.png){kind=icon}
.png){kind=icon}
Return
DTraComp: Comparing distributed execution traces for understanding intermittent latency sources
.png){kind=icon}
PRE
AI
M
F
M
M
N
M
DOI：10.1016/j.jss.2026.112990.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Microservice architectures significantly enhance software development by offering flexibility in programming languages and deployment infrastructures. They facilitate the isolation of failures within individual services and simplify the process of debugging and fixing issues in independent services. Nonetheless, pinpointing performance degradation becomes a challenging task due to the complex interactions and parallelism among numerous service instances. While end-to-end tracing aids in mapping execution paths across services and identifying latencies, its ability is limited to gathering high-level information, not fully capable of pinpointing the root causes of performance degradation among processes. Additionally, a notable gap in many current performance analysis tools is the absence of a comparison feature, which is crucial for providing developers with a detailed perspective on the performance variations between different sets of requests. This paper introduces DTraComp (Distributed Trace Compare), an open-source framework, compatible with various microservice trace standards, and integrated with Eclipse Trace Compass™1. The demo is available on YouTube2. Our framework offers robust visual comparison capability for two groups of requests within distributed systems, which includes nested spans executed in parallel. Furthermore, it provides system kernel details for each thread involved in the execution of each span, allowing it to pinpoint the reasons for performance degradation across distributed systems. We used our proposed framework to analyze five practical use cases. By evaluating the efficiency of our tool, it was determined that the overall time complexity scales linearly O(n) with the trace size, indicating its suitability for deployment in production environments. It is currently used within Ericsson company for performance evaluation purposes.
Journal
IF:
4.1
Papers:
5.4K
Citations:
8.4K
Organization
Cited Papers