Software Automatic Configuration Model And Methods In Cloud

With the development of cloud computing, an increasing number of basicsoftware applications for calculating, storing and more are migrated to cloud platform.And more and more software applications are set up on cloud platform or rely on theservices provided by cloud platform to support its calculating. With the help ofdeveloping virtual technology, the running of software are no longer limited byparticular hardware environment. People can get same software experience ondifferent terminals. Traditional design of software structure need to consider thehardware environment where the software runs, the hardware environment of theclients where software service is received and limited hardware conditions. Therefore,it’s hard to try the high expandability of it. When the hardware conditions areimproved, traditional softwares usually need to be reconstructed to adaptedthemselves to the new hardware and make full use of it. With cloud computingtechnology, we can consider higher expandability, more reliable service and morecomfortable software experience of software structure design, regardless thelimitation of hardware environment. For the reasons above, we hope present asoftware automatic configuration model in the cloud. Cloud platform relies onvirtualization technology and distributed computing technology. Nowadays, mostrelated researches focuses on these two technologies. However, we notice that cloudplatform itself has provided a good environment for researches of distributedcomputing technology so we tried a research of distributed system structure on cloudplatform.This essay defines a general model of software firstly. a general softwareautomatic configuration model are represented by us on the base of this model. Withthis model, we can focus more on the implementation of the services of this particularfield and the improvement pf software quality, ignoring the environment andresources needed for its running. We referenced some designs of main strain PaaSplatform, such as Google App Engine、Heroku、Amazon S2, etc. Google App Engineand Heroku are merely platform services for web applications; Amazon S2providesplatform services for general softwares, but the software configuration andmanagement of the running environment have to be done by users themselves, whichcosts a lot. None of these models mentioned above is our ideal model. What we wantis a kind of plug-in based model, which abstracts the structure model by OVF expansion and has expandability. We describe the whole life cycle and automaticconfiguration of the software of this model in detail.What’s more, we hope our model has the general characteristics of distributedsoftware and each of its components has expandability. In cloud, the virtualization ofhardware resources helps the running environment be automatically set up with apredesigned configuration file. When one of the components needs more calculatingsupport and one nor more cannot answer the user requires, we let one component tomultiply and become a group of components which has same function and canreasonably deal the calculating request through an intermediate organizer. In this part,we analyze the existing load-balancing algorithm and propose an improved algorithmsuitable for our model.Finally, we make a simulation which based on Eucalyptus, an IaaS platform, tovalidate a general software automatic configuration model, GAICM, in the cloud, andprove the automatic extensibility of the software components.