Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Performant Synchronization in Geo-Distributed Databases

Duling Xu, Tong Li, Zegang Sun, Zheng Chen, Weixing Zhou, Yanfeng Zhang, Wei Lu, Xiaoyong Du

TL;DR

The paper addresses the high WAN synchronization cost in geo-distributed databases by introducing GeoCoCo, a latency-aware, group-based synchronization framework. It combines three core ideas—latency-aware grouping, redundancy-aware filtering, and consistency-preserving hierarchical transmission—to reduce cross-region communication and improve throughput. The approach is demonstrated across trace-driven simulations and real deployments, showing up to 40.3% reduction in synchronization WAN cost and up to 14.1% throughput gains in GeoGauss (and 11.5% in CockroachDB), while maintaining strong consistency guarantees via CRDT epoch semantics. The work highlights the practical significance of topology-aware design for scaling geo-distributed databases and shows compatibility with existing systems and optimizations.

Abstract

The deployment of databases across geographically distributed regions has become increasingly critical for ensuring data reliability and scalability. Recent studies indicate that distributed databases exhibit significantly higher latency than single-node databases, primarily due to consensus protocols maintaining data consistency across multiple nodes. We argue that synchronization cost constitutes the primary bottleneck for distributed databases, which is particularly pronounced in wide-area networks (WAN). Fortunately, we identify opportunities to optimize synchronization costs in real production environments: (1) network clustering phenomena, (2) triangle inequality violations in transmission, and (3) redundant data transfers. Based on these observations, we propose GeoCoCo, a synchronization acceleration framework for cross-region distributed databases. First, GeoCoCo presents a group rescheduling strategy that adapts to real-time network conditions to maximize WAN transmission efficiency. Second, GeoCoCo introduces a task-preserving data filtering method that reduces data volume transmitted over the WAN. Finally, GeoCoCo develops a consistency-guaranteed transmission framework integrating grouping and pruning. Extensive evaluations in both trace-driven simulations and real-world deployments demonstrate that GeoCoCo reduces synchronization cost-primarily by lowering WAN bandwidth usage-by up to 40.3%, and increases system throughput by up to 14.1% in GeoGauss.

Performant Synchronization in Geo-Distributed Databases

TL;DR

The paper addresses the high WAN synchronization cost in geo-distributed databases by introducing GeoCoCo, a latency-aware, group-based synchronization framework. It combines three core ideas—latency-aware grouping, redundancy-aware filtering, and consistency-preserving hierarchical transmission—to reduce cross-region communication and improve throughput. The approach is demonstrated across trace-driven simulations and real deployments, showing up to 40.3% reduction in synchronization WAN cost and up to 14.1% throughput gains in GeoGauss (and 11.5% in CockroachDB), while maintaining strong consistency guarantees via CRDT epoch semantics. The work highlights the practical significance of topology-aware design for scaling geo-distributed databases and shows compatibility with existing systems and optimizations.

Abstract

The deployment of databases across geographically distributed regions has become increasingly critical for ensuring data reliability and scalability. Recent studies indicate that distributed databases exhibit significantly higher latency than single-node databases, primarily due to consensus protocols maintaining data consistency across multiple nodes. We argue that synchronization cost constitutes the primary bottleneck for distributed databases, which is particularly pronounced in wide-area networks (WAN). Fortunately, we identify opportunities to optimize synchronization costs in real production environments: (1) network clustering phenomena, (2) triangle inequality violations in transmission, and (3) redundant data transfers. Based on these observations, we propose GeoCoCo, a synchronization acceleration framework for cross-region distributed databases. First, GeoCoCo presents a group rescheduling strategy that adapts to real-time network conditions to maximize WAN transmission efficiency. Second, GeoCoCo introduces a task-preserving data filtering method that reduces data volume transmitted over the WAN. Finally, GeoCoCo develops a consistency-guaranteed transmission framework integrating grouping and pruning. Extensive evaluations in both trace-driven simulations and real-world deployments demonstrate that GeoCoCo reduces synchronization cost-primarily by lowering WAN bandwidth usage-by up to 40.3%, and increases system throughput by up to 14.1% in GeoGauss.

Paper Structure

This paper contains 22 sections, 12 equations, 19 figures, 1 table, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. BACKGROUND
  3. Geo-distributed Database System
  4. WAN Bandwidth and Latency
  5. Distributed Database System on WAN
  6. MOTIVATION
  7. GeoCoCo DESIGN
  8. GeoCoCo Overview
  9. Latency-aware Grouping Strategy Orchestrator
  10. Task-Preserved Data Filtering
  11. Consistency-Guaranteed Transmission
  12. IMPLEMENTATION
  13. EVALUATION
  14. Setup
  15. Micro Benchmark
  16. ...and 7 more sections

Figures (19)

  • Figure 1: Cross-Domain Communication Complexity in three Database Architectures: (a) Single Point vs. (b) Single Master vs. (c) Multi-Master.
  • Figure 2: Measured network latency between Amazon EC2 sites in 10 different regions.
  • Figure 3: Performance Impact of WAN Bandwidth and Latency on Multi-Master Systems Transactions.
  • Figure 4: Observed Geographical Clustering of Data Centers Worldwide.
  • Figure 5: Proportion of nodes that violate the triangle inequality (%).
  • ...and 14 more figures