Mozilla foundation has not committed the resources necessary to keep Thunderbird contemporary, the Mailpile initiative holds a lot of promise as a cross-platform mail client that seeks to redesign how we interact with email. + +**Keys:** Mailpile email encryption is based on OpenPGP (it uses your GPG keyring). Key discovery will be handled using OpenPGP keyservers and including public keys as attachments to outgoing email. Public keys are trusted on first use, with plans for validation via DANE and manual fingerprint verification (future support for a P2P protocol might include additional methods, such as Certificate Transparency or Short Authentication Strings). Currently, keys are not backed up. + +**Application:** Mailpile UI is written using HTML5 and Javascript, running against a self-hosted Python application (that typically lives locally on the device, but might be running on your own server). + +**Limitations:** Mailpile does not currently have a scheme for recovery if your device is destroyed or a method for securely synchronizing keys among devices. Although the search index is stored encrypted on disk (if the user already has GPG installed and a key pair generated), it is encrypted in a way that requires the index to be loaded entirely into memory. Mailpile is under very active development, so these and other issues may change in the near future. + +* Written in: Python, Javascript +* Source code: http://github.com/pagekite/mailpile +* Design documentation: http://github.com/pagekite/mailpile +* License: AGPL & Apache +* Platforms: Windows, Mac, Linux (with Android and iOS planned). +* Contact: team@mailpile.is + +<a name="parley"></a>Parley +----------------------------------------------------------- + +[parley.co](https://parley.co) + +Parley is a desktop mail client with a UI written using HTML5 and Javascript, with a local backend written in Python. + +**Keys:** Although Parley can be used with any service provider, the Parley servers are used to publish public keys and back up client-encrypted private keys. For key discovery, Parley uses a central repository and the OpenPGP keyservers. For key validation, Parley relies on trust on first use and Parley key endorsement. + +**Infrastructure:** All users of the Parley client also sign up for the Parley service, but they use their existing email account. The Parley server acts as a proxy that uses [context.io](http://context.io) for email storage (context.io is a commercial service that provides a REST API for a traditional IMAP account). The Parley server also handles key discovery, validation, and backup. Both the client and server are released as free software. + +**Application:** Parley is currently bundled into an executable using [Appcelerator](http://www.appcelerator.com/). The Parley client does not speak IMAP or SMTP directly. Rather, uses the email REST API of context.io. + +**Limitations:** All user email is stored by context.io, albeit in OpenPGP format. Metadata is exposed to context.io, however (in addition to your service provider). + +* Written in: Python, Javascript +* Source code: https://github.com/blackchair/parley +* Design documentation: https://parley.co/#how-it-works +* License: BSD +* Platforms: Windows, Mac, Linux (with Android and iOS planned). + +<a name="self-hosted-email"></a>Self-Hosted Email +=========================================================== + +Traditionally, email is a federated protocol: when you send an email it travels from your computer, to the server of your email provider, to the server of the recipient's provider, and finally to the recipient's computer. The key idea with self-hosted email is to cut out the middleman and run your own email server. + +In the United States, much of the interest in self-hosted email is driven by the Supreme Court's current (and particularly odd) interpretation of the 4th amendment called the ["Third-Party Doctrine"](https://en.wikipedia.org/wiki/Third-Party_Doctrine). Essentially, you have much weaker privacy protections in the US if you entrust any of your data to a third party. Additionally, the Court has so far afforded much greater protections to items physically inside your home. "Aha!" say the hackers and the lawyers, "we will just put email in the home." + +Unfortunately, it is not so simple. There are some major challenges to putting email servers in everyone's home: + +* **Delegated reputation**: The current email infrastructure is essentially a system of delegated reputation. In order to be able to send mail to most providers and not have a large percentage of it marked as Spam, a service provider must gradually build up a good reputation. Users are able to send mail because their provider has cultivated this reputation and maintained it by closing abusive accounts. It is certainly possible to run an email provider with a single user, but it is much harder to build up a good reputation. Also, many email providers block all relay attempts from IP addresses that have been flagged as "home" addresses, on the (probable) assumption that the message is coming from