[ACM SIGCOMM 2024] "m3: Accurate Flow-Level Performance Estimation using Machine Learning" by Chenning Li, Arash Nasr-Esfahany, Kevin Zhao, Kimia Noorbakhsh, Prateesh Goyal, Mohammad Alizadeh, Thomas Anderson.
Proof of concept lib for creating finagle-like composable Services/Filters in node.
[TBD] "m4: A Learned Flow-level Network Simulator" by Chenning Li, Anton A. Zabreyko, Om Chabra, Arash Nasr-Esfahany, Kevin Zhao, Prateesh Goyal, Mohammad Alizadeh, Thomas Anderson.
Simulation of adversarial queueing (high packet loss) events in computer networks.
Reverse proxy which eliminates the tail latency caused by non-deterministic garbage collection
Chasing the long tail of Clojure HTTP servers.
Linux kernel with CTS scheduler enabled.
A deployment-oriented study of latency bias in real-time face recognition systems, showing that fairness violations emerge in tail inference latency rather than mean performance, with label-free auditing and mitigation analysis.
TailBench (Updated version)
Offline, fail-closed verifier for JSONL telemetry event logs. Emits deterministic audit certificates + human summaries with explicit claims/non-claims for bottleneck and integrity review.
A zero-overhead, transparent sidecar for UDP Request Hedging built with eBPF/BCC. Bypasses Python GC and Scheduler jitter by handling retries directly in the SoftIRQ context.
Request hedging for tail latency reduction in distributed systems
WASL: Multi-Module Coordination in Adaptive Multi-Tenant Clouds (ACM/SPEC ICPE'26)
Add a description, image, and links to the tail-latency topic page so that developers can more easily learn about it.
To associate your repository with the tail-latency topic, visit your repo's landing page and select "manage topics."