Mining And Application Of Big Performance Data In Cloud Computing

In contemporary cloud computing platforms,thousands of loads are running at the same time,and these complex workloads come from different users,sharing the basic resources in the cloud computing platform.Different types of loads often have different requirements for basic shared resources.Therefore,in cloud computing,different types of loads running on the same server often have performance interference,but commercial server systems have no solution to performance interference.With the advent of 5G technology,various new applications have begun to emerge,which has exacerbated the problem of performance interference in the data center.How to effectively solve the performance isolation problem has become a key challenge in cloud computing.The server in the cloud computing generates a large amount of software and hardware performance data at runtime.These performance data records server micro-system events and software performance events of different processes,which contain many detailed and deep performance information.This paper mines and analyzes various software and hardware performance data generated in the server,and proposes load performance indicators based on hardware and software performance data,load classification method based on software and hardware performance data,and designing performance isolation method based on the above.Performance isolation method could reduce performance interference.In addition,based on the hardware and software performance data,we designs a load monitoring method,which can be used to monitor the performance interference phenomenon in cloud computing and locate the antagonist in time.Through the mining and application of hardware and software performance data,this paper proposes two effective solutions to deal with performance interference.