XFTP is a new file transfer protocol focussed on meta-data protection - it is based on the same principles as SimpleX Messaging Protocol used in SimpleX Chat messenger:
asynchronous file delivery - the sender does not need to be online for file to be received, it is stored on XFTP relays for a limited time (currently, it is 48 hours) or until deleted by the sender.
padded e2e encryption of file content.
content padding and fixed size chunks sent via different XFTP relays, assembled back into the original file by the receiving client.
efficient sending to multiple recipients (the file needs to be uploaded only once).
no identifiers or ciphertext in common between sent and received relay traffic, same as for messages delivered by SMP relays.
protection of sender IP address from the recipients.
Open tor configuration with your editor of choice (nano,vim,emacs,etc.):
vim /etc/tor/torrc
And insert the following lines to the bottom of configuration. Please note lines starting with #: this is comments about each individual options.
# Enable log (otherwise, tor doesn't seemd to deploy onion address)
Log notice file /var/log/tor/notices.log
# Enable single hop routing (2 options below are dependencies of third). Will reduce latency in exchange of anonimity (since tor runs alongside xftp-server and onion address will be displayed in clients, this is totally fine)
SOCKSPort 0
HiddenServiceNonAnonymousMode 1
HiddenServiceSingleHopMode 1
# xftp-server hidden service host directory and port mappings
HiddenServiceDir /var/lib/tor/simplex-xftp/
HiddenServicePort 443 localhost:443
Enable systemd service and start tor. Offical tor is a bit flunky on the first start and may not create onion host address, so we're restarting it just in case.
systemctl enable tor && systemctl start tor && systemctl restart tor
Display onion host:
Execute the following command to display your onion host address:
cat /var/lib/tor/simplex-xftp/hostname
Configuration
To see which options are available, execute xftp-server without flags:
sudo su xftp -c xftp-server
...
Available commands:
init Initialize server - creates /etc/opt/simplex-xftp and
/var/opt/simplex-xftp directories and configuration
files
start Start server (configuration:
/etc/opt/simplex-xftp/file-server.ini)
delete Delete configuration and log files
You can get further help by executing su xftp -c "xftp-server <command> -h"
After that, we need to configure xftp-server:
sudo su xftp -c "xftp-server init -h"
...
Available options:
-l,--store-log Enable store log for persistence
-a,--sign-algorithm ALG Signature algorithm used for TLS certificates:
ED25519, ED448 (default: ED448)
--ip IP Server IP address, used as Common Name for TLS online
certificate if FQDN is not supplied
(default: "127.0.0.1")
-n,--fqdn FQDN Server FQDN used as Common Name for TLS online
certificate
-p,--path PATH Path to the directory to store files
-q,--quota QUOTA File storage quota (e.g. 100gb)
-h,--help Show this help text
You should determine which flags are needed for your use-case and then execute xftp-server init:
sudo su xftp -c "xftp-server init -<your flag> <your option>"
to initialize your xftp-server configuration with:
restoring connections when the server is restarted (-l flag),
IP address 192.168.1.5 (--ip flag),
set overall storage quota to 10Gb (-q flag),
store files in /srv/xftp directory (-p flag).
To password-protect your xftp-server, change it in the configuration:
Open configuration with:
sudo su xftp -c "vim /etc/opt/simplex-xftp/file-server.ini"
Under [AUTH] section uncomment create_password and change it:
...
[AUTH]
# Set new_files option to off to completely prohibit uploading new files.
# This can be useful when you want to decommission the server, but still allow downloading the existing files.
new_files: on
# Use create_password option to enable basic auth to upload new files.
# The password should be used as part of server address in client configuration:
# xftp://fingerprint:password@host1,host2
# The password will not be shared with file recipients, you must share it only
# with the users who you want to allow uploading files to your server.
create_password: your_very_secure_password
...
After that, your installation is complete and you should see in your teminal output something like this:
Certificate request self-signature ok
subject=CN = 192.168.1.5
Server initialized, you can modify configuration in /etc/opt/simplex-xftp/file-server.ini.
Run `file-server start` to start server.
----------
You should store CA private key securely and delete it from the server.
If server TLS credential is compromised this key can be used to sign a new one, keeping the same server identity and established connections.
