CHISEL: An optical slice of the wide-area network
Abhishek Vijaya Kumar, Bill Owens, Nikolaj S. Bjørner, Binbin Guan, Yawei Yin, Paramvir Bahl, Rachee Singh
Abstract
Network slicing reserves a portion of the physical resources of radio access networks and makes them available to consumers. Slices guarantee traffic isolation, strict bandwidth and quality of service. However, the abstraction of slicing has been limited to access networks. We develop CHISEL, a system that dynamically carves slices of the wide-area network (WAN), enabling an end-to-end network slicing abstraction. CHISEL creates optical slices between WAN endpoints to avoid queueing and congestion delays inherent in packet switched paths in WANs. CHISEL incrementally allocates optical spectrum on long-haul fiber to provision slices. This task is made challenging by the co-existence of data-carrying channels on the fiber and numerous physical constraints associated with provisioning optical paths e.g., spectrum contiguity, continuity and optical reach constraints. CHISEL leverages the empirical finding that cloud WANs have abundant optical spectrum to spare -75% of optical spectrum on 75% of fiber spans is unused. CHISEL can optimally allocate terabits of slice requests while consuming minimal optical spectrum within seconds without increasing spectral fragmentation on fiber. CHISEL trades-off optimality of slice bandwidth allocation for faster run-time, provisioning slices within 2% of optimal in less than 30 seconds in a commercial cloud WAN. Finally, CHISEL reduces the latency of provisioning optical slices on hardware by 10X. Compared to IP tunnels of equivalent capacity, CHISEL consumes 3.3X fewer router ports. 1
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 cae9814d-0b75-438e-845e-cd7c98667b20Cited by top-tier papers2
- HEDGE: Traffic Engineering with Probabilistic Link CapacitiesArjun Devraj, Bill Owens, Umesh Krishnaswamy, Ying Zhang et al.NSDI 2026
- Learning to Tune Optical WANs: A Field Deployment of Noise Models in Optical NetworksBhaskar Kataria, Howard Hua, Andrea D'Amico, Bill Owens et al.NSDI 2026
Builds on6
- Cost-effective Cloud Edge Traffic Engineering with CascaraRachee Singh, Sharad Agarwal, Matt Calder, Paramvir BahlNSDI 2021 · 79 citations
- ARROW: restoration-aware traffic engineeringZhizhen Zhong, Manya Ghobadi, Alaa Khaddaj, Jonathan Leach et al.SIGCOMM 2021 · 37 citations
- OneWAN is better than two: Unifying a split WAN architectureUmesh Krishnaswamy, Rachee Singh, Paul Mattes, Paul-Andre C. Bissonnette et al.NSDI 2023 · 34 citations
- Cost-effective capacity provisioning in wide area networks with ShooflyRachee Singh, Nikolaj S. Bjørner, Sharon Shoham, Yawei Yin et al.SIGCOMM 2021 · 32 citations
- Toward formally verifying congestion control behaviorVenkat Arun, Mina Tahmasbi Arashloo, Ahmed Saeed, Mohammad Alizadeh et al.SIGCOMM 2021 · 31 citations
Related papers
- FSA: fronthaul slicing architecture for 5G using dataplane programmable switchesNishant Budhdev, Raj Joshi, Pravein Govindan Kannan, Mun Choon Chan et al.MobiCom 2021 · 16 citations
- AZTEC: Anticipatory Capacity Allocation for Zero-Touch Network SlicingDario Bega, Marco Gramaglia, Marco Fiore, Albert Banchs et al.INFOCOM 2020 · 71 citations
- FlexWAN: Software Hardware Co-design for Cost-Effective and Resilient Optical BackbonesCongcong Miao, Zhizhen Zhong, Ying Zhang, Kunling He et al.SIGCOMM 2023 · 10 citations
- Channel-Aware 5G RAN Slicing with Customizable SchedulersYongzhou Chen, Ruihao Yao, Haitham Hassanieh, Radhika MittalNSDI 2023 · 52 citations
- Scalable Real-Time Bandwidth Fairness in SwitchesRobert MacDavid, Xiaoqi Chen, Jennifer RexfordINFOCOM 2023 · 15 citations
