Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cold-Start Anti-Patterns and Refactorings in Serverless Systems: An Empirical Study

Syed Salauddin Mohammad Tariq, Foyzul Hassan, Amiangshu Bosu, Probir Roy

TL;DR

This study reframes serverless cold-start latency as a developer-centered design problem, deriving cross-layer anti-patterns, remediation strategies, and localization challenges from real-world issues. It then delivers SCABENCH, a reproducible benchmark, and InitScope, a hybrid import-time and runtime profiling framework, to enable explainable localization of initialization costs. Through a four-phase empirical workflow and a developer study, the work demonstrates that this approach improves localization accuracy and reduces diagnostic effort, validating an evidence-driven pathway for performance-aware serverless design. The findings advocate explicit initialization contracts and cross-layer tooling to address Initialization Debt and guide practitioners toward more predictable cold-start performance.

Abstract

Serverless computing simplifies deployment and scaling, yet cold-start latency remains a major performance bottleneck. Unlike prior work that treats mitigation as a black-box optimization, we study cold starts as a developer-visible design problem. From 81 adjudicated issue reports across open-source serverless systems, we derive taxonomies of initialization anti-patterns, remediation strategies, and diagnostic challenges spanning design, packaging, and runtime layers. Building on these insights, we introduce SCABENCH, a reproducible benchmark, and INITSCOPE, a lightweight analysis framework linking what code is loaded with what is executed. On SCABENCH, INITSCOPE improved localization accuracy by up to 40% and reduced diagnostic effort by 64% compared with prior tools, while a developer study showed higher task accuracy and faster diagnosis. Together, these results advance evidence-driven, performance-aware practices for cold-start mitigation in serverless design. Availability: The research artifact is publicly accessible for future studies and improvements.

Cold-Start Anti-Patterns and Refactorings in Serverless Systems: An Empirical Study

TL;DR

This study reframes serverless cold-start latency as a developer-centered design problem, deriving cross-layer anti-patterns, remediation strategies, and localization challenges from real-world issues. It then delivers SCABENCH, a reproducible benchmark, and InitScope, a hybrid import-time and runtime profiling framework, to enable explainable localization of initialization costs. Through a four-phase empirical workflow and a developer study, the work demonstrates that this approach improves localization accuracy and reduces diagnostic effort, validating an evidence-driven pathway for performance-aware serverless design. The findings advocate explicit initialization contracts and cross-layer tooling to address Initialization Debt and guide practitioners toward more predictable cold-start performance.

Abstract

Serverless computing simplifies deployment and scaling, yet cold-start latency remains a major performance bottleneck. Unlike prior work that treats mitigation as a black-box optimization, we study cold starts as a developer-visible design problem. From 81 adjudicated issue reports across open-source serverless systems, we derive taxonomies of initialization anti-patterns, remediation strategies, and diagnostic challenges spanning design, packaging, and runtime layers. Building on these insights, we introduce SCABENCH, a reproducible benchmark, and INITSCOPE, a lightweight analysis framework linking what code is loaded with what is executed. On SCABENCH, INITSCOPE improved localization accuracy by up to 40% and reduced diagnostic effort by 64% compared with prior tools, while a developer study showed higher task accuracy and faster diagnosis. Together, these results advance evidence-driven, performance-aware practices for cold-start mitigation in serverless design. Availability: The research artifact is publicly accessible for future studies and improvements.

Paper Structure

This paper contains 28 sections, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Cold-Start Latency: Application vs. Platform Causes
  4. Developer Practices and Diagnostic Gaps
  5. Empirical Studies and Taxonomy Gaps
  6. Research Questions and Study Design
  7. Research Questions
  8. Empirical Framework
  9. Phase I - Mining-Driven Taxonomy Construction
  10. Phase II: SCABench Microbenchmark Construction
  11. Phase III: Profiling and Comparative Evaluation on SCABench
  12. Phase IV: Developer User Study
  13. Evaluation
  14. RQ1: Which initialization anti-patterns occur in practice?
  15. RQ2: Which refactoring strategies remediate these anti-patterns, and how are they applied?
  16. ...and 13 more sections

Figures (3)

  • Figure 1: Empirical workflow and RQ mapping. Phase I: issue mining $\rightarrow$ taxonomies. Phase II: SCABench construction. Phase III: profiling and comparative evaluation on SCABench. Phase IV: developer study.
  • Figure 2: Unified taxonomy: initialization anti-patterns (A1–A6), refactoring strategies (R1–R6), and localization challenges (L1–L5).
  • Figure 3: Co-occurrence heatmap of initialization anti-patterns.