CA private key location:
/etc/opt/simplex-xftp/ca.key
----------
SimpleX XFTP server v0.1.0
Fingerprint: ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=
Server address: xftp://ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=@<hostnames>
The server address above should be used in your client configuration and if you added server password it should only be shared with the other people when you want to allow them to use your server to upload files. If you passed IP address or hostnames during the initialisation, they will be printed as part of server address, otherwise replace <hostnames> with the actual server addresses.
Documentation
All necessary files for xftp-server are located in /etc/opt/simplex-xftp/ folder.
Stored messages, connections, statistics and server log are located in /var/opt/simplex-xftp/ folder.
Location of uploaded files is configured by the user. In our guide we're using /srv/xftp/
Your xftp-server fingerprint of certificate. You can check your certificate fingerprint in /etc/opt/simplex-xftp/fingerprint.
optional<password>
Your configured password of xftp-server. You can check your configured pasword in /etc/opt/simplex-xftp/file-server.ini, under [AUTH] section in create_password: field.
<public_hostname>, optional<onion_hostname>
Your configured hostname(s) of xftp-server. You can check your configured hosts in /etc/opt/simplex-xftp/file-server.ini, under [TRANSPORT] section in host: field.
Systemd commands
To start xftp-server on host boot, run:
sudo systemctl enable xftp-server.service
Created symlink /etc/systemd/system/multi-user.target.wants/xftp-server.service → /etc/systemd/system/xftp-server.service.
To start xftp-server, run:
sudo systemctl start xftp-server.service
To check status of xftp-server, run:
sudo systemctl status xftp-server.service
● xftp-server.service - XFTP server systemd service
Loaded: loaded (/etc/systemd/system/xftp-server.service; enabled; vendor preset: enabled)
Active: active (running) since Sat 2023-03-11 13:11:55 UTC; 1 months 10 days ago
Main PID: 110770 (xftp-server)
Tasks: 14 (limit: 4611)
Memory: 2.4G
CGroup: /system.slice/xftp-server.service
└─110770 /usr/local/bin/xftp-server start +RTS -N -RTS
Feb 27 19:21:11 localhost systemd[1]: Started XFTP server systemd service.
Feb 27 19:21:11 localhost xftp-server[2350]: SimpleX XFTP server v0.1.0
Feb 27 19:21:11 localhost xftp-server[2350]: Fingerprint: ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=
Feb 27 19:21:11 localhost xftp-server[2350]: Server address: xftp://ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=@<hostnames>
Feb 27 19:21:11 localhost xftp-server[2350]: Store log: /var/opt/simplex-xftp/file-server-store.log
Feb 27 19:21:11 localhost xftp-server[2350]: Uploading new files allowed.
Feb 27 19:21:11 localhost xftp-server[2350]: Listening on port 443...
Feb 27 19:21:11 localhost xftp-server[2350]: [INFO 2023-02-27 19:21:11 +0000 src/Simplex/FileTransfer/Server/Env.hs:85] Total / available storage: 64424509440 / 64424509440
To stop xftp-server, run:
sudo systemctl stop xftp-server.service
To check tail of xftp-server log, run:
sudo journalctl -fu xftp-server.service
Feb 27 19:21:11 localhost systemd[1]: Started XFTP server systemd service.
Feb 27 19:21:11 localhost xftp-server[2350]: SimpleX XFTP server v0.1.0
Feb 27 19:21:11 localhost xftp-server[2350]: Fingerprint: ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=
Feb 27 19:21:11 localhost xftp-server[2350]: Server address: xftp://ioyYeRyy4SqJkNvb_7nM04MuLasOM4c-acVyVnqw248=@<hostnames>
Feb 27 19:21:11 localhost xftp-server[2350]: Store log: /var/opt/simplex-xftp/file-server-store.log
Feb 27 19:21:11 localhost xftp-server[2350]: Uploading new files allowed.
Feb 27 19:21:11 localhost xftp-server[2350]: Listening on port 443...
Feb 27 19:21:11 localhost xftp-server[2350]: [INFO 2023-02-27 19:21:11 +0000 src/Simplex/FileTransfer/Server/Env.hs:85] Total / available storage: 64424509440 / 64424509440
Monitoring
You can enable xftp-server statistics for Grafana dashboard by setting value on in /etc/opt/simplex-xftp/file-server.ini, under [STORE_LOG] section in log_stats: field.
