A computing-aware routing protocol for Computing Force Network

Abstract

Facing the development trend of computing and network deep convergence, this paper proposes a computing-aware routing protocol(CARP), which introduces computing information as a cost into the network. CARP provides the capability to jointly schedule service requests to optimal service endpoints along network path, and guarantee the quality of service (QoS) for end users. Moreover, in order to make the network aware of the computing resource information, we develop the computing resource discovery and advertisement scheme, which achieves a better trade-off between network resource consumed by routing path and signaling load for computing resource information updates. Finally, The test results show that under the ideal network load, the system capacity of the proposed system can be increased by 33.17% compared with the existing MEC system, and the end-to-end average latency can be increased by 35.29%. Therefore, the performance analyses show that our proposed scheme can substantially achieve the joint optimization of computing and networking resources through a tighter integration of computing and networking.