Important Dates (Times are AoE)
- Feb 28, 2020: Paper Submission
- April 3, 2020: Notification of Acceptance
- April 17, 2020: Camera Ready Deadline
Modern computer systems such as serverless systems and microservice systems are growing in popularity and complexity. Developers today often design systems that integrate 100s to 1000s of components provided by other teams or third-party sources. Thus, it is often difficult to understand the system-wide interactions between all the different components within these modern systems. As a result, there has been a rise in research on diagnosing bugs in these systems. This workshop focuses specifically on performance bugs, which are notoriously difficult to detect and diagnose. As performance bugs do not cause crashes or lead to functionality flaws, they do not readily lend themselves to traditional debugging techniques. The goal of this workshop is to brainstorm new directions for performance debugging in modern systems with a focus on discussion and providing feedback to the presenters.
The Workshop on Performance Debugging in Modern Computer Systems aims to bring together researchers interested in performance debugging from a variety of communities such as systems/networking, analytical performance modeling, machine learning, programming languages, and visualization. Researchers from these communities have previously provided unique insights, ideas, and techniques for performance debugging. Bringing together researchers from each of these communities will spur collaboration and inspire new approaches that may revolutionize the way people debug performance issues.
Topics of interest include but are not limited to:
- Serverless or microservice architecture designs with explicit debugging support
- Methodologies for debugging cross-layer performance problems in the system stack
(e.g., application, system framework, kernel, network, etc.)
- Performance modeling and/or analysis of serverless and/or microservice systems
- New approaches to performance debugging in distributed systems
- Root cause analysis of performance issues
- Performance anomaly detection
- Distributed tracing/log analysis
- System support for distributed tracing
- Performance profiling
- Performance regression testing
- System design for observability and debuggability
- Addressing public cloud computing performance variability issues
- Visualization for performance debugging
The workshop will be a full day and include a keynote and three sessions where authors of accepted papers will present their ideas. There will also be substantial time dedicated to questions and discussion to foster collaboration and provide feedback.
Authors must submit original papers that are not previously published or concurrently submitted elsewhere. The workshop welcomes a broad range of paper types including but not limited to research, vision, industry, position, etc. Papers should be at most 5 pages excluding references. Reviewing is double-blind, so authors should not identify themselves either explicitly or implicitly. All submissions must be in PDF format using the IEEE 8.5″ x 11″ twocolumn format with a 10-point Times-Roman or similar typeface font and numbered pages. A template is available at https://www.ieee.org/conferences/publishing/templates.html, but authors can use other templates or editors that conform with the requirements. Submissions that do not follow the formatting requirements may be rejected without review.
Submissions site: TBD
Conference/Workshop Attendance Policy
The workshop is co-located with ICDCS 2020 (July 8 – 10th) and will run in parallel with the conference sessions. There is no workshop-only registration, so each accepted paper must have at least one author registered for the ICDCS conference. For authors with multiple papers accepted by workshops or the main conference, a separate author registration is required for each paper.