Research On Load Balancing Scheduling For BaaS

Baa S(Blockchain as a Service)is a fusion product of cloud computing and blockchain technology.It is based on the unique management advantages of cloud platforms,abstract resource pools and highly scalable capabilities.Blockchain users and developers provide convenient and fast application environments.With the significant growth of the Baa S cluster size and the number of users,how to carry high-load services under a complex network topology has become a major difficulty in load balancing in Baa S.In order to improve the resource load balancing capability,enhance the quality of service,and achieve high availability of Baa S,take the core project Fabric of Hyperledger deployed on the Kubernetes platform as an example,and design a more targeted scheduling algorithm without changing the basic structure of Fabric.Baa S and cloud computing related technologies are mainly discuccsed in this paper,the main innovations are reflected in the following two aspects: 1)Two static scheduling algorithms for Baa S channels are proposed to improve resource load balancing.The first is the channel-oriented general scheduling algorithm to enhance disaster tolerance,which has the characteristics of high cohesion and low coupling and easy deployment;the second is the NJW spectral clustering algorithm based on weighted average k-means to orchestrate the peers in the Fabric The static scheduling of the system not only improves the convergence level,but also does not cause the situation that a very small number of peers are divided into one category alone,which enhances the usability of the platform.2)For dynamic deployment scenarios,referring to the system architecture of the platform,it is proposed to use a dynamic warning module equipped with an improved least squares-based structure function method to calculate the fractal dimension and make resource predictions.Baa S is an extension of blockchain technology in the field of cloud computing.Although the essential attributes of decentralization,trust,and non-tampering are still its core competitiveness after "cloudification," cloud computing has given it a stronger usability and wider application space are also worthy of attention.According to this,this article completes the following three aspects of work:(1)Briefly analyze the origin and development status of blockchain technology,in-depth discussion of Fabric's architecture,functional modules,transaction flow and other system elements,and on the basis of it,it proposes Kubernetes-based Optimization plan,and completed the analysis of Baa S load balancing needs.(2)Build a multi-machine environment based on Fabric 1.0.5,and use Hyperledger Caliper's test framework to first test the platform's channel-oriented,endorsement strategy,and chain code scheduling schemes.Secondly,build a deployment scenario,conduct a comparative test on the static scheduling algorithm proposed in this paper,verify the feasibility and have high load balancing performance.(3)Initiate short-term high-load services to Baa S where the dynamic warning module is deployed.The results show that the deployment of the dynamic warning module in this scenario can reduce the user's request waiting time without affecting the throughput,and further improve the service quality.