A Gittins Policy for Optimizing Tail Latency
Service level objectives (SLOs) for queueing systems typically relate to the tail of the system's response time distribution T. The tail is the function mapping a time t to the probability P[T > t]. SLOs typically ask that high percentiles of T are not too large, i.e. that P[T > t] is small for large t.
Authors
- Ziv Scully (ORCID: https://orcid.org/0000-0002-8547-1068)
- George Yu (ORCID: https://orcid.org/0009-0005-3013-8558)
- Amit Harlev (ORCID: https://orcid.org/0000-0002-2627-0012)
Institutions
- Cornell University (US)
Publication Details
- Journal
- ACM SIGMETRICS Performance Evaluation Review
- Published
- 2024-09-05
- DOI
- https://doi.org/10.1145/3695411.3695418
- Citations
- 2
- Primary Topic
- Advanced Queuing Theory Analysis
- Type
- article
- Field-Weighted Citation Impact
- 1.06