*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title
PerfDB + PerfML: Supporting Big Data-Driven Research on Fine-Grained Performance Phenomena
Joshua Kimball
Ph.D. Student in Computer Science
School of Computer Science
Georgia Institute of Technology
Date: November 20, 2020
Time: 10:00 AM to 12:00PM EST
Location: https://bluejeans.com/287403017
**Note: this proposal is remote-only. complete meeting info. below.**
Committee
Dr. Calton Pu (Advisor) - School of Computer Science, Georgia Institute of Technology
Dr. Arulaj Joy - School of Computer Science, Georgia Institute of Technology
Dr. Ling Liu - School of Computer Science, Georgia Institute of Technology
Dr. Sham Navathe - School of Computer Science, Georgia Institute of Technology
Dr. Qingyang Wang - School of Computer Science, Louisiana State University
Abstract
The long-tail latency problem is a well-documented problem in large-scale system topologies like cloud platforms. Long-tail latency can lead to degraded quality of experience, potential economic loss and less predictable system performance overall. In the past, platforms have over-provisioned to mitigate tail latency and its effects. Instead, we propose studying these performance bugs systematically and addressing their underlying root cause. In this proposal, we present the first-generation population study of VLRT requests using PerfDB, our data management system. Specifically, we present the first study of VLRT periods on integrated data from 500 experiments. We find evidence that many VLRT coincide with Cross-Tier Queue Overflow (CTQO) induced by millibottlenecks. In our second analysis, we conduct the first study of a phenomenon called Localized Latency Requests (LLR)—requests with latency between 100-500ms. Finally, we propose a machine learning system, PerfML, which employs a teamed-classifier method to automatically isolate and diagnose fine-grained performance anomalies found in a collection of fine-grained (sampling at 50ms intervals) resource monitoring data covering a wide variety of hardware and software configurations spanning hundreds of experiments and terabytes of experimental data.
-----------------------------------------
Meeting URL
https://bluejeans.com/287403017
Meeting ID
287 403 017
Want to dial in from a phone?
Dial one of the following numbers:
+1.408.419.1715 (United States(San Jose))
+1.408.915.6290 (United States(San Jose))
(see all numbers - https://www.bluejeans.com/numbers)
Enter the meeting ID and passcode followed by #
Connecting from a room system?
Dial: bjn.vc or 199.48.152.152 and enter your meeting ID & passcode