a virus and not a legitimate email server. +* **Servers are on a hostile network**: Because a server needs to have open ports that are publicly accessible from the internet at all times, running one is much trickier than a simple desktop computer. It is much more critical to make sure security upgrades are applied in a timely manner, and that you are able to respond to external attacks, such as "Spam Bombs". Any publicly addressable IP that is put on the open internet will be continually probed for vulnerabilities. Self-hosting will probably work great for a protocol like Pond, where there are strict restrictions on who may deliver incoming messages. Email, however, is a protocol that is wide open and prone to abuse. +* **Sysadmins are not robots**: No one has yet figured out how to make self-healing servers that don't require a skilled sysadmin to keep them healthy. Once someone does, a lot of sysadmins will be out of work, but they are presently not very worried. There are many things that commonly go wrong with servers, such as upgrades failing, drives filling up, daemons crashing, memory leaks, hardware failures, and so on. +* **Does not address the important problems**: Moving the physical location of a device does nothing to solve the hard problems associated with easy-to-use email security (such as data availability and key validation). Some of the approaches to these problems rely on service provider infrastructure that would be infeasible to self host. +* **DNS is hard**: One of the important security problems with traditional email is the vulnerability MX DNS records. Doing DNS correctly is hard, and not something that can be expected of the common user. + +Self-hosted email is an intriguing "legal hack", albeit one that faces many technical challenges. + +<a name="self-hosted-dark-mail"></a>Dark Mail Alliance +----------------------------------------------------------- + +The Dark Mail Alliance has said they want to support self-hosting for the server component of the system. No details yet. + +<a name="freedombox"></a>FreedomBox +----------------------------------------------------------- + +[freedomboxfoundation.org](https://freedomboxfoundation.org) + +From its early conception, part of FreedomBox was "email and telecommunications that protects privacy and resists eavesdropping". Email, however, is not currently being worked on as part of FreedomBox. (as far as I can tell). + +<a name="self-hosted-mailpile"></a>Mailpile +----------------------------------------------------------- + +Although Mailpile is primarily a mail client, the background Python component can read the Maildir format for email. This means you could install Mailpile on your own server running a Mail Transfer Agent (MTA) like postfix or qmail. You would then access your mail remotely by connecting to your server via a web browser. + +<a name="Mail-in-a-box"></a>Mail-in-a-box +----------------------------------------------------------- + +<a href="https://github.com/JoshData/mailinabox">github.com/JoshData/mailinabox</a> + +Mail-in-a-box helps people set up self-hosted email for linux hobbyists and email developers. It will install and configure the necessary Debian packages required to turn a machine running Ubuntu into a self-hosted email server. It provides a fairly straightforward, standard email server with IMAP, SMTP, greylisting, DKIM and SPF. It also includes a command line tool for adding and removing accounts. + +**Advantages:** Something quick for anyone with some linux skill who wants to experiment with email. + +**Limitations:** Setting up an email server is the easy part, maintaining the service over time is the tricky part. Without any automation recipes using something like Puppet, Chef, Salt, or CFEngine, mail-in-a-box is unlikely to be useful to anyone but the curious hobbyist. + +* Written in: Bash +* Source code: https://github.com/JoshData/mailinabox +* License: CC0 1.0 Universal + +<a name="kinko"></a>kinko +----------------------------------------------------------- + +[kinko](https://kinko.me) implements an en/decrypting SMTP- and IMAP-proxy on ARM-class hardware, the kinko box. Emails are synced from the users' email accounts via IMAP to the box and are stored in plaintext in a secure storage area on the box. The kinko box also includes a webmailer to be able to use email with the browser. + +Connections to the kinko box are secured by TLS using a private key only known to the box itself. Furthermore, the kinko box is tunnelled to a public internet location. Consequently, users can access secure email from everywhere, using IMAP compatible email clients and/or browsers, including mobile clients. + +kinko uses GnuPG for encryption, with the addition of encrypting the email subject. Further additions should allow "Post-email alternatives" (a la bitmessage) to be used with the email clients that users are using today already. Other, privacy-related additions are planned as well. + +**Key discovery and validation:** Users can upload existing PGP keyrings. PGP keys are discovered via email +addresses, email