Logs will be stored as csv file in /var/opt/simplex-xftp/file-server-stats.daily.log. Fields for the csv file are:
Similarly to phone number portability (the ability of the customer to transfer the service to another provider without changing the number), the address portability means the ability of a communication service customer to change the service provider without changing the service address. Many federated networks support SRV records to provide address portability, but allowing service users to set up their own domains for the addresses is not as commonly supported by the available server and client software as for email.
Federated network
Federated network is provided by several entities that agree upon the standards and operate the network collectively. This allows the users to choose their provider, that will hold their account, their messaging history and contacts, and communicate with other providers' servers on behalf of the user. The examples are email, XMPP, Matrix and Mastodon.
The advantage of that design is that there is no single organization that all users depend on, and the standards are more difficult to change, unless it benefits all users. There are several disadvantages: 1) the innovation is slower, 2) each user account still depends on a single organization, and in most cases can't move to another provider without changing their network address – there is no address portability, 3) the security and privacy are inevitably worse than with the centralized networks.
The credential that allows proving something, e.g. the right to access some resource, without identifying the user. This credential can either be generated by a trusted party or by the user themselves and provided together with the request to create the resource. The first approach creates some centralized dependency in most cases. The second approach does not require any trust - this is used in SimpleX network to authorize access to the messaging queues.
In a wide sense, blockchain means a sequence of blocks of data, where each block contains a cryptographic hash of the previous block, thus providing integrity to the whole chain. Blockchains are used in many communication and information storage systems to provide integrity and immutability of the data. For example, BluRay disks use blockchain. SimpleX messaging queues also use blockchain - each message includes the hash of the previous message, to ensure the integrity – if any message is modified it will be detected by the recipient when the next message is received. Blockchains are a subset of Merkle directed acyclic graphs.
In a more narrow sense, particularly in media, blockchain is used to refer specifically to distributed ledger, where each record also includes the hash of the previous record, but the blocks have to be agreed by the participating peers using some consensus protocol.
Also known as Merkle DAG, a data structure based on a general graph structure where node contains the cryptographic hashes of the previous nodes that point to it. Merkle trees are a subset of Merkle DAGs - in this case each leaf contains a cryptographic hash of the parent.
This structure by design allows to verify the integrity of the whole structure by computing its hashes and comparing with the hashes included in the nodes, in the same way as with blockchain.
The motivation to use DAG in distributed environments instead of a simpler linear blockchain is to allow concurrent additions, when there is no requirement for a single order of added items. Merkle DAG is used, for example, in IPFS and will be used in decentralized SimpleX groups.
Also known as break-in recovery, it is the quality of the end-to-end encryption scheme allowing to recover security against a passive attacker who observes encrypted messages after compromising one (or both) of the parties. Also known as recovery from compromise or break-in recovery. Double-ratchet algorithm has this quality.
Double Ratchet algorithm provides perfect forward secrecy and post-compromise security. It is designed by Signal, and used in SimpleX Chat and many other secure messengers. Most experts consider it the state-of-the-art encryption protocol in message encryption.
Centralized network
Centralized networks are provided or controlled by a single entity. The examples are Threema, Signal, WhatsApp and Telegram. The advantage of that design is that the provider can innovate faster, and has a centralized approach to security. But the disadvantage is that the provider can change or discontinue the service, and leak, sell or disclose in some other way all users' data, including who they are connected with.
Content padding
Also known as content padding, it is the process of adding data to the beginning or the end of a message prior to encryption. Padding conceals the actual message size from any eavesdroppers. SimpleX has several encryption layers, and prior to each encryption the content is padded to a fixed size.
Decentralized network is often used to mean "the network based on decentralized blockchain". In its original meaning, decentralized network means that there is no central authority or any other point of centralization in the network, other than network protocols specification. The advantage of decentralized networks is that they are resilient to censorship and to the provider going out of business. The disadvantage is that they are often slower to innovate, and the security may be worse than with the centralized network.
The examples of decentralized networks are email, web, DNS, XMPP, Matrix, BitTorrent, etc. All these examples have a shared global application-level address space. Cryptocurrency blockchains not only have a shared address space, but also a shared state, so they are more centralized than email. Tor network also has a shared global address space, but also a central authority. SimpleX network does not have a shared application-level address space (it relies on the shared transport-level addresses - SMP relay hostnames or IP addresses), and it does not have any central authority or any shared state.
