tlock gives you time based encryption and decryption capabilities by relying on a drand threshold network as described in our tlock paper.
It is implemented here as a Go library, which is used to implement the tle
command line tool enabling anybody to leverage timelock encryption. A compatible Typescript implementation can be found in tlock-js and a third-party implementation in Rust is available at tlock-rs.
Our timelock encryption system relies on an "unchained drand network".
Working endpoints to access it are, on mainnet:
- https://api.drand.sh/ (US)
- https://api2.drand.sh/ (EU)
- https://api3.drand.sh/ (Asia)
- https://drand.cloudflare.com/ (load-balanced across regions)
On mainnet, the only chainhash supporting timelock encryption, with a 3s frequency and signatures on the G1 group is:
dbd506d6ef76e5f386f41c651dcb808c5bcbd75471cc4eafa3f4df7ad4e4c493
This is a production-ready network with high-availability guarantees. It is considered fully secure by the drand team and ran by the same League of Entropy that has been running drand in production since 2019.
On testnet:
- https://pl-us.testnet.drand.sh/
- https://pl-eu.testnet.drand.sh/
- https://testnet0-api.drand.cloudflare.com/ where we have two networks supporting timelock:
- running with a 3 seconds frequency with signatures on G1:
f3827d772c155f95a9fda8901ddd59591a082df5ac6efe3a479ddb1f5eeb202c
- running with a 3 seconds frequency with signatures on G2:
7672797f548f3f4748ac4bf3352fc6c6b6468c9ad40ad456a397545c6e2df5bf
Note these are relying on the League of Entropy Testnet, which should not be considered secure.
You can also spin up a new drand network and run your own, but note that the security guarantees boil down to the trust you have in your network.
For more details about the scheme and maths, please refer to our paper on eprint.
- [Install the
tle
tool](#install-or build-the-cli) - CLI usage
- Library usage
- Applying another layer of encryption
- Security considerations
- Get in touch
The tle
tool is pure Go, it works without CGO (CGO_ENABLED=0
).
To install using Go:
go install github.com/drand/tlock/cmd/tle@latest
or locally with git:
git clone https://github.com/drand/tlock
go build cmd/tle/tle.go
Note: if you need to decrypt old ciphertexts produced before v1.0.0 against our testnet, you'll need to install the old binary using:
go install github.com/drand/tlock/cmd/tle@v0.1.0
We have changed in v1.0.0 the way we are mapping a hash to field element and this is not retro-compatible with the old ciphertext format. You can still produce ciphertexts against our testnet as explained below using the v1.0.0, but it will use the new hash to field mechanism.
Usage:
tle [--encrypt] (-r round)... [--armor] [-o OUTPUT] [INPUT]
tle --decrypt [-o OUTPUT] [INPUT]
Options:
-e, --encrypt Encrypt the input to the output. Default if omitted.
-d, --decrypt Decrypt the input to the output.
-n, --network The drand API endpoint to use.
-c, --chain The chain to use. Can use either beacon ID name or beacon hash. Use beacon hash in order to ensure public key integrity.
-r, --round The specific round to use to encrypt the message. Cannot be used with --duration.
-f, --force Forces to encrypt against past rounds.
-D, --duration How long to wait before the message can be decrypted.
-o, --output Write the result to the file at path OUTPUT.
-a, --armor Encrypt to a PEM encoded format.
If the OUTPUT exists, it will be overwritten.
NETWORK defaults to the drand mainnet endpoint https://api.drand.sh/.
CHAIN defaults to the chainhash of quicknet:
52db9ba70e0cc0f6eaf7803dd07447a1f5477735fd3f661792ba94600c84e971
it used to be that of the fastnet network, but that one is being deprecated:
dbd506d6ef76e5f386f41c651dcb808c5bcbd75471cc4eafa3f4df7ad4e4c493
You can also use the drand test network:
https://pl-us.testnet.drand.sh/
and its unchained network on G2 with chainhash 7672797f548f3f4748ac4bf3352fc6c6b6468c9ad40ad456a397545c6e2df5bf
Note that if you encrypted something prior to March 2023, this was the only available network and used to be the default.
DURATION, when specified, expects a number followed by one of these units:
"ns", "us" (or "µs"), "ms", "s", "m", "h", "d", "M", "y".
Example:
$ tle -D 10d -o encrypted_file data_to_encrypt
After the specified duration:
$ tle -d -o decrypted_file.txt encrypted_file
Files can be encrypted using a duration (--duration/-D
) in which the encrypted_data
can be decrypted.
Example using the testnet network and a duration of 5 seconds:
$ tle -n="https://pl-us.testnet.drand.sh/" -c="7672797f548f3f4748ac4bf3352fc6c6b6468c9ad40ad456a397545c6e2df5bf" -D=5s -o=encrypted_data data.txt
If a round (--round/-R
) number is known, it can be used instead of the duration. The data can be decrypted only when that round becomes available in the network.
Example using the fastnet mainnet network and a given round:
$ tle -n="https://api.drand.sh/" -c="dbd506d6ef76e5f386f41c651dcb808c5bcbd75471cc4eafa3f4df7ad4e4c493" -r=123456 -o=encrypted_data data.txt
It is also possible to encrypt the data to a PEM encoded format using the armor (--armor/-a
) flag,
and to rely on the default network and chain hash (which is the fastnet
one on api.drand.sh
):
$ tle -a -D 20s -o=encrypted_data.PEM data.txt
For decryption, it's only necessary to specify the network if you're not using the default one.