content, and PGP key servers. Keys are trusted on first use (but this policy can be changed +to explicit fingerprint validation.) + +**Project status:** An alpha prototype exists. We are preparing for the release of a beta package in Q2/2014. + +**Languages:** The kinko base system is implemented in ruby and shell, with minor portions in native code. +Applications can be implemented in more or less any language. + +**Webmail:** The currently included webmail application is roundcube webmail. That might change in the future. + +**Licenses:** All portions of the kinko system will be released under the AGPL license. (Included 3rd party +applications will use their respective open source licenses). The hardware is open sourced as +per [olimex](https://www.olimex.com/wiki/A10-OLinuXino-LIME). + +<a name="email-infrastructure"></a>Email Infrastructure +=========================================================== + +The "infrastructure" projects give a service provider the opportunity to offer secure email accounts to end-users. By modifying how both email clients and email servers work, these projects have the potential to deploy greater security measures than are possible with a client-only approach. For example: + +* Encrypted relay: A secure email provider is able to support, and enforce, encrypted transport when relaying mail to other providers. This is an important mechanism for preventing mass surveillance of metadata (which is otherwise not protected by OpenPGP client-side encryption of message contents). +* Easier key management: A secure email provider can endorse the public keys of its users, and provide assistance to various schemes for automatic validation. Additionally, a secure email provider, coupled with a custom client, can make it easy to securely manage and back up the essential private keys which are otherwise cumbersome for most users to manage. +* Invisible upgrade to better protocols: A secure email provider has the potential to support multiple protocols bound to a single user@domain address, allowing automatic and invisible upgrades to more secure post-email protocols when both parties detect the capability. +* A return to federation: The recent concentration of email to a few giant providers greatly reduces the health and resiliency of email as an open protocol, since now only a few players essentially monopolize the medium. Projects that seek to make it easier to offer secure email as a service have the potential to reverse this trend. +* Secure DNS: A secure provider can support DNSSEC and DANE, while most other email providers are unlikely to anytime soon. This is very important, because it is easy to hijack the MX records of a domain without DNSSEC. +* Minimal data retention: A service provider that follows "best practices" will choose to retain less personally identifiable information on their users, such as their home IP addresses. + +The goal of both projects in this category is to build systems where the service provider is untrusted and cannot compromise the security of its users. + +Despite the potential of this approach, there are several unknown factors that might limit its appeal: + +* In order to benefit from a more secure provider, a user will need to switch their email account and email address, a very high barrier to adoption. +* Where once there were many ISPs that offered email service, it is no longer clear if there is either the demand to sustain many email providers or the supply of providers interested in offering email as a service. +* Users must download and install a custom application. + +<a name="dark-mail-alliance"></a>Dark Mail Alliance +----------------------------------------------------------- + +[darkmail.info](https://darkmail.info) + +The Dark Mail Alliance will include both a client application and server software. The plan is to support traditional encrypted email (both OpenPGP and S/MIME), a new federated email-like protocol adapted from SilentCircle's instant message protocol (SCIMP), and a pure peer-to-peer messaging protocol. Both the client and server will be made available as free software. + +**Keys:** Key pairs will be generated on the user's device and uploaded to the service provider. [Certificate Transparency](http://certificate-transparency.org) will be used to automatically validate the service provider's endorsement of these public keys. Dark Mail additionally plans to support fingerprint confirmation, short authentication strings, and shared secret for manual key validation. Automatic discovery of public keys will happen using DNS, HTTPS, and via the messages themselves. + +**Routing:** The post-email messaging protocol promises to have forward secrecy and protection from metadata analysis (details have not yet been posted, and SCIMP does not currently support meta-data protection). Dark Mail Alliance plans to additionally support pure peer-to-peer messaging using a key fingerprint as the user identifier. + +**Infrastructure:** Dark Mail plans to support three types of architectures: traditional client/server, self-hosted, and pure