# Scalable Tail Latency Estimation for Data Center Networks

arXiv: https://arxiv.org/abs/2205.01234 · HTML: https://mlsanity.com/paper/2205.01234/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 May 2022.

## Built on by (most cited first, 4 of 4)

- [RouteNet-Gauss: Hardware-Enhanced Network Modeling with Machine Learning](https://mlsanity.com/paper/2501.08848/) (2025)
- [Janus: An Algorithm-Evaluator Co-Evolution Framework for LLM-Driven Discovery under Expensive Evaluation Budgets](https://mlsanity.com/paper/2608.08189/) (2026)
- [Building a Graph-based Deep Learning network model from captured traffic traces](https://mlsanity.com/paper/2310.11889/) (2023)
- [SwiftQueue: Optimizing Low-Latency Applications with Swift Packet Queuing](https://mlsanity.com/paper/2410.06112/) (2024)

## Abstract (arXiv)

In this paper, we consider how to provide fast estimates of flow-level tail latency performance for very large scale data center networks. Network tail latency is often a crucial metric for cloud application performance that can be affected by a wide variety of factors, including network load, inter-rack traffic skew, traffic burstiness, flow size distributions, oversubscription, and topology asymmetry. Network simulators such as ns-3 and OMNeT++ can provide accurate answers, but are very hard to parallelize, taking hours or days to answer what if questions for a single configuration at even moderate scale. Recent work with MimicNet has shown how to use machine learning to improve simulation performance, but at a cost of including a long training step per configuration, and with assumptions about workload and topology uniformity that typically do not hold in practice. We address this gap by developing a set of techniques to provide fast performance estimates for large scale networks with general traffic matrices and topologies. A key step is to decompose the problem into a large number of parallel independent single-link simulations; we carefully combine these link-level simulations to produce accurate estimates of end-to-end flow level performance distributions for the entire network. Like MimicNet, we exploit symmetry where possible to gain additional speedups, but without relying on machine learning, so there is no training delay. On large-scale networks where ns-3 takes 11 to 27 hours to simulate five seconds of network behavior, our techniques run in one to two minutes with 99th percentile accuracy within 9% for flow completion times.