Drand (pronounced "dee-rand") is a distributed randomness beacon daemon written in Golang.
Linked drand nodes collectively produce publicly verifiable, unbiased and unpredictable random values at fixed intervals using bilinear pairings and threshold cryptography.
Drand was first developed within the DEDIS organization, and as of December 2019, is now under the drand organization.
The need for digital randomness is paramount in multiple digital applications ([e]voting, lottery, cryptographic parameters, embedded devices bootstrapping randomness, blockchain systems etc) as well in non-digital such as statistical sampling (used for example to check results of an election), assigning court cases to random judges, random financial audits, etc. However, constructing a secure source of randomness is anything but easy: there are countless examples of attacks where the randomness generation was the culprit (static keys, non-uniform distribution, biased output, etc). drand aims to fix that gap by providing a Randomness-as-a-Service network (similar to NTP servers for time, or Certificate Authority servers for CAs verification), providing continuous source of randomness which is:
- Decentralized: drand is a software ran by a diverse set of reputable entities on the Internet and a threshold of them is needed to generate randomness, there is no central point of failure.
- Publicly verifiable & unbiased: drand periodically delivers publicly verifiable and unbiased randomness. Any third party can fetch and verify the authenticity of the randomness and by that making sure it hasn't been tampered with.
- And "private" as well: drand nodes can also deliver encrypted randomness to be used in a local applications, for example to seed the OS's PRNG.
A drand network is operated by a group of organizations around the world that includes Cloudflare, EPFL, Kudelski Security, Protocol Labs, Celo, UCL, and UIUC. You can learn more by visiting the League of Entropy website, where you can also see the random values being generated by the network in real time.
Generating public randomness is the primary functionality of drand. Public randomness is generated collectively by drand nodes and publicly available. The main challenge in generating good randomness is that no party involved in the randomness generation process should be able to predict or bias the final output. Additionally, the final result has to be third-party verifiable to make it actually useful for applications like lotteries, sharding, or parameter generation in security protocols.
A drand randomness beacon is composed of a distributed set of nodes and has two phases:
- Setup: Each node first generates a long-term public/private key pair. Then all of the public keys are written to a group file together with some further metadata required to operate the beacon. After this group file has been distributed, the nodes perform a distributed key generation (DKG) protocol to create the collective public key and one private key share per server. The participants NEVER see/use the actual (distributed) private key explicitly but instead utilize their respective private key shares for the generation of public randomness.
- Generation: After the setup, the nodes switch to the randomness generation mode. Any of the nodes can initiate a randomness generation round by broadcasting a message which all the other participants sign using a t-of-n threshold version of the Boneh-Lynn-Shacham (BLS) signature scheme and their respective private key shares. Once any node (or third-party observer) has gathered t partial signatures, it can reconstruct the full BLS signature (using Lagrange interpolation). The signature is then hashed using SHA-256 to ensure that there is no bias in the byte representation of the final output. This hash corresponds to the collective random value and can be verified against the collective public key.
Private randomness generation is the secondary functionality of drand. Clients can request private randomness from some or all of the drand nodes which extract it locally from their entropy pools and send it back in encrypted form. This can be useful to gather randomness from different entropy sources, for example in embedded devices.
In this mode we assume that a client has a private/public key pair and
encapsulates its public key towards the server's public key using the ECIES
encryption scheme. After receiving a request, the drand node produces 32 random
bytes locally (using Go's crypto/rand
interface), encrypts them using the
received public key and sends it back to the client.
Note: Assuming that clients without good local entropy sources (such as embedded devices) use this process to gather high entropy randomness to bootstrap their local PRNGs, we emphasize that the initial client key pair has to be provided by a trusted source (such as the device manufacturer). Otherwise we run into the chicken-and-egg problem of how to produce on the client's side a secure ephemeral key pair for ECIES encryption without a good (local) source of randomness.
Please go use the latest drand binary in the release page.
Drand can be installed via Golang or
Docker. By default, drand saves the configuration
files such as the long-term key pair, the group file, and the collective public
key in the directory $HOME/.drand/
.
Make sure that you have a working Golang installation and that your GOPATH is set.
Then install drand via:
git clone https://github.com/drand/drand
cd drand
make install
The setup is explained in docker/README.md.
To run a local demo, you can simply run:
make demo
The script spins up a few drand local processes, performs resharing and other operations and will continue to print out new randomness every Xs (currently 6s). For more information, look at the demo README.
A drand beacon provides several public services to clients. A drand node
exposes its public services on a gRPC endpoint as well as a REST JSON endpoint,
on the same port. The latter is especially useful if one wishes to retrieve
randomness from a JavaScript application. Communication is protected through
TLS by default. If the contacted node is using a self-signed certificate, the
client can use the --tls-cert
flag to specify the server's certificate.