peer-to-peer. No details yet on how these will work. + +**Application:** The client application will work with any existing MUA by exposing a local IMAP/SMTP server that the MUA can connect to. + +**Limitations:** Dark Mail has not yet released any code or design documents. However, they certainly have the resources to carry out their plans. + +* Written in: C +* Source code: none yet +* Design documentation: none yet +* License: planned to be OSI-compatible +* Platforms: initially Android and iOS, followed by Windows, OS X, Linux, and Windows Phone. +* Contact: press@darkmail.info + +<a name="leap"></a>LEAP Encryption Access Project +----------------------------------------------------------- + +[leap.se](https://leap.se) + +LEAP includes both a client application and turn-key system to automate the process of running a secure service provider. Currently, this includes user registration and management, help tickets, billing, VPN service, and secure email service. The secure email service is based on OpenPGP. + +**Keys:** Key pairs are generated on the user's device. Keys, and all user data, are stored in a client-encrypted database that is synchronized among the user's devices and backed up to the service provider. Keys are automatically validated using a combination of provider endorsement and network perspective (coming soon). Keys are discovered via the OpenPGP keyservers, the OpenPGP header, email footers, and a custom HTTP-based discovery protocol. + +**Infrastructure:** LEAP follows a traditional federated client/server architecture. The client is designed to work with any LEAP-compatible service provider (with plans to support legacy IMAP providers in the future). For security reasons, users are encouraged to get the application from LEAP and not their service provider. + +**Application:** The client application works with any existing MUA by exposing a local IMAP/SMTP server that the MUA can connect to. There is a Thunderbird extension to automate configuration of the account in Thunderbird. The client application communicates with the service provider using a custom protocol for synchronizing encrypted databases. The application is a very small C program that launches the Python code. The user interface is written using Qt. + +**Limitations:** In the current implementation, security properties of forward secrecy and metadata production are not end-to-end. Instead, the client relies on the service provider to ensure these properties. This limitation is due to some inherent limitations in the existing protocols for secure email. As with many of the other projects, LEAP's plan is to invisibly upgrade to a post-email protocol when possible in order to overcome these limitations. + +* Written in: Python +* Source code: https://leap.se/source +* Design documentation: https://leap.se/docs +* License: mostly GPL v3, some MIT and AGPL. + +<a name="post-email-alternatives"></a>Post-email alternatives +=========================================================== + +There are several projects to create alternatives to email that are more secure yet still email-like. + +These projects share some common advantages: + +1. **Trust no one:** These projects share an approach that treats the network, and all parties on the network, as potentially hostile and not to be trusted. With this approach, a user's security can only be betrayed if their own device is compromised or the software is flawed or tampered with, but the user is protected from attacks against any service provider (because there typically is not one). +1. **Fingerprint as identifier:** All these projects also use the fingerprint of the user's public key as the unique routing identifier for a user, allowing for decentralized and unique names. This neatly solves the problem of validating public keys, because every identifier basically *is* a key, so there is no need to establish a mapping from an identifier to a key. + +Except for Pond, all these alternatives take a pure peer-to-peer approach. As such, they face particular challenges: + +1. **The "Natural" Network**: Many advocates of peer-to-peer networking advance the notion that decentralized networks are the most efficient networks and are found everywhere in nature (in the neurons in our brain, in how mold grows, in how insects communicate, etc). This notion is only partially true. Networks are found in nature, but these network are not radically decentralized. Instead, natural networks tend to follow a power law distribution (aka "[scale free networks](https://en.wikipedia.org/wiki/Scale-free_network)"), where there is a high degree of partial centralization that balances "brokerage" (ability to communicate far in the network) with "closure" (ability to communicate close in the network). Thus, in practice, digital networks rely on "super hubs" that process most of the traffic. These hubs need to be maintained and hosted by someone, often at great expense (and making the network much more vulnerable to Sybil attacks). +1. **The Internet:** Sadly, the physical internet infrastructure is actually very polycentric