Research And Application Of Fine Granular Application Container Model For Cloud Computing

The approaches to software development continue to evolve with the rapid development of cloud computing.Various service platforms have become to be software products composited with a set of fine granular service applications in loosely-coupled model by the pattern of microarchitecture and micro service.It is the main feature to develop a software product.On the one hand,it fundamentally changed the pattern of development,deployment and scheduling of an application;on the other hand,the complex applications and the diversity resource impacted the needs of resource allocation and scheduling in the Iaas for the development and scheduling of an application.Research on parallel processing with container technology based on lightweight virtualization is an important and inevitable trend research topic in cloud computing.In order to use of computing resources more flexible,finer and more fully,to realize the application virtualization under the cloud computing infrastructure,and to deploy,schedule and execute the user's various types of application services adaptively in a heterogeneous cloud data center,a novel fine granular application container parallel processing framework based on the technology of application virtualization was presented.The main contribution and innovation goes as follows:First,it presented to use the technology of application visualization to build an Application as a Service layer between the layers of PaaS and IaaS,to realize the application management platform with multi-core aware parallel scheduling model.To implement the collaborative work between the logical layers and meet the needs of tasks scheduling,it designed a fine Granular Application Container(GAC)parallel processing framework based on the lightweight virtualization container technology.We analyzed the logical architecture of the framework,the description of the resources set and the parallel processing mechanism,and the results of the experiments validated the feasibility and effectiveness of the model of GAC.Various complex and diverse functions can be combined by application services with different granularity,and it further can be handled in the multi-core aware application management platform.On the one hand,the platform based on GAC will improve the flexibility of the system and its industry adaptability;on the other hand,it can also allocate the resources with fine granularity to improve the utilization of the resources.To resolve the problems on description the architecture of the fine granular application container model and its basic mapping relations,and the problems on definition the processing of assembly and reuse with using different granularity application services,a formal description approach was proposed on the basis of type category theory.It laid a preliminary foundation for the theoretical formation of the formal description framework and the specification definition of the fine granular application model,which is used to support to build a cloud computing service platform by using fine granular applications.The formal framework and the specification definition is used to definite the business logic and the access interface of a services for the system.Third,it presented a performance analysis model of heterogeneous data centers in cloud computing.Based on the model of the fine granular application container,we investigate the heterogeneity of modern data centers and the service process used in these heterogeneous data centers.Using queuing theory,we construct a complex queuing model composed of two concatenated queuing systems and present this as an analytical model for evaluating the performance of heterogeneous datacenters.Based on this complex queuing model,we analyze the mean response time,the mean waiting time,and other important performance indicators.We also conduct simulation experiments to confirm the validity of the complex queuing model.We further conduct numerical experiments to demonstrate that the traffic intensity(or utilization)of each execution server,as well as the configuration of server clusters,in a heterogeneous data center will impact the performance of the system.Our results indicate that our analytical model is effective in accurately estimating the performance of the heterogeneous data center.Forth,it provided a novel multi-server control model with dynamic feedback to acquire dynamic states of the cloud system,and with queue waiting cost-awareness to optimize the queue wait time and load distribution in task assignment and server configuration management.By using the technology of speed scaling,each server in a data center is configured as an M/M/1 queue system with variable service rate,and the service rate is a function of the length of the task queue.We formulate two optimization problems,the optimal load distribution problem and the optimal controlling service rate problem,and provide algorithms to solve these problems,facilitating load distributions and service rate adjustments.We also present numerical simulations to validate our model.The results show our model to be efficient in multi-server dynamic configurations and task assignments according to the feedback information for the tradeoff between the system cost and performance.Fifth,a novel multi-resource scheduling strategy named Multi-dimensional Quantify Polygon algorithm(MQP)based on the quantify polygon model of resources and the theory of quantify polygon complementarity scheduling was presented to allocate multi-resource for the requirement of the fine granular applications.With the theoretical of quantify polygon model of resources and the multilateral complementarity strategy,it analyzed the fundamental principle of the quantify polygon complementarity scheduling and the application architecture.The MQP algorithm can be used to allocate multi-resources from 3D to 6D for application services deployment in the scenarios of static,dynamic and balanced tasks deployment.The results of the simulation experiments confirmed that it can minimize the nodes about 2%-5% better than the other strategies(Eg.FFDProd,L2,DP et al.)for improving the utilization of the resources and saving energy and operation costs.