Defense in depth
Originally, it is a military strategy that seeks to delay rather than prevent the advance of an attacker, buying time and causing additional casualties by yielding space.
In information security, defense in depth represents the use of multiple computer security techniques to help mitigate the risk of one component of the defense being compromised or circumvented. An example could be anti-virus software installed on individual workstations when there is already virus protection on the firewalls and servers within the same environment.
SimpleX network applies defense in depth approach to security by having multiple layers for the communication security and privacy:
additional layer of end-to-end encryption for each messaging queue and another encryption layer of encryption from the server to the recipient inside TLS to prevent correlation by ciphertext,
TLS with only strong ciphers allowed,
mitigation of man-in-the-middle attack on client-server connection via server offline certificate verification,
mitigation of replay attacks via signing over transport channel binding,
multiple layers of message padding to reduce efficiency of traffic analysis,
mitigation of man-in-the-middle attack on client-client out-of-band channel when sending the invitation,
rotation of delivery queues to reduce efficiency of traffic analysis,
A communication system where only the communicating parties can read the messages. It is designed to protect message content from any potential eavesdroppers – telecom and Internet providers, malicious actors, and also the provider of the communication service.
End-to-end encryption requires agreeing cryptographic keys between the sender and the recipient in a way that no eavesdroppers can access the agreed keys. See key agreement protocol. This key exchange can be compromised via man-in-the-middle attack, particularly if key exchange happens via the same communication provider and no out-of-band channel is used to verify key exchange.
Also known as perfect forward secrecy, it is a feature of a key agreement protocol that ensures that session keys will not be compromised even if long-term secrets used in the session key exchange are compromised. Forward secrecy protects past sessions against future compromises of session or long-term keys.
Also known as break-in recovery, it is the quality of the end-to-end encryption scheme allowing to recover security against a passive attacker who observes encrypted messages after compromising one (or both) of the parties. Also known as recovery from compromise or break-in recovery. Double-ratchet algorithm has this quality.
Man-in-the-middle attack
The attack when the attacker secretly relays and possibly alters the communications between two parties who believe that they are directly communicating with each other.
This attack can be used to compromise end-to-end encryption by intercepting public keys during key exchange, substituting them with the attacker's keys, and then intercepting and re-encrypting all messages, without altering their content. With this attack, while the attacker does not change message content, but she can read the messages, while the communicating parties believe the messages are end-to-end encrypted.
Such attack is possible with any system that uses the same channel for key exchange as used to send messages - it includes almost all communication systems except SimpleX, where the initial public key is always passed out-of-band. Even with SimpleX, the attacker may intercept and substitute the key sent via another channel, gaining access to communication. This risk is substantially lower, as attacker does not know in advance which channel will be used to pass the key.
To mitigate such attack the communicating parties must verify the integrity of key exchange - SimpleX and many other messaging apps, e.g. Signal and WhatsApp, have the feature that allows it.
Also known as content padding, it is the process of adding data to the beginning or the end of a message prior to encryption. Padding conceals the actual message size from any eavesdroppers. SimpleX has several encryption layers, and prior to each encryption the content is padded to a fixed size.
Also known as key exchange, it is a process of agreeing cryptographic keys between the sender and the recipient(s) of the message. It is required for end-to-end encryption to work.
Also known as key exchange, it is a process of agreeing cryptographic keys between the sender and the recipient(s) of the message. It is required for end-to-end encryption to work.
The attack when the attacker secretly relays and possibly alters the communications between two parties who believe that they are directly communicating with each other.
This attack can be used to compromise end-to-end encryption by intercepting public keys during key exchange, substituting them with the attacker's keys, and then intercepting and re-encrypting all messages, without altering their content. With this attack, while the attacker does not change message content, but she can read the messages, while the communicating parties believe the messages are end-to-end encrypted.
Such attack is possible with any system that uses the same channel for key exchange as used to send messages - it includes almost all communication systems except SimpleX, where the initial public key is always passed out-of-band. Even with SimpleX, the attacker may intercept and substitute the key sent via another channel, gaining access to communication. This risk is substantially lower, as attacker does not know in advance which channel will be used to pass the key.
