Poseidon: Efficient, Robust, and Practical Datacenter CC via Deployable INT
Poseidon: Efficient, Robust, and Practical Datacenter CC via Deployable INT
复制标题
DOI：
--
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Weitao Wang;M. Moshref;Yuliang Li;G. Kumar;T. Ng;Neal Cardwell;Nandita Dukkipati
中科院分区：
文献类型：
--
作者：
Weitao Wang;M. Moshref;Yuliang Li;G. Kumar;T. Ng;Neal Cardwell;Nandita Dukkipati
几十年来，越来越多的人可以在网络内部的CC中呈现更多的blibl cc，而不是在网络中，我们可以在bliblits interc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc copsoct的，这很难获得的杂货级交通状态（cc），这是一个造成了更多的con，blogys intecoits and this Protips and blime cc cc cc cc and the the topers and cc exts and toccots cc exts and的困难已成为杂货店。几个从根本上有利属性。 C模式，包括多人和反向路线交通拥堵。绩效收益在几种情况下，Poseidon将织物RTT降低了50％，将收敛时间最多减少到12倍，并将跨流的吞吐量变化降低了70％。
The difficulty in gaining visibility into the fine-timescale hop-level congestion state of networks has been a key chal-lenge faced by congestion control (CC) protocols for decades. However, the emergence of commodity switches supporting in-network telemetry (INT) enables more advanced CC. In this paper, we present Poseidon , a novel CC protocol that ex-ploits INT to address blind spots of CC algorithms and realize several fundamentally advantageous properties. First, Poseidon is efficient : it achieves low queuing delay, high throughput, and fast convergence. Furthermore, Poseidon decouples bandwidth fairness from the traditional AIMD control law, using a novel adaptive update scheme that converges quickly and smooths out oscillations. Second, Poseidon is robust : it realizes CC for the actual bottleneck hop , and achieves max-min fairness across traffic patterns, including multi-hop and reverse-path congestion. Third, Poseidon is practical : it is amenable to incremental brownfield deployment in networks that mix INT and non-INT switches. We show, via testbed and simulation experiments, that Poseidon provides significant improvements over the state-of-the-art Swift CC algorithm across key metrics – RTT, throughput, fairness, and convergence – resulting in end-to-end application performance gains. Evaluated across several scenarios, Poseidon lowers fabric RTT by up to 50%, reduces time to converge up to 12 × , and decreases throughput variation across flows by up to 70%. Collectively, these improvements reduce message transfer time by more than 61% on average and 14.5 × at 99.9p.