Search

Applications that run in large-scale data center networks (DCNs) rely on the DCN's ability to deliver application requests in a performant manner. DCNs expose a complex design and operational space, and network designers and operators care how different options along this space affect application performance. One might run controlled experiments and measure the corresponding application-facing performance, but such experiments become progressively infeasible at a large scale, and simulations risk yielding inaccurate or incomplete results. Instead, we show that we can predict application-facing performance through more easily measured network metrics. For example, network telemetry metrics (e.g., link utilization) can predict application-facing metrics (e.g., transfer latency). Through large-scale measurements of production networks, we study the correlation between the two types of metrics, and construct predictive, interpretable models that serve as a suggestive guideline to network designers and operators. We show that no single network metric is universally the best predictor (even though some prior work has focused on a single predictor). We found that simple linear models often have the lowest error, while queueing-based models are better in a few cases., Comment: 17 (main body) + 5 (appendix) pages

To reduce cost, datacenter network operators are exploring blocking network designs. An example of such a design is a "spine-free" form of a Fat-Tree, in which pods directly connect to each other, rather than via spine blocks. To maintain application-perceived performance in the face of dynamic workloads, these new designs must be able to reconfigure routing and the inter-pod topology. Gemini is a system designed to achieve these goals on commodity hardware while reconfiguring the network infrequently, rendering these blocking designs practical enough for deployment in the near future. The key to Gemini is the joint optimization of topology and routing, using as input a robust estimation of future traffic derived from multiple historical traffic matrices. Gemini "hedges" against unpredicted bursts, by spreading these bursts across multiple paths, to minimize packet loss in exchange for a small increase in path lengths. It incorporates a robust decision algorithm to determine when to reconfigure, and whether to use hedging. Data from tens of production fabrics allows us to categorize these as either low-or high-volatility; these categories seem stable. For the former, Gemini finds topologies and routing with near-optimal performance and cost. For the latter, Gemini's use of multi-traffic-matrix optimization and hedging avoids the need for frequent topology reconfiguration, with only marginal increases in path length. As a result, Gemini can support existing workloads on these production fabrics using a spine-free topology that is half the cost of the existing topology on these fabrics.

Predictably sharing the network is critical to achieving high utilization in the datacenter. Past work has focussed on providing bandwidth to endpoints, but often we want to allocate resources among multi-node services. In this paper, we present Parley, which provides service-centric minimum bandwidth guarantees, which can be composed hierarchically. Parley also supports service-centric weighted sharing of bandwidth in excess of these guarantees. Further, we show how to configure these policies so services can get low latencies even at high network load. We evaluate Parley on a multi-tiered oversubscribed network connecting 90 machines, each with a 10Gb/s network interface, and demonstrate that Parley is able to meet its goals.