Multitenant In-Network Acceleration with SwitchVM
Sajy Khashab, Alon Rashelbach, Mark Silberstein
Abstract
We propose a practical approach to implementing multitenancy on programmable network switches to make in-network acceleration accessible to cloud users. We introduce a Switch Virtual Machine (SwitchVM), that is deployed on the switches and offers an expressive instruction set and program state abstractions. Tenant programs, called Data-Plane Filters (DPFs), are executed on top of SwitchVM in a sandbox with memory, network and state isolation policies controlled by network operators. The packets that trigger DPF execution include the code to execute or a reference to the DPFs deployed in the switch. DPFs are Turing-complete, may maintain state in the packet and in switch virtual memory, may form a dynamic chain, and may steer packets to desired destinations, all while enforcing the operator's policies.
We demonstrate that this idea is practical by prototyping SwitchVM in P4 on Intel Tofino switches. We describe a variety of use cases that SwitchVM supports, and implement three complex applications from prior works -key-value store cache, load balancer and Paxos accelerator. We also show that SwitchVM provides strong performance isolation, zerooverhead runtime programmability, may hold two orders of magnitude more in-switch programs than existing techniques, and may support thousands of concurrent tenants each with its private state.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 83ecdfd6-bf10-491f-8db7-9dbda9181504Cited by top-tier papers3
- P4runpro: Enabling Runtime Programmability for RMT Programmable SwitchesYifan Yang, Lin He, Jiasheng Zhou, Xiaoyi Shi et al.SIGCOMM 2024 · 12 citations
- Enabling In-Network Acceleration Over the CloudHao Wang, Decang Sun, Jinbin Hu, Kai ChenINFOCOM 2025 · 3 citations
- Performance Prediction of On-NIC Network Functions with Multi-Resource Contention and Traffic AwarenessShaofeng Wu, Qiang Su, Zhixiong Niu, Hong XuASPLOS 2025 · 3 citations
Builds on10
- NetLock: Fast, Centralized Lock Management Using Programmable SwitchesZhuolong Yu, Yiwen Zhang, Vladimir Braverman, Mosharaf Chowdhury et al.SIGCOMM 2020 · 60 citations
- RackSched: A Microsecond-Scale Scheduler for Rack-Scale ComputersHang Zhu, Kostis Kaffes, Zixu Chen, Zhenming Liu et al.OSDI 2020 · 58 citations
- Switches for HIRE: resource scheduling for data center in-network computingMarcel Blöcher, Lin Wang, Patrick Eugster, Max SchmidtASPLOS 2021 · 34 citations
- Disaggregating Stateful Network FunctionsDeepak Bansal, Gerald DeGrace, Rishabh Tewari, Michal Zygmunt et al.NSDI 2023 · 33 citations
- NetRPC: Enabling In-Network Computation in Remote Procedure CallsBohan Zhao, Wenfei Wu, Wei XuNSDI 2023 · 23 citations
Related papers
- Unlocking the Power of Inline Floating-Point Operations on Programmable SwitchesYifan Yuan, Omar Alama, Jiawei Fei, Jacob Nelson et al.NSDI 2022 · 33 citations
- Lucid: a language for control in the data planeJohn Sonchack, Devon Loehr, Jennifer Rexford, David WalkerSIGCOMM 2021 · 45 citations
- NetVRM: Virtual Register Memory for Programmable NetworksHang Zhu, Tao Wang, Yi Hong, Dan R. K. Ports et al.NSDI 2022 · 21 citations
- Monic: In-Network Mixture-of-Experts Inference on Programmable Data PlanesXiaoquan Zhang, Bowen Liang, Fung Po Tso, Yuhui Deng et al.INFOCOM 2026
- Flowrest: Practical Flow-Level Inference in Programmable Switches with Random ForestsAristide Tanyi-Jong Akem, Michele Gucciardo, Marco FioreINFOCOM 2023 · 63 citations