To mitigate such attack the communicating parties must verify the integrity of key exchange - SimpleX and many other messaging apps, e.g. Signal and WhatsApp, have the feature that allows it.
A technique for anonymous communication over a computer network that uses multiple layers of message encryption, analogous to the layers of an onion. The encrypted data is transmitted through a series of network nodes called "onion routers," each of which "peels" away a single layer, revealing the data's next destination. The sender remains anonymous because each intermediary knows only the location of the immediately preceding and following nodes.
Some elements of SimpleX network use similar ideas in their design - different addresses for the same resource used by different parties, and additional encryption layers. Currently though, SimpleX messaging protocol does not protect sender network address, as the relay server is chosen by the recipient. The delivery relays chosen by sender that are planned for the future would make SimpleX design closer to onion routing.
Nodes in the overlay network can be thought of as being connected by virtual or logical links, each of which corresponds to a path, perhaps through many physical links, in the underlying network. Tor, for example, is an overlay network on top of IP network, which in its turn is also an overlay network over some underlying physical network.
SimpleX Clients also form a network using SMP relays and IP or some other overlay network (e.g., Tor), to communicate with each other. SMP relays, on another hand, do not form a network.
The property of the cryptographic or communication system that allows the recipient of the message to prove to any third party that the sender identified by some cryptographic key sent the message. It is the opposite to repudiation. While in some context non-repudiation may be desirable (e.g., for contractually binding messages), in the context of private communications it may be undesirable.
The property of the cryptographic or communication system that allows the sender of the message to plausibly deny having sent the message, because while the recipient can verify that the message was sent by the sender, they cannot prove it to any third party - the recipient has a technical ability to forge the same encrypted message. This is an important quality of private communications, as it allows to have the conversation that can later be denied, similarly to having a private face-to-face conversation.
Generalizing the definition from NIST Digital Identity Guidelines, it is an opaque unguessable identifier generated by a service used to access a resource by only one party.
In the context of SimpleX network, these are the identifiers generated by SMP relays to access anonymous messaging queues, with a separate identifier (and access credential) for each accessing party: recipient, sender and and optional notifications subscriber. The same approach is used by XFTP relays to access file chunks, with separate identifiers (and access credentials) for sender and each recipient.
Peer-to-peer
Peer-to-peer (P2P) is the network architecture when participants have equal rights and communicate directly via a general purpose transport or overlay network. Unlike client-server architecture, all peers in a P2P network both provide and consume the resources. In the context of messaging, P2P architecture usually means that the messages are sent between peers, without user accounts or messages being stored on any servers. Examples are Tox, Briar, Cwtch and many others.
The advantage is that the participants do not depend on any servers. There are multiple downsides to that architecture, such as no asynchronous message delivery, the need for network-wide peer addresses, possibility of network-wide attacks, that are usually mitigated only by using a centralized authority. These disadvantages are avoided with proxied P2P architecture.
Network topology of the communication system when peers communicate via proxies that do not form the network themselves. Such design is used in Pond, that has a fixed home server for each user, and in SimpleX, that uses multiple relays providing temporary connections.
Perfect forward secrecy
Also known as perfect forward secrecy, it is a feature of a key agreement protocol that ensures that session keys will not be compromised even if long-term secrets used in the session key exchange are compromised. Forward secrecy protects past sessions against future compromises of session or long-term keys.
Any of the proposed cryptographic systems or algorithms that are thought to be secure against an attack by a quantum computer. It appears that as of 2023 there is no system or algorithm that is proven to be secure against such attacks, or even to be secure against attacks by massively parallel conventional computers, so a general recommendation is to use post-quantum cryptographic systems in combination with the traditional cryptographic systems.
Also known as break-in recovery, it is the quality of the end-to-end encryption scheme allowing to recover security against a passive attacker who observes encrypted messages after compromising one (or both) of the parties. Also known as recovery from compromise or break-in recovery. Double-ratchet algorithm has this quality.
User identity
In a communication system it refers to anything that uniquely identifies the users to the network. Depending on the communication network, it can be a phone number, email address, username, public key or a random opaque identifier. Most messaging networks rely on some form of user identity. SimpleX appears to be the only messaging network that does not rely on any kind of user identity - see this comparison.