2021/09/03 by Murugan Sankaradas, Sankaradas, Murugan, Kunal Rao +3
Computer Science · Social Sciences · #Advanced Computing and Algorithms #FOS: Computer and information sciences #IoT and Edge/Fog Computing #Networking and Internet Architecture (cs.NI) #Software-Defined Networks and 5G
paper · pdf · doi:10.48550/arxiv.2109.01737
openalex publication_date 2021/09/03 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Applications that use edge computing and 5G to improve response times consume\nboth compute and network resources. However, 5G networks manage only network\nresources without considering the application's compute requirements, and\ncontainer orchestration frameworks manage only compute resources without\nconsidering the application's network requirements. We observe that there is a\ncomplex coupling between an application's compute and network usage, which can\nbe leveraged to improve application performance and resource utilization. We\npropose a new, declarative abstraction called AppSlice that jointly considers\nthe application's compute and network requirements. This abstraction leverages\ncontainer management systems to manage edge computing resources, and 5G network\nstacks to manage network resources, while the joint consideration of coupling\nbetween compute and network usage is explicitly managed by a new runtime\nsystem, which delivers the declarative semantics of the app slice. The runtime\nsystem also jointly manages the edge compute and network resource usage\nautomatically across different edge computing environments and 5G networks by\nusing two adaptive algorithms. We implement a complex, real-world, real-time\nmonitoring application using the proposed app slice abstraction, and\ndemonstrate on a private 5G/LTE testbed that the proposed runtime system\nsignificantly improves the application performance and resource usage when\ncompared with the case where the coupling between the compute and network\nresource usage is ignored.\n