Design and Evaluation of a Novel Telemetry System for AI/ML Datacenter Networks - 2026_IDR_DEIB_24
Scientific-Disciplinary Group
09/IINF-03 - Telecommunications
Description
This research activity focuses on the design and evaluation of network telemetry systems tailored to AI/ML datacenter workloads. It will begin with an analysis of existing in-device data structures and key telemetry use cases, such as packet loss, congestion, and failure detection, under AI-driven traffic patterns. The candidate will then conduct an empirical study of representative AI clusters to characterize flow granularity, temporal behavior, and cross-switch traffic distribution, highlighting differences from traditional elephant-and-mice workloads. These insights will inform the redesign of sketch-based telemetry mechanisms to address emerging monitoring challenges in AI/ML datacenter networks, particularly in the presence of advanced multipath forwarding.
Job posting website
Number of positions
1
Funding body
Politecnico di Milano
Selection process
Click to expand
View the original posting on the MUR website: Go to MUR website