Distributed Request Tracing
Project Status:
(roll over for info)
Write
The working group is currently writing the document. This is a collaborative approach using Google Docs.
Start:
October 1, 2020
Estimated Completion:
June 30, 2023
Problem Statement
Streaming operators struggle to provide exceptional viewer quality of experience (QoE) due to observability challenges. As an industry, we can significantly improve viewer QoE by leveraging distributed request tracing and the insight it enables. Streaming operators and service providers working together can collaboratively integrate logs, metrics, and request traces across the streaming video distribution workflow, dramatically increasing its overall observability for mutual benefit.
Project Description
This project focuses on the integration of distributed request tracing into the streaming video workflow, from player to CDN edge to CDN middle tier to origin. Once instrumented baseline traces will be captured to inform normal request/response patterns through the workflow. This will be followed by simulation of real-world degradations and failures to identify and codify corresponding trace patterns. If successful, these patterns can be leveraged for rapid root cause analysis in response to higher level QoE signals.
The first phase of the project will be simple, focusing on VOD delivery workflows. The purpose is to demonstrate the potential of distributed request tracing to significantly improve viewer delight.
Project Type
Document and Proof of Concept
Project Leads
Advisors
There are no SMEs associated with this project.
Goals and Objectives
The goals and objectives of this project are to:
- Identify baseline request/response patterns in the end to end workflow to understand normal behaviors across multiple players and CDNs.
- Simulate interesting degradations and failures across the workflow to identify corresponding request trace and aggregate patterns.
- Correlate high level QoE signals (slow start, low video quality, rebuffers) to underlying root causes and trace patterns.
- Develop metrics and visualizations which illustrate the analysis process.
Project Scope
The document produced by this project WILL PROVIDE:
- Technical details about trace instrumentation throughout the VOD delivery workflow.
- Identification of trace patterns related to interesting, real-world degradations and failures.
- A guide for leveraging trace patterns for rapid root cause analysis.
- CDN request-collapsing scenarios, live workflows, DRM encryption/decryption, ad delivery.
Contributors
The following members have contributed to this project. Click on their name to visit their profile. If they have not published their profile, the link will redirect to their LinkedIn profile.
Additional References
The following are recommended readings prior to participating in this project: