Sep 2026· Proceedings of the 17th ACM SIGOPS Asia-Pacific Workshop on Systems· 0 citations· 30 references
TL;DR
It is shown that TCP performance can be dramatically improved by following the design philosophy of RoCE, which integrates kernel bypassing, segment offloading, and zero-copy buffer management, and software TCP stacks can achieve performance parity with modern RNICs.
Abstract
RDMA delivers a high-performance data plane with minimal CPU overhead in datacenters, but its hardware-integrated control plane suffers from serious scalability limitations. Recent work on improving RoCEv2-based NIC (RNIC) scalability has helped, yet these solutions depend on specialized hardware features, limiting their practical adoption. In this paper, we ask the opposite question: can a software-based data plane match RDMA's performance while preserving the flexibility of a software control plane? We show that TCP performance can be dramatically improved by following the design philosophy of RoCE, which integrates kernel bypassing, segment offloading, and zero-copy buffer management. Our preliminary evaluation shows that a single CPU core can sustain over 1 Tbps on TCP transmission (TX) and 752 Gbps on reception (RX). These results demonstrate that software TCP stacks can achieve performance parity with modern RNICs, providing a flexible and practical alternative for high-performance networked systems.
Many data centers adopt Remote Direct Memory Access (RDMA) to allow data center applications to achieve low latency and high throughput, while keeping minimal CPU overhead. The upper-layer applications keep evolving rapidly, and thus need congestion control algorithms (CCAs) that exist in the NIC hardware also to react...
Hong-Jing Huang, Xu-Zheng Chen, Zi-Yu Song et al.· IEEE Transactions on Network...· 0 citations
With the rapid development of network intensive applications such as LLM training and cloud storage, RoCE has been widely deployed in data center networks to enhance the underlying high-speed network and meet the high-performance requirements of upper layer applications. The existing host-based RoCE load balancing sche...
Hui-Min Luo, Jiao Zhang, Jia-Feng Jiang et al.· IEEE Transactions on Network...· 0 citations
Results demonstrate that RSC can make RDMA economically viable for serverless services, and workflow, request‑mix, resource, noisy‑neighbor, endpoint‑scale, and QP‑density diagnostics confirm that connection reuse removes millisecond‑scale RC setup from warm service invocation paths.
Wen-Di Song, Guang-Ping Xu, Yang-Yang Fan et al.· Proceedings of the Internati...· 0 citations
Erasure coding provides efficient fault tolerance for large-scale distributed storage systems. However, its data repair process is well-known to be resource-intensive. We find that conventional host-centric, TCP-based repair architectures suffer from severe resource contention. Even in high-bandwidth networks, such int...
Falling launch costs, together with advances in radiation-tolerant electronics, optical inter-satellite links, and high-efficiency power and thermal systems, have made Space Data Centers (SDCs) an increasingly credible proposition, whether for deep-space autonomy, on-board processing of Earth-observation data, or as or...
Soyed Tuhin Ahmed, B. Farahani, Krishnendu Chakrabarty et al.· Coins· 0 citations
This work presents ActiveRDMA, an active RDMA model that extends traditional one-sided semantics by leveraging on-NIC programmability, and enables complex, multi-step RDMA patterns to execute directly on the target side, eliminating host CPU involvement and reducing communication round-trips.
Jerónimo Sánchez García, Peter-Jan Gootzen, Raphael Frantz et al.· Proceedings of the Internati...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
