Architecture of the Hyperledger Blockchain Fabric∗

Christian Cachin

IBM Research - Zurich

CH-8803 Rüschlikon, Switzerland

[email protected]

July 2016

Overview. A blockchain is best understood in the model of state-machine replication [8], where a
service maintains some state and clients invoke operations that transform the state and generate outputs.
A blockchain emulates a “trusted” computing service through a distributed protocol, run by nodes con-
nected over the Internet. The service represents or creates an asset, in which all nodes have some stake.
The nodes share the common goal of running the service but do not necessarily trust each other for
more. In a “permissionless” blockchain such as the one underlying the Bitcoin cryptocurrency, anyone
can operate a node and participate through spending CPU cycles and demonstrating a “proof-of-work.”
On the other hand, blockchains in the “permissioned” model control who participates in validation and
in the protocol; these nodes typically have established identities and form a consortium. A report of
Swanson compares the two models [9].

Hyperledger. The Hyperledger Project (www.hyperledger.org) is a collaborative effort to cre-

ate an enterprise-grade, open-source distributed ledger framework and code base. It aims to advance
blockchain technology by identifying and realizing a cross-industry open standard platform for dis-
tributed ledgers, which can transform the way business transactions are conducted globally. Established
as a project of the Linux Foundation in early 2016, the Hyperledger Project currently has more than 50
members.

Hyperledger Fabric. Hyperledger Fabric (github.com/hyperledger/fabric) is an imple-

mentation of a distributed ledger platform for running smart contracts, leveraging familiar and proven
technologies, with a modular architecture allowing pluggable implementations of various functions.
It is one of multiple projects currently in incubation under the Hyperledger Project. A developer-
preview of the Hyperledger Fabric (called “v0.5-developer-preview”) has been released in June 2016
(github.com/hyperledger/fabric/wiki/Fabric-Releases).
The distributed ledger protocol of the fabric is run by peers. The fabric distinguishes between
two kinds of peers: A validating peer is a node on the network responsible for running consensus,
∗
This text represents the view of the author and is directly based on the current Hyperledger Fabric documentation and
code, available online at github.com/hyperledger/fabric/.
The author is grateful to many collaborators at IBM and in the Hyperledger Project, including Sheehan Anderson, Elli An-
droulaki, Angelo De Caro, Konstantinos Christidis, Jason Yellick, Chet Murthy, Binh Nguyen, Michael Osborne, Simon
Schubert, Gari Singh, Alessandro Sorniotti, Marko Vukolić, Sharon Weed, and John Wolpert.
validating transactions, and maintaining the ledger. On the other hand, a non-validating peer is a node
that functions as a proxy to connect clients (issuing transactions) to validating peers. A non-validating
peer does not execute transactions but it may verify them.
Some key features of the current fabric release are:
• A permissioned blockchain with immediate finality;
• Runs arbitrary smart contracts (called chaincode) implemented in Go (golang.org):
– User-defined chaincode is encapsulated in a Docker container;
– System chaincode runs in the same process as the peer;
• Consensus protocol is pluggable, currently an implementation of Byzantine fault-tolerant con-
sensus using the PBFT protocol [4] is supported, a prototype of SIEVE [3] to address non-
deterministic chaincode is available, and a protocol stub (named NOOPS) serves for development
on a single node;
• Security support through certificate authorities (CAs) for TLS certificates, enrollment certificates,
and transaction certificates;
• Persistent state using a key-value store interface, backed by RocksDB (rocksdb.org);
• An event framework that supports pre-defined and custom events;
• A client SDK (Node.js) to interface with the fabric;
• Support for basic REST APIs and CLIs.
Support for non-validating peers is minimal in the developer preview release.

