A recent approach to building consensus protocols on top of Directed Acyclic Graphs (DAGs) shows much promise due to its simplicity and stable throughput. However, as each node in the DAG typically includes a linear number of references to the nodes in the previous round, prior DAG protocols only scale up to a certain point when the overhead of maintaining the graph becomes the bottleneck. To enable large-scale deployments of DAG-based protocols, we propose a sparse DAG architecture, where each node includes only a constant number of references to random nodes in the previous round. We present a sparse version of Bullshark - one of the most prominent DAG-based consensus protocols - and demonstrate its improved scalability. Remarkably, unlike other protocols that use random sampling to reduce communication complexity, we manage to avoid sacrificing resilience: the protocol can tolerate up to f < n/3 Byzantine faults (where n is the number of participants), same as its less scalable deterministic counterpart. The proposed "sparse" methodology can be applied to any protocol that maintains disseminated system updates and causal relations between them in a graph-like structure. Our simulations show that the considerable reduction of transmitted metadata in sparse DAGs results in more efficient network utilization and better scalability.

The majority of fault-tolerant distributed algorithms are designed assuming a nominal corruption model, in which at most a fraction fn of parties can be corrupted by the adversary. However, due to the infamous Sybil attack, nominal models are not sufficient to express the trust assumptions in open (i.e., permissionless) settings. Instead, permissionless systems typically operate in a weighted model, where each participant is associated with a weight and the adversary can corrupt a set of parties holding at most a fraction fw of the total weight.

A typical blockchain protocol uses consensus to make sure that mutually mistrusting users agree on the order in which their operations on shared data are executed. However, it is known that asset transfer systems, by far the most popular application of blockchains, can be implemented without consensus. Assuming that no account can be accessed concurrently and every account belongs to a single owner, one can efficiently implement an asset transfer system in a purely asynchronous, consensus-free manner. It has also been shown that implementing asset transfer with shared accounts is impossible without consensus.

International audience

Randomisation is a critical tool in designing distributed systems. The common coin primitive, enabling the system members to agree on an unpredictable random number, has proven to be particularly useful. We observe, however, that it is impossible to implement a truly random common coin protocol in a fault-prone asynchronous system. To circumvent this impossibility, we introduce two relaxations of the perfect common coin: (1) approximate common coin generating random numbers that are close to each other; and (2) Monte Carlo common coin generating a common random number with an arbitrarily small, but non-zero, probability of failure. Building atop the approximate agreement primitive, we obtain efficient asynchronous implementations of the two abstractions, tolerating up to one third of Byzantine processes. Our protocols do not assume trusted setup or public key infrastructure and converge to the perfect coin exponentially fast in the protocol running time. By plugging one of our protocols for Monte Carlo common coin in a well-known consensus algorithm, we manage to get a binary Byzantine agreement protocol with O(n³ log n) communication complexity, resilient against an adaptive adversary, and tolerating the optimal number f < n/3 of failures without trusted setup or PKI. To the best of our knowledge, the best communication complexity for binary Byzantine agreement achieved so far in this setting is O(n⁴). We also show how the approximate common coin, combined with a variant of Gray code, can be used to solve an interesting problem of Intersecting Random Subsets, which we introduce in this paper.

Replicated services are inherently vulnerable to failures and security breaches. In a long-running system, it is, therefore, indispensable to maintain a reconfiguration mechanism that would replace faulty replicas with correct ones. An important challenge is to enable reconfiguration without affecting the availability and consistency of the replicated data: the clients should be able to get correct service even when the set of service replicas is being updated. In this paper, we address the problem of reconfiguration in the presence of Byzantine failures: faulty replicas or clients may arbitrarily deviate from their expected behavior. We describe a generic technique for building asynchronous and Byzantine fault-tolerant reconfigurable objects: clients can manipulate the object data and issue reconfiguration calls without reaching consensus on the current configuration. With the help of forward-secure digital signatures, our solution makes sure that superseded and possibly compromised configurations are harmless, that slow clients cannot be fooled into reading stale data, and that Byzantine clients cannot cause a denial of service by flooding the system with reconfiguration requests. Our approach is modular and based on dynamic Byzantine lattice agreement abstraction, and we discuss how to extend it to enable Byzantine fault-tolerant implementations of a large class of reconfigurable replicated services.

International audience

It is a common belief that Byzantine fault-tolerant solutions for consensus are significantly slower than their crash fault-tolerant counterparts. Indeed, in PBFT, the most widely known Byzantine fault-tolerant consensus protocol, it takes three message delays to decide a value, in contrast with just two in Paxos. This motivates the search for fast Byzantine consensus algorithms that can produce decisions after just two message delays in the common case, e.g., under the assumption that the current leader is correct and not suspected by correct processes. The (optimal) two-step latency comes with the cost of lower resilience: fast Byzantine consensus requires more processes to tolerate the same number of faults. In particular, 5f+1 processes were claimed to be necessary to tolerate f Byzantine failures. In this paper, we present a fast Byzantine consensus algorithm that relies on just 5f-1 processes. Moreover, we show that 5f-1 is the tight lower bound, correcting a mistake in the earlier work. While the difference of just 2 processes may appear insignificant for large values of f, it can be crucial for systems of a smaller scale. In particular, for f=1, our algorithm requires only 4 processes, which is optimal for any (not necessarily fast) partially synchronous Byzantine consensus algorithm.

Most modern asset transfer systems use consensus to maintain a totally ordered chain of transactions. It was recently shown that consensus is not always necessary for implementing asset transfer. More efficient, asynchronous solutions can be built using reliable broadcast instead of consensus. This approach has been originally used in the closed (permissioned) setting. In this paper, we extend it to the open (permissionless) environment. We present Pastro, a permissionless and asynchronous asset-transfer implementation, in which quorum systems, traditionally used in reliable broadcast, are replaced with a weighted Proof-of-Stake mechanism. Pastro tolerates a dynamic adversary that is able to adaptively corrupt participants based on the assets owned by them.

We address the problem of online payments, where users can transfer funds among themselves. We introduce Astro, a system solving this problem efficiently in a decentralized, deterministic, and completely asynchronous manner. Astro builds on the insight that consensus is unnecessary to prevent double-spending. Instead of consensus, Astro relies on a weaker primitive---Byzantine reliable broadcast---enabling a simpler and more efficient implementation than consensus-based payment systems. In terms of efficiency, Astro executes a payment by merely broadcasting a message. The distinguishing feature of Astro is that it can maintain performance robustly, i.e., remain unaffected by a fraction of replicas being compromised or slowed down by an adversary. Our experiments on a public cloud network show that Astro can achieve near-linear scalability in a sharded setup, going from 10K payments/sec (2 shards) to 20K payments/sec (4 shards). In a nutshell, Astro can match VISA-level average payment throughput, and achieves a 5× improvement over a state-of-the-art consensus-based solution, while exhibiting sub-second 95^th percentile latency.

Reliable broadcast is a communication primitive guaranteeing, intuitively, that all processes in a distributed system deliver the same set of messages. The reason why this primitive is appealing is twofold: (i) we can implement it deterministically in a completely asynchronous environment, unlike stronger primitives like consensus and total-order broadcast, and yet (ii) reliable broadcast is powerful enough to implement important applications like payment systems. The problem we tackle in this paper is that of dynamic reliable broadcast, i.e., enabling processes to join or leave the system. This property is desirable for long-lived applications (aiming to be highly available), yet has been precluded in previous asynchronous reliable broadcast protocols. We study this property in a general adversarial (i.e., Byzantine) environment. We introduce the first specification of a dynamic Byzantine reliable broadcast (DBRB) primitive that is amenable to an asynchronous implementation. We then present an algorithm implementing this specification in an asynchronous network. Our DBRB algorithm ensures that if any correct process in the system broadcasts a message, then every correct process delivers that message unless it leaves the system. Moreover, if a correct process delivers a message, then every correct process that has not expressed its will to leave the system delivers that message. We assume that more than $2/3$ of processes in the system are correct at all times, which is tight in our context. We also show that if only one process in the system can fail---and it can fail only by crashing---then it is impossible to implement a stronger primitive, ensuring that if any correct process in the system broadcasts or delivers a message, then every correct process in the system delivers that message---including those that leave.