EUVIMEDEUVIMED entdecken
Crossref

dVRM: Cross-Switch Memory Sharing and Self-Adaptive Allocation in Distributed Data Plane

Mimi Qian, Lin Cui, Fung Po Tso, Yuhui Deng, Zhen Zhang, Weijia Jia

IEEE Transactions on Computers · 2026-07 · 75 (7) · 2382-2393

DOI: 10.1109/tc.2026.3678701 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Computers, 2026-07. Verfasst von Mimi Qian und 5 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Innovate UK.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Programmable switches have revolutionized networking by enabling a new spectrum of applications, such as network telemetry, in-network computation, and machine learning. These applications heavily utilize register memory but their performance is significantly constrained by the scarcity of on-chip resources, such as the 15 MB of SRAM available on a Tofino switch. To effectively accommodate increasingly memory-demanding applications, we aim to pool register resources across multiple switches, creating a larger unified register memory space. This resource pooling approach addresses the limitations of existing single-switch Virtual Register Memory (VRM) solutions, which cannot meet the demands of these applications in distributed environments. To achieve this, we propose dVRM, a distributed VRM deployment framework that enables cross-switch memory sharing and self-dynamic memory allocation on the data plane. dVRM introduces three innovations: (1) Grouped Multi-Switch Registers (GMRs), virtualizing distributed pipeline stages into a unified memory pool; (2) a self-adaptive, bit-width allocation mechanism driven by real-time data-plane feedback; and (3) lightweight heuristics for concurrent application deployment with distributed VRM, formulated as a mixed-integer linear programming (MILP) problem. We have implemented dVRM on both P4 hardware switches (with Intel Tofino ASIC) and BMv2. Experimental results show that dVRM significantly reduces hash unit consumption by up to 26% and achieves an improvement in accuracy (ARE) of up to 57.3% across various workloads.

Förderung

  • National Natural Science Foundation of China
  • Innovate UK

Zitieren

APA

Qian, M., Cui, L., Tso, F. P., Deng, Y., Zhang, Z., & Jia, W. (2026). dVRM: Cross-Switch Memory Sharing and Self-Adaptive Allocation in Distributed Data Plane. IEEE Transactions on Computers, 75(7), 2382-2393. https://doi.org/10.1109/tc.2026.3678701

Vancouver

Qian M, Cui L, Tso FP, Deng Y, Zhang Z, Jia W. dVRM: Cross-Switch Memory Sharing and Self-Adaptive Allocation in Distributed Data Plane. IEEE Transactions on Computers. 2026;75(7):2382-2393. doi:10.1109/tc.2026.3678701

BibTeX
@article{Qian2026,
  author = {Qian, Mimi and Cui, Lin and Tso, Fung Po and Deng, Yuhui and Zhang, Zhen and Jia, Weijia},
  title = {dVRM: Cross-Switch Memory Sharing and Self-Adaptive Allocation in Distributed Data Plane},
  journal = {IEEE Transactions on Computers},
  year = {2026},
  volume = {75},
  number = {7},
  pages = {2382--2393},
  doi = {10.1109/tc.2026.3678701}
}
Kostenlos starten →