Research On The Architecture For The Next-generation Highly Efficient Cloud Computing Paradigm

Cloud computing has been in fast growing in this several years and has been the fundamental of IT industry,making great changes and impact.More and more services are being provided via cloud computing and the technology involved in cloud computing is developing fast.Generally,the fundunmental infrastrature of cloud computing is large scale data center.Cloud computing manufacturers make a great resource pool by installing lots of hardwares and proper software,and provide services according the requirement of customers.This kind of traditional cloud computing paradigm can serve the IT industry well in current condition.But with the development of IT industry,data are gaining in a burst mode and the requirement of big data and Internet of things makes great challenges on traditional cloud computing.The current centralized cloud computing paradigm will face the problem of lack of resources and network delay when dealing with massive data processing requirement.Simply scaling the cloud data center can not solve the problem and may lead to more serious problems,such as energy consumption and network bandwidth occupation.Thus new development direction of cloud computing needs to be studied for next-generation cloud computing.This paper researches on future cloud computing trends.We first summary the trends of related industries and analyse the challenges cloud computing facing,then we make deep research on two important trends.The first trend we study is containerized data center.We propose a hierarchical management strategy of containerized data center.Then we simulate a small scale data center do some test.The experiment results show that our startegy gets better performance and can save energy.The second trend we study is how to provide edge computing services for future smart home in the fog computing era.We propose a solution by using PCs.And after analyzing the feasibility and challenges,we build a prototype system and evaluate it by choosing some applications in smart home to evaluate single process performance,system scalability and energy consumption.The results show that our desgin has better scalability and can save lots of energy with proper energy control algorithm.