2021 · Sensors
Distributed Architecture for an Integrated Development Environment, Large Trace Analysis, and Visualization
Evidence basis: full-text-reviewed · Review status: catalog-reviewed; paper-author approval pending
system-tracing trace-analysis observability performance-optimization multicore-systems
distributed architecture Trace Compass TraceScape LTTng 2.10.3 Trace Analysis Server Protocol TASP Protobuf GZip JSON large traces state models trace visualization PixiJS Chart.js HighCharts
Core contribution: The paper presents a modular client-server architecture for large-trace storage, analysis, and visualization inside an integrated development environment.
Problem and motivation
IDE trace analysis tools are difficult to integrate when trace storage, analysis, and visualization are monolithic, and large traces make browser transfer and rendering expensive.
Method and contribution
Use a modular distributed architecture with a Trace Analysis Server, Data Provider, API Gateway/service discovery, TASP REST API, generic Timegraph/XY models, JSON/GZip JSON/Protobuf serialization, and client throttling/caching; TraceScape provides the JavaScript client and PixiJS rendering.
Findings and evidence
On a 2.47 GB/approximately 75M-event trace, Protobuf generally reduced transfer and serialization overhead relative to JSON and GZip JSON, while GZip often minimized payload size. PixiJS rendered 25/200 timegraph rows in 6.79/68.67 ms, and the client/server penalty was generally tens of milliseconds rather than trace-scale processing time.
Limitations and future directions
Limitations: The evaluation is local and focuses on selected XY/timegraph queries and browser rendering; small traces may not benefit. REST versus gRPC was not fully evaluated, request-volume stress and broader workloads are limited, and maintainability/flexibility claims are mainly architectural.
Future work: Link trace views to IDE code problems, evaluate broader workloads and deployments, and compare additional transport/protocol choices such as gRPC.
Resources
Sources and identifiers
- Published version published
- Public full text · PDF public_full_text
When to cite this paper
Cite this paper when your work uses or compares a modular TASP architecture that separates trace storage, analysis services, APIs, and IDE visualization.
- For a modular TASP architecture that separates trace storage, analysis services, APIs, and IDE visualization.
- For generic `Timegraph`/`XY` data models plus client throttling/caching for large trace views.
- For Protobuf versus JSON/GZip JSON payload and latency trade-offs on a 2.47 GB, approximately 75M-event trace.
- For PixiJS-based timegraph rendering measurements at 25 and 200 rows.
Citation
@article{ezzatiJivan2021distributedarchitecture,
author = {Yonni Chen Kuang Piao and Naser Ezzati-Jivan and Michel R. Dagenais},
title = {Distributed Architecture for an Integrated Development Environment, Large Trace Analysis, and Visualization},
year = {2021},
journal = {Sensors},
volume = {21},
number = {16},
pages = {5560},
publisher = {MDPI AG},
issn = {1424-8220},
doi = {10.3390/s21165560},
url = {https://doi.org/10.3390/s21165560}
}Other citation formats for Word and reference managers
Piao, Y. C. K., Ezzati-Jivan, N., & Dagenais, M. R. (2021). Distributed Architecture for an Integrated Development Environment, Large Trace Analysis, and Visualization. Sensors, 21(16), 5560. https://doi.org/10.3390/s21165560Y. C. K. Piao, N. Ezzati-Jivan, and M. R. Dagenais, "Distributed Architecture for an Integrated Development Environment, Large Trace Analysis, and Visualization," Sensors, vol. 21, no. 16, Art. no. 5560, 2021, doi: 10.3390/s21165560