Architecture. The validating peers run a BFT consensus protocol for executing a replicated state
machine that accepts three types of transactions as operations:
Deploy transaction: Takes a chaincode (representing a smart contract) written in Go as a parameter; the
chaincode is installed on the peers and ready to be invoked.
Invoke transaction: Invokes a transaction of a particular chaincode that has been installed earlier through
a deploy transaction; the arguments are specific to the type of transaction; the chaincode executes the
transaction, may read and write entries in its state accordingly, and indicates whether it succeeded or
failed.
Query transaction: Returns an entry of the state directly from reading the peer’s persistent state; this
may not ensure linearizability.
Each chaincode may define its own persistent entries in the state. The blockchain’s hash chain is
computed over the executed transactions and the resulting persistent state.
Validation of transactions occurs through the replicated execution of the chaincode and given the
fault assumption underlying BFT consensus, i.e., that among the n validating peers at most f < n/3
may “lie” and behave arbitrarily, but all others execute the chaincode correctly. When executed on top
of PBFT consensus, it is important that chaincode transactions are deterministic, otherwise the state
of the peers might diverge. A modular solution to filter out non-deterministic transactions that are
demonstrably diverging is available and has been implemented in the SIEVE protocol [3].
Membership among the validating nodes running BFT consensus is currently static and the setup
requires manual intervention. Support for dynamically changing the set of nodes running consensus is
planned for a future version.
As the fabric implements a permissioned ledger, it contains a security infrastructure for authen-
tication and authorization. It supports enrollment and transaction authorization through public-key
certificates, and confidentiality for chaincode realized through in-band encryption.
 More precisely, for connecting to the network every peer needs to obtain an enrollment certificate
from an enrollment CA that is part of the membership services. It authorizes a peer to connect to the
network and to acquire transaction certificates, which are needed to submit transactions. Transaction
certificates are issued by a transaction CA and support pseudonymous authorization for the peers sub-
mitting transactions, in the sense that multiple transaction certificates issued to the same peer (that is, to
the same enrollment certificate) cannot be linked with each other.
Confidentiality for chaincodes and state is provided through symmetric-key encryption of transac-
tions and states with a blockchain-specific key that is available to all peers with an enrollment certificate
for the blockchain. Extending the encryption mechanisms towards more fine-grained confidentiality for
transactions and state entries is planned for a future version.

Discussion. Consensus protocols for blockchain are currently being debated very actively, in re-
search [6, 10] but also by fintech startup companies (e.g., tendermint.com, kadena.io). The
fabric’s design uses a modular notion of consensus, which is aligned with the well-established concept
of consensus in distributed computing. This ensures that the blockchain-related features of the fabric
can be developed independently of the specific consensus protocol. The PBFT protocol [4] has been
implemented as the first one in the fabric because of its prominence: it benefits from the experience
of almost 20 years of systems-level research on Byzantine consensus, is closely related to well-known
protocols like viewstamped replication and Paxos [7], has been analyzed in many environments [5, 1],
and is described in textbooks [2].

Conclusion. The Hyperledger Fabric is a permissioned blockchain platform aimed at business use. It
is open-source and based on standards, runs user-defined smart contracts, supports strong security and
identity features, and uses a modular architecture with pluggable consensus protocols.
The fabric is currently evolving and being actively developed under the governance of the Hyper-
ledger Project. More information about the fabric is available online at
github.com/hyperledger/fabric/blob/master/docs/protocol-spec.md

References
[1] Y. Amir, B. A. Coan, J. Kirsch, and J. Lane. Prime: Byzantine replication under attack. IEEE
Transactions on Dependable and Secure Computing, 8(4):564–577, 2011.

[2] C. Cachin, R. Guerraoui, and L. Rodrigues. Introduction to Reliable and Secure Distributed Pro-
gramming (Second Edition). Springer, 2011.

[3] C. Cachin, S. Schubert, and M. Vukolić. Non-determinism in Byzantine fault-tolerant replication.

e-print, arXiv:1603.07351 [cs.DC], 2016. URL: http://arxiv.org/abs/1603.07351.

[4] M. Castro and B. Liskov. Practical Byzantine fault tolerance and proactive recovery. ACM Trans-
actions on Computer Systems, 20(4):398–461, Nov. 2002.

[5] A. Clement, E. L. Wong, L. Alvisi, M. Dahlin, and M. Marchetti. Making Byzantine fault tolerant
systems tolerate Byzantine faults. In Proc. 6th Symp. Networked Systems Design and Implemen-
tation (NSDI), pages 153–168, 2009.

[6] J. A. Garay, A. Kiayias, and N. Leonardos. The bitcoin backbone protocol: Analysis and appli-
cations. In Advances in Cryptology: Eurocrypt 2015, volume 9057 of Lecture Notes in Computer
Science, pages 281–310. Springer, 2015.
 [7] B. Liskov. From viewstamped replication to Byzantine fault tolerance. In B. Charron-Bost, F. Pe-
done, and A. Schiper, editors, Replication: Theory and Practice, volume 5959 of Lecture Notes in
Computer Science, pages 121–149. Springer, 2010.

[8] F. B. Schneider. Implementing fault-tolerant services using the state machine approach: A tutorial.
ACM Computing Surveys, 22(4):299–319, Dec. 1990.

[9] T. Swanson. Consensus-as-a-service: A brief report on the emergence of permissioned, distributed

ledger systems. Report, available online, Apr. 2015. URL: http://www.ofnumbers.com/
wp-content/uploads/2015/04/Permissioned-distributed-ledgers.pdf.

[10] M. Vukolić. The quest for scalable blockchain fabric: Proof-of-work vs. BFT replication. In Open
Problems in Network Security, Proc. IFIP WG 11.4 Workshop (iNetSec 2015), volume 9591 of
Lecture Notes in Computer Science, pages 112–125. Springer, 2016.