Consult full instructions at DEPLOYMENT
To get the latest public random value, run
drand get public --round <i> <group.toml>
where <group.toml>
is the group identity file of a drand node. You can
specify the round number when the public randomness has been generated. If not
specified, this command returns the most recent random beacon.
The JSON-formatted output produced by drand is of the following form:
{
"round": 367,
"signature": "b62dd642e939191af1f9e15bef0f0b0e9562a5f570a12a231864afe468377e2a6424a92ccfc34ef1471cbd58c37c6b020cf75ce9446d2aa1252a090250b2b1441f8a2a0d22208dcc09332eaa0143c4a508be13de63978dbed273e3b9813130d5",
"previous_signature": "afc545efb57f591dbdf833c339b3369f569566a93e49578db46b6586299422483b7a2d595814046e2847494b401650a0050981e716e531b6f4b620909c2bf1476fd82cf788a110becbc77e55746a7cccd47fb171e8ae2eea2a22fcc6a512486d",
"randomness": "d7aed3686bf2be657e6d38c20999831308ee6244b68c8825676db580e7e3bec6"
}
Here Signature
is the threshold BLS signature on the previous signature value
Previous
and the current round number. Randomness
is the hash of
Signature
, to be used as the random value for this round. The field Round
specifies the index of Randomness
in the sequence of all random values
produced by this drand instance. The message signed is therefore the
concatenation of the round number treated as a uint64
and the previous
signature. At the moment, we are only using BLS signatures on the bls12-381 curves
and the signature is made over G1.
(Note that this command expects access to a drand group member, this won't work with the current League of Entropy nodes, since they are not exposing their GRPC endpoints directly.)
To get a private random value, run the following:
drand get private group.toml
The JSON-formatted output produced by drand should look like the following:
{
"Randomness": "764f6e3eecdc4aba8b2f0119e7b2fd8c35948bf2be3f87ebb5823150c6065764"
}
The command outputs a 32-byte hex-encoded random value generated from the local randomness engine of the contacted server. If the encryption is not correct, the command outputs an error instead.
(Note that this is also not available with the League of Entropy nodes, we recommend using this only with your own nodes, since private randomness is usually used for private keys or secrets and can be very sensitive.)
This is the recommended way of using drand randomness, but don't forget to validate the beacons' signatures against the group public key.
One may want to get the distributed key or public randomness by issuing a GET to a HTTP endpoint instead of using a gRPC client. Here is a basic example on how to do so with curl.
To get the distributed key, you can use:
curl <address>/group
Similarly, to get the latest round of randomness from the drand beacon, you can use
curl <address>/public/latest
To facilitate the use of drand's randomness in JavaScript-based applications,
we provide drand-client
.
For more details on the procedure and instructions on how to use it, refer to the readme.
Here is a list of all documentation related to drand:
- To learn more about the protocol, the motivation and its backgronund
- For a high level presentation of motivations and background, here are some public slides about drand or online video.
- A basic explainer of the cryptography behind drand,
- Protocol Specification
- API documentation
- DevOps/Deployment documentation
As well, here is a list of background readings w.r.t to the cryptography used in drand:
- Pairing-based cryptography and Barreto-Naehrig curves.
- Pedersen's distributed key generation protocol for the setup.
- Threshold BLS signatures for the generation of public randomness.
- The resharing scheme used comes from the paper from Y. Desmedt and S. Jajodia.
- ECIES for the encryption of private randomness.
Note that drand was originally a DEDIS-owned project that is now spinning off on its own Github organization. For related previous work on public randomness, see DEDIS's academic paper Scalable Bias-Resistant Distributed Randomness.
Although being already functional, drand is still at an early development stage and there is a lot left to be done. The list of opened issues is a good place to start. On top of this, drand would benefit from higher-level enhancements such as the following:
- Implement a more failure-resilient DKG protocol or an approach based on verifiable succinct computations (zk-SNARKs, etc).
- Use / implement a faster pairing based library in JavaScript
- Add more unit tests
- Add a systemd unit file
Feel free to submit feature requests or, even better, pull requests ;)
Thanks to @herumi for providing support on his optimized pairing-based cryptographic library used in the first version.
Thanks to Apostol Vassilev for its interest in drand and the extensive and helpful discussions on the drand design.
Thanks to @Bren2010 and @grittygrease for providing the native Golang bn256 implementation and for their help in the design of drand and future ideas.
Finally, a special note for Bryan Ford from the DEDIS lab for letting me work on this project and helping me grow it.
- EPFL blog post
- Cloudflare crypto week introduction post and the more technical post.
- Kudelski Security blog post
- OneZero post on the league of entropy
- SlashDot post
- Duo post
- Liftr
- (French) nextimpact
The drand project is dual-licensed under Apache 2.0 and MIT terms:
- Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)