Speaking Skills Talk - Christoper Canel June 18, 2024 3:00pm — 4:00pm Location: In Person - Gates Hillman 8102 Speaker: CHRISTOPHER CANEL, Ph.D. Student, Computer Science Department, Carnegie Mellon University https://www.ccanel.com/ Measuring and Mitigating Incast Bursts in Modern Datacenters In datacenter networks, common many-to-one traffic patterns known as incast are challenging because they violate the basic premise of bandwidth stability on which Transmission Control Protocol (TCP) congestion control is built, overwhelming shallow switch buffers and causing packet loss and high application latency. First, to understand why these challenges remain despite decades of research on datacenter transport, this talk presents an in-depth investigation into high-degree incasts both in production workloads at a major datacenter operator and in simulation. In addition to characterizing the bursty nature of these incasts and their impacts on the network, our findings demonstrate the shortcomings of widely deployed window-based congestion control techniques used to address incast problems. Second, we show how congestion control can leverage existing features in TCP to mitigate high-degree incast without a new protocol design. Drawing on simple receiver-assisted congestion control techniques, we demonstrate that tuning an incast receiver's TCP advertised window (RWND) provides the coordination necessary to help independent senders reduce the volume of in-flight data to keep queues short and prevent packet loss. Furthermore, we find that hosts associated with a specific application or service exhibit similar and predictable incast traffic properties across hours, enabling us to avoid computationally expensive online adaptation and instead adopt a static RWND tuning approach with low overheads. We draw on three years of production deployment experience to demonstrate that RWND tuning for incast bursts reduces retransmissions by 3.8x, leading to a 2—8x reduction in application tail latency. Presented in Partial Fulfillment of the CSD Speaking Skills Requirement Event Website: https://csd.cmu.edu/calendar/speaking-skills-talk-christoper-canel