Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a multi-agent framework for optimizing the performance of microservice-based systems by integrating control-flow, data-flow, architectural, and cross-component dependency information. The framework decomposes the optimization process into coordinated agent roles (summarization, analysis, optimization, verification) that collaboratively identify bottlenecks and construct multi-step optimization strategies. Results on a proof-of-concept microservice system demonstrate a 36.58% improvement in throughput and a 27.81% reduction in average response time, showcasing the potential of system-level reasoning for software optimization.
LLMs can now orchestrate system-wide optimizations across microservices, boosting throughput by 36% and slashing response times by 27%.
Large language models and AI agents have recently shown promise in automating software performance optimization, but existing approaches predominantly rely on local, syntax-driven code transformations. This limits their ability to reason about program behavior and capture whole system performance interactions. As modern software increasingly comprises interacting components - such as microservices, databases, and shared infrastructure - effective code optimization requires reasoning about program structure and system architecture beyond individual functions or files. This paper explores the feasibility of whole system optimization for microservices. We introduce a multi-agent framework that integrates control-flow and data-flow representations with architectural and cross-component dependency signals to support system-level performance reasoning. The proposed system is decomposed into coordinated agent roles - summarization, analysis, optimization, and verification - that collaboratively identify cross-cutting bottlenecks and construct multi-step optimization strategies spanning the software stack. We present a proof-of-concept on a microservice-based system that illustrates the effectiveness of our proposed framework, achieving a 36.58% improvement in throughput and a 27.81% reduction in average response time.