Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Benchmarking Data Management Systems for Microservices

Rodrigo Laigner, Yongluan Zhou

TL;DR

The paper addresses the gap between practical microservice deployments and existing benchmarks by proposing Online Marketplace, a benchmark that embeds distributed transaction processing, data replication semantics, event processing, and data integrity constraints for data-intensive microservice workloads. It formalizes the workload, driver, and data-management criteria, and implements four benchmark variants on Orleans and Statefun to reveal platform capabilities and limitations. The findings indicate that no single data platform can satisfy all core data-management requirements simultaneously; a customized architecture that combines transactional guarantees with external storage can broaden functionality. Overall, the benchmark aims to drive the design of next-generation data-management platforms for microservices and provides a practical framework for evaluating end-to-end data consistency, performance, and reliability in real-world-like workloads.

Abstract

Microservice architectures are a popular choice for deploying large-scale data-intensive applications. This architectural style allows microservice practitioners to achieve requirements related to loose coupling, fault contention, workload isolation, higher data availability, scalability, and independent schema evolution. Although the industry has been employing microservices for over a decade, existing microservice benchmarks lack essential data management challenges observed in practice, including distributed transaction processing, consistent data querying and replication, event processing, and data integrity constraint enforcement. This gap jeopardizes the development of novel data systems that embrace the complex nature of data-intensive microservices. In this talk, we share our experience in designing Online Marketplace, a novel benchmark that embraces core data management requirements intrinsic to real-world microservices. By implementing the benchmark in state-of-the-art data platforms, we experience the pain practitioners face in assembling several heterogeneous components to realize their requirements. Our evaluation demonstrates Online Marketplace allows experimenting key properties sought by microservice practitioners, thus fomenting the design of novel data management systems.

Benchmarking Data Management Systems for Microservices

TL;DR

The paper addresses the gap between practical microservice deployments and existing benchmarks by proposing Online Marketplace, a benchmark that embeds distributed transaction processing, data replication semantics, event processing, and data integrity constraints for data-intensive microservice workloads. It formalizes the workload, driver, and data-management criteria, and implements four benchmark variants on Orleans and Statefun to reveal platform capabilities and limitations. The findings indicate that no single data platform can satisfy all core data-management requirements simultaneously; a customized architecture that combines transactional guarantees with external storage can broaden functionality. Overall, the benchmark aims to drive the design of next-generation data-management platforms for microservices and provides a practical framework for evaluating end-to-end data consistency, performance, and reliability in real-world-like workloads.

Abstract

Microservice architectures are a popular choice for deploying large-scale data-intensive applications. This architectural style allows microservice practitioners to achieve requirements related to loose coupling, fault contention, workload isolation, higher data availability, scalability, and independent schema evolution. Although the industry has been employing microservices for over a decade, existing microservice benchmarks lack essential data management challenges observed in practice, including distributed transaction processing, consistent data querying and replication, event processing, and data integrity constraint enforcement. This gap jeopardizes the development of novel data systems that embrace the complex nature of data-intensive microservices. In this talk, we share our experience in designing Online Marketplace, a novel benchmark that embraces core data management requirements intrinsic to real-world microservices. By implementing the benchmark in state-of-the-art data platforms, we experience the pain practitioners face in assembling several heterogeneous components to realize their requirements. Our evaluation demonstrates Online Marketplace allows experimenting key properties sought by microservice practitioners, thus fomenting the design of novel data management systems.
Paper Structure (4 sections, 1 figure)

This paper contains 4 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Online Marketplace Benchmark
  3. Implementations and Experiments
  4. Conclusion

Figures (1)

  • Figure 1: Customized Orleans-based Solution. HTTP Layer parses HTTP requests and forwards them to the correct grains. Orleans Silo provides location and life-cycle transparency for grains. Events are modeled as asynchronous messages exchanged by grains. Storage Layer contains a primary-secondary deployment based on Redis to support causal replication of product updates, grain storage to manage grain states and log storage to store audit logging.