International Journal For Multidisciplinary Research

E-ISSN: 2582-2160     Impact Factor: 9.24

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Call for Paper Volume 7, Issue 4 (July-August 2025) Submit your research before last 3 days of August to publish your research paper in the issue of July-August.

Efficient Reduction of Commit Delays in Consensus Based Systems

Author(s) Kanagalakshmi Murugan
Country United States
Abstract ETCD is a highly available key-value store often used in distributed systems for configuration management and service discovery. It is based on the Raft consensus algorithm, which ensures that the data in the system is consistently replicated across all nodes. One of the critical performance factors for etcd is the latency involved in committing writes, which directly affects the overall responsiveness and availability of the system. The latency of commit operations can be influenced by multiple factors, including network delays, disk I/O, and the underlying consensus mechanism used. SMR (State Machine Replication) is a fundamental technique used in distributed systems to ensure consistency across multiple nodes. In an SMR-based system, a set of replicas maintains an identical state, and each replica processes the same sequence of operations in the same order. The Raft consensus algorithm, which is used by etcd, is a popular implementation of SMR. While SMR ensures strong consistency and fault tolerance, it introduces overhead in terms of latency, particularly as the number of nodes increases. The latency of commit operations in SMR-based systems is influenced by the need to communicate with a quorum of nodes before a commit can be finalized. As the system scales, the time required to reach consensus increases, leading to higher commit latency. For example, in a system with fewer nodes, the communication required for consensus is minimal, resulting in low latency. However, as the number of nodes grows, the consensus process becomes more complex, and network delays between nodes can further increase latency. In distributed systems that require low-latency operations, minimizing commit latency is crucial. However, there is often a trade-off between latency and consistency. While SMR offers strong consistency guarantees, achieving low latency becomes more challenging as the system scales. Therefore, optimizing SMR algorithms to reduce the latency of consensus while maintaining fault tolerance and consistency remains a key area for future research. In conclusion, while etcd’s use of SMR ensures high reliability and consistency, the associated latency in commit operations can become a bottleneck in large-scale distributed systems. Efforts to optimize commit latency in these systems are essential for improving performance without compromising consistency. This paper addresses the commit latency issue using write ahead log WAL.
Published In Volume 6, Issue 3, May-June 2024
Published On 2024-06-05
DOI https://doi.org/10.36948/ijfmr.2024.v06i03.48249
Short DOI https://doi.org/

Share this