rather than decentralized (more akin to a tree than a spider's web). One reason for the rise of cloud computing is that resources are much cheaper near the core of the internet than near the periphery. Technical strategies that attempt to leverage the periphery will always be disadvantaged from an efficiency standpoint. +1. **Traffic Analysis:** Most of the peer-to-peer approaches directly relay messages from sender's device to recipient's device, or route messages through the participant's contacts. Such an approach to message routing makes it potentially very easy for a network observer to map the network of associations, even if the message protocol otherwise offers very strong metadata protection. +1. **Sybil Attacks:** By their nature, peer-to-peer networks do not have a method of blocking participation in the network. This makes them potentially very vulnerable to [Sybil attacks](https://en.wikipedia.org/wiki/Sybil_attack), where an attacker creates a very large number of fake participants in order to control the network or reveal the identity of other network participants. +1. **Mobile:** Peer-to-peer networks are resource intensive, typically with every node in the network responsible for continually relaying traffic and keeping the network healthy. Unfortunately, this kind of thing is murder on the battery life of a mobile device, and requires a lot of extra network traffic. +1. **Identifiers**: Using key fingerprints as unique identifiers has some advantages, but it also makes user identifiers impossible to remember. There is a lot of utility in the convenience of memorable username handles, as evidence in the use of email addresses and twitter handles. +1. **Data Availability**: Unless also paired with a cloud component, peer-to-peer networks have much lower data availability than other approaches. For example, it takes much longer to update message deliveries from a peer network than from a server, particularly when the device has been offline for a while. Also, if a device is lost or destroyed, generally the user loses all their data. + +Most of these challenges have possible technological solutions that might make peer-to-peer approaches the most attractive option in the long run. For example, researchers may discover ways to make P2P networks less battery intensive. For this reason, it is important that research continue in this area. However, [in the long run we are all dead](https://en.wikiquote.org/wiki/John_Maynard_Keynes) and peer-to-peer approaches face serious hurdles before they can achieve the kind of user experience demanded today. + +<a name="bitmessage"></a>Bitmessage +----------------------------------------------------------- + +[Bitmessage](https://bitmessage.org) + +Bitmessage is a peer-to-peer email-like communication protocol. It is totally decentralized and places no trust on any organization for services or validation. + +Advantages: + +* resistant to metadata analysis +* relatively easy to use +* works and is actively used by many people. + +Disadvantages: + +* no forward secrecy +* unsolved scaling issues: all messages are broadcast to everyone +* because there is no forward secrecy, it is especially problematic that anyone can grab an encrypted copy of any message in the system. This means if the private key is ever compromised, then all the past messages can be decrypted easily by anyone using the system. +* relies on proof of work for spam prevention, which is probably not actually that preventative (spammers often steal CPU anyway). + +<a name="bote-mail"></a>Bote mail +----------------------------------------------------------- + +[i2pbote.i2p.us](http://i2pbote.i2p.us) (or [i2pbote.i2p](http://i2pbote.i2p) if using i2p) + +Bote mail (aka [IMail](https://en.wikipedia.org/wiki/IMail)) is an email-like communication protocol that uses the anonymizing network I2p for transport. Bote mail stores messages in a global distribute hash table for up to 100 days, during which time the client has an opportunity to download and store the message. + +**Keys**: Bote mail uses public-key based addresses. You can create as many identities as you want, each identity corresponding to a ECDSA or NTRU key-pair. + +**Application**: Users interact with the user interface through a webmail interface, although the client is running locally. + +* Written in: Java +* License: GPLv3 + +<a name="cables"></a>Cables +----------------------------------------------------------- + +[github.com/mkdesu/cables](https://github.com/mkdesu/cables) + +* Written in: C, Bash +* License: GPL v2 + +<a name="p2p-dark-mail-alliance"></a>Dark Mail Alliance +----------------------------------------------------------- + +The Dark Mail Alliance plans to incorporate traditional email, a federated email alternative, and a second email alternative that is pure peer-to-peer. Details are not yet forthwith. + +<a name="enigmabox"></a>Enigmabox +----------------------------------------------------------- + +[enigmabox.net](https://enigmabox.net) + +Enigmabox is a device that you install on your local network between your