Using the default ("fastnet" network on mainnet) and printing on stdout:
$ tle -d encrypted_data
Using the old testnet unchained network and storing the output in a file named "decrypted_data":
$ tle -d -n="https://pl-us.testnet.drand.sh/" -c="7672797f548f3f4748ac4bf3352fc6c6b6468c9ad40ad456a397545c6e2df5bf"
-o=decrypted_data encrypted_data
Note it will overwrite the decrypted_data
file if it already exists.
If decoding an armored source you don't need to specify -a
again.
These example show how to use the API to time lock encrypt and decrypt data.
// Open an io.Reader to the data to be encrypted.
in, err := os.Open("data.txt")
if err != nil {
log.Fatalf("open: %s", err)
return
}
defer in.Close()
// Construct a network that can talk to a drand network. Example using the mainnet fastnet network.
// host: "https://api.drand.sh/"
// chainHash: "dbd506d6ef76e5f386f41c651dcb808c5bcbd75471cc4eafa3f4df7ad4e4c493"
network,err := http.NewNetwork(host, chainHash)
if err != nil {
log.Fatalf("new network: %s", err)
return
}
// Specify how long we need to wait before the file can be decrypted.
duration := 10 * time.Second
// Use the network to identify the round number that represents the duration.
roundNumber := network.RoundNumber(time.Now().Add(duration))
// Write the encrypted file data to this buffer.
var cipherData bytes.Buffer
// Encrypt the data for the given round.
if err := tlock.New(network).Encrypt(&cipherData, in, roundNumber); err != nil {
log.Fatalf("encrypt: %v", err)
return
}
// Open an io.Reader to the data to be decrypted.
in, err := os.Open("data.tle")
if err != nil {
log.Fatalf("open: %v", err)
return
}
defer in.Close()
// Construct a network that can talk to a drand network.
// host: "https://api.drand.sh/"
// chainHash: "dbd506d6ef76e5f386f41c651dcb808c5bcbd75471cc4eafa3f4df7ad4e4c493"
network,err := http.NewNetwork(host, chainHash)
if err != nil {
log.Fatalf("new network: %s", err)
return
}
// Write the decrypted file data to this buffer.
var plainData bytes.Buffer
// Decrypt the data. If you try to decrypt the data *before* the specified
// duration, it will fail with the message: "too early to decrypt".
if err := tlock.New(network).Decrypt(&plainData, in); err != nil {
log.Fatalf("decrypt: %v", err)
return
}
The recommended way of doing "hybrid" encryption where you both encrypt your data using timelock encryption, but also with another encryption scheme, such as a public-key or a symmetric-key scheme is to simple re-encrypt your encrypted data using tlock.
For example, you can use the age cli to encrypt your data with a passphrase as follows.
$ cat data.txt | age -p | tle -D 30s -o encrypted_data
$ cat encrypted_data | tle -d | age -d -o data.txt
Note that you could do the same with PGP or any other encryption tool.
The security of our timelock encryption mechanism relies on four main things:
- The security of the underlying Identity Encryption Scheme (proposed in 2001) and its implementation that we're using.
- The security of the threshold BLS scheme (proposed in 2003), and its impementation by the network you're relying on.
- The security of age's underlying primitives, and that of the age implementation we're using to encrypt the data, since we rely on the hybrid encryption principle, where we only timelock encrypt ("wrap") a random symmetric key that is used by age to actually symmetrically encrypt the data using Chacha20Poly1305).
- The security of the threshold network providing you with its BLS signatures at a given frequency, for instance the default for
tle
is to rely on drand and its existing League of Entropy network.
In practice this means that if you trust there are never more than the threshold t
malicious nodes on the network you're relying on, you are guaranteed that you timelocked data cannot be decrypted earlier than what you intended.
Please note that neither BLS nor the IBE scheme we are relying on are "quantum resistant", therefore shall a Quantum Computer be built that's able to threaten their security, our current design wouldn't resist. There are also no quantum resistant scheme that we're aware of that could be used to replace our current design since post-quantum signatures schemes do not "thresholdize" too well in a post-quantum IBE-compatible way.
However, such a quantum computer seems unlikely to be built within the next 5-10 years and therefore we currently consider that you can expect a "long term security" horizon of at least 5 years by relying on our design.
Finally, relying on the League of Entropy Testnet should not be considered secure and be used only for testing purposes. We recommend relying on the League of Entropy fastnet
beacon chain running on Mainnet for securing timelocked content.
Our timelock scheme and code was reviewed by cryptography and security experts from Kudelski and the report is available on IPFS at QmWQvTdiD3fSwJgasPLppHZKP6SMvsuTUnb1vRP2xM7y4m
.
- Open an issue for feature requests or to report a bug.
- Join the drand Slack to discuss Timelock, randomness beacons and more.
- Follow the drand blog for our articles.
- Follow the @drand_loe account on Twitter to stay tuned.
This project is licensed using the Permissive License Stack which means that all contributions are available under the most permissive commonly-used licenses, and dependent projects can pick the license that best suits them.
Therefore, the project is dual-licensed under Apache 2.0 and MIT terms:
- Apache License, Version 2.0, (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT)