Cheap and Available State Machine Replication
نویسندگان
چکیده
This paper presents that, by combining on-demand instantiation and lazy recovery, we can reduce the cost of asynchronous state machine replication protocols, such as Paxos and UpRight, while maintaining their high availability. To reduce cost, we incorporate on-demand instantiation, which activates a subset of replicas first and activates backup ones when active ones fail. To solve its key limitation—the system can be halted for long when activating a backup replica, we apply lazy recovery, allowing the system to proceed while recovering backup nodes in the background. The key contribution of this paper is to identify that, when agreement nodes and execution nodes are logically separated, they each presents a unique property that enables lazy recovery. We have applied this idea to Paxos and built ThriftyPaxos, which, as shown in the evaluation, can achieve higher throughput and similar availability comparing to standard Paxos, despite the fact that ThriftyPaxos activates fewer replicas.
منابع مشابه
Generalized Universality
State machine replication reduces distributed to centralized computing. Any sequential service, modeled by a state machine, can be replicated over any number of processes and made highly available to all of them. At the heart of this fundamental reduction lies the so called universal consensus abstraction, key to providing the illusion of single shared service, despite replication. Yet, as univ...
متن کاملDemocratizing the Parliament (extended Abstract)
This paper presents a replication algorithm that implements a highly-available, non-deterministic state machine. Our algorithm generalizes the Paxos parliament algorithm of Lamport to cope with non-deterministic computations, while preserving its nice resilience and e ciency properties. The algorithm is surprisingly simple, thanks to the use of two powerful underlying abstractions: weak consens...
متن کاملAnalysis of Static and Dynamic Configurability of Existing Group Communication Systems
Active replication following the state machine replication (SMR) approach is a way to make existing systems and services more reliable and fault-tolerant. The additional communication overhead has a negative impact on the system’s throughput and overall request latency. Today’s systems should be highly optimized to their execution environment and usage scenario in order to remedy the performanc...
متن کاملScaling Leader-Based Agreement Protocols for State Machine Replication
State machine replication is a technique used to guarantee the availability of a system even in the presence of faults. Agreement protocols are often used to implement state machine replication. However, the throughput of many agreement protocols, such as Paxos, does not scale with the number of machines available to the system. Systems whose throughput does scale often provide weaker consisten...
متن کاملAbstractions for Devising Byzantine-Resilient State Machine Replication
State machine replication is a common approach for making a distributed service highly available and resilient to failures, by replicating it on different processes. It is well-known, however, that the difficulty of ensuring the safety and liveness of a replicated service increases significantly when no synchrony assumptions are made, and when processes can exhibit Byzantine behaviors. The cont...
متن کامل