computer and the internet. It acts as secure proxy, providing VPN, and communication services analogous to email and VoIP. In order to communicate with another user, they must also have an enigmabox. + +Data is routed peer-to-peer directly from one enigmabox to another using cjdns, a system of virtual mesh networking in which IP addresses are derived from public keys. End to end encryption of messages is provided entirely by the cjdns transport layer. + +With this scheme, message are forward secret, but not entirely asynchronous. At some point, both the sender and recipient must have their enigmaboxes online at the same time. + +<a name="flowingmail"></a>FlowingMail +----------------------------------------------------------- + +[flowingmail.com](http://flowingmail.com) + +P2P secure, encrypted email system. + +<a name="goldbug"></a>Goldbug +----------------------------------------------------------- + +[goldbug.sf.net](http://goldbug.sf.net) + +* Written in: C++, Qt +* License: BSD + +<a name="pond"></a>Pond +----------------------------------------------------------- + +[pond.imperialviolet.org](https://pond.imperialviolet.org/) + +Pond is an email-like messaging application with several unique architectural and cryptographic features that make it stand out in the field. + +**Message Encryption**: Pond uses [Axolotl](https://github.com/trevp/axolotl/wiki) for asynchronous forward secret messages where the key is frequently ratcheted (akin to OTR, but more robust). + +**Routing**: Pond uses a unique architecture where every user relies on a service provider for receiving messages, but sent messages are delivered directly to the recipient's server (over Tor). This allows for strong metadata protection, but does not suffer from the other problems that peer-to-peer systems typically do. In order to prevent excessive Spam under this scheme, Pond uses a clever system of group signatures to allow the server to check if a sender is authorized to deliver to a particular user without leaking any information to the server. + +**Keys**: Pond uses Panda, a system for secure peer validation using short authentication strings. + +Pond's advantages include: + +* Very high security: forward secrecy, metadata protection, resistant to traffic analysis. +* Pond hybrid federated and peer-to-peer approach is cool and holds a lot of promise. +* Written in Go, and thus probably has many fewer security flaws than programs written in C or C++. +* Pond is written by Adam Langley, an extremely well respected crypto-engineer. + +Pond's disadvantages include: + +* Uses non-human memorial unique identifiers, although this is not a necessary element of the design. +* Requires that you set up contacts in advance before you can communicate with them (via a Short Authentication String or full public key exchange). +* Pond is still very difficult to install and use. + +Pond is an exciting experiment in how you could build a very secure post-email protocol. Although Pond currently uses non-human identifiers, Pond could be easily modified to use traditional email username@domain.org identifiers (because it relies on service providers for message reception). The requirement in Pond that both parties pre-exchange keys could also be modified to allow users to set up addresses that could receive messages from anyone, albeit at the cost of likely Spam. Currently, Pond uses Tor to anonymize message routing, but the Tor network was designed for low-latency. Pond could potentially use a more secure anonymization network that was designed for higher-latency asynchronous messages. + +Ultimately, Pond's unique design makes it a very strong candidate for incorporation into a messaging application that could automatically upgrade from email to Pond should it detect that both parties support it. + +* Written in: Go +* Source code: https://github.com/agl/pond +* License: BSD +* Platforms: anything you can compile Go on (for command line interface) or anything you can compile Go + Gtk (for GUI interface). + +<a name="related-works"></a>Related Works +=========================================================== + +There are many technologies that don't belong in this document because they either (a) are not trying to make encrypted email-like communication easier, (b) use some kind of weird proprietary escrow system, or (c) we just don't know enough about them yet. Here is a place to store links to such projects. + +* [Virtru](https://www.virtru.com) has a secure email product that relies on a centralized key escrow. For details, see [here](http://www.theregister.co.uk/Print/2014/01/24/ex_nsa_cloud_guru_email_privacy_startup) and [here](https://www.virtru.com/how-virtru-works). +* [OpenCom](http://opencom.io) is a secure email and email-like communication in the planning stages. +* [Ubiquitous Encrypted Email](https://github.com/tomrittervg/uee) is a protocol draft for standards that could lead to universal adoption of encrypted email. +* [Redecentralize](https://github.com/redecentralize/alternative-internet) has a list of decentralized networks, such as Tor. |