Design and Implementation of Computing Based Service Chain Orchestration Framework

The computing power network intends to accomplish on-demand allocation and flexible scheduling of network computing resources as well as the unification of the administration of multiple computing resources. Two significant obstacles stand in the way of its progress, nevertheless. First off, heterogeneous computing resources like CPU, GPU, and FPGA lack a consistent measure. Second, service function chains created by breaking down individual large-scale computing activities lack an effective orchestration method. This paper designs and implements a computing based service chain framework to effectively utilize diverse computing resources in the network and deliver efficient computing services. The framework has the ability to efficiently handle user computing tasks, assess user computing demands, and standardize the evaluation of computing resources. In addition, we propose a Stackelberg method to deal with the price of resources and task distribution issues between computing service users and providers. When the user scale is less than 30, the algorithm finishes the computation in one minute, showing its effectiveness in handling tasks using the computing service chain. We confirm its efficacy in lowering user costs and raising service provider revenue through a comparison analysis with single-node processing of the same jobs.