On the Burstiness of Distributed Machine Learning Traffic
Traffic from distributed training of machine learning (ML) models makes up a large and growing fraction of the traffic mix in enterprise data centers. While work on distributed ML abounds, the network traffic generated by distributed ML has received little attention. Using measurements on a testbed network, we investigate the traffic characteristics generated by the training of the ResNet-50 neural network with an emphasis on studying its short-term burstiness. For the latter we propose metrics that quantify traffic burstiness at different time scales. Our analysis reveals that distributed ML traffic exhibits a very high degree of burstiness on short time scales, exceeding a 60:1 peak-to-mean ratio on time intervals as long as 5~ms. We observe that training software orchestrates transmissions in such a way that burst transmissions from different sources within the same application do not result in congestion and packet losses. An extrapolation of the measurement data to multiple applications underscores the challenges of distributed ML traffic for congestion and flow control algorithms.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Using Bursty Announcements for Detecting BGP Routing Anomalies
Despite the robust structure of the Internet, it is still susceptible to disruptive routing updates that prevent network traffic from reaching its destination. Our research shows that BGP announcements that are associate…
Anomaly DetectionCommon TF-IDF variants arise as key components in the test statistic of a penalized likelihood-ratio test for word burstiness
TF-IDF is a classical formula that is widely used for identifying important terms within documents. We show that TF-IDF-like scores arise naturally from the test statistic of a penalized likelihood-ratio test setup captu…
Document ClassificationOn the LRD of the Aggregated Traffic Flows in High-Speed Computer Networks
This paper studies and analyses the behavior of the Long-Range Dependence in network traffic after classifying traffic flows in aggregated time series. Following Differentiated Services architecture principles, the gener…
Time SeriesTime Series AnalysisOn the LRD of the Aggregated Traffic Flows in High-Speed Computer Networks
This paper studies and analyses the behavior of the Long-Range Dependence in network traffic after classifying traffic flows in aggregated time series. Following Differentiated Services architecture principles, the gener…
Time SeriesTime Series AnalysisMeasurement-based Online Available Bandwidth Estimation employing Reinforcement Learning
An accurate and fast estimation of the available bandwidth in a network with varying cross-traffic is a challenging task. The accepted probing tools, based on the fluid-flow model of a bottleneck link with first-in, firs…
reinforcement-learningReinforcement LearningReinforcement Learning (RL)