Cloud Computing Distributed Internet Computing for IT and Scientific Research

One vision of 21st century computingis that users will accessInternet services over lightweightportable devices rather than throughsome descendant of the traditionaldesktop PC. Because users won’t have(or be interested in) powerful machines,who will supply the computing power?The answer to this question lies withcloud computing.Cloud computing is a recent trendin IT that moves computing and dataaway from desktop and portable PCsinto large data centers. It refers toapplications delivered as services overthe Internet as well as to the actualcloud infrastructure — namely, thehardware and systems software in datacenters that provide these services.The key driving forces behind cloudcomputing are the ubiquity of broadbandand wireless networking, fallingstorage costs, and progressive improvementsin Internet computing software.Cloud-service clients will be able toadd more capacity at peak demand,reduce costs, experiment with new services,and remove unneeded capacity,whereas service providers will increaseutilization via multiplexing, and allowfor larger investments in software andhardware.Currently, the main technical underpinningsof cloud computing infrastructuresand services include virtualization,service-oriented software, grid computingtechnologies, management of largefacilities, and power efficiency. Consumerspurchase such services in the formof infrastructure-as-a-service (IaaS),platform-as-a-service (PaaS), or software-as-a-service(SaaS) and sell valueaddedservices (such as utility services)to users. Within the cloud, the laws ofprobability give service providers greatleverage through statistical multiplexingof varying workloads and easier management— a single software installationcan cover many users’ needs.We can distinguish two differentarchitectural models for clouds:the first one is designed to scale outby providing additional computinginstances on demand. Clouds can use these instances to supply services in the formof SaaS and PaaS. The second architecturalmodel is designed to provide data and computeintensiveapplications via scaling capacity. Inmost cases, clouds provide on-demand computinginstances or capacities with a “pay-as-yougo”economic model. The cloud infrastructurecan support any computing model compatiblewith loosely coupled CPU clusters. Organizationscan provide hardware for clouds internally(internal clouds), or a third party can provideit externally (hosted clouds). A cloud might berestricted to a single organization or group (privateclouds), available to the general public overthe Internet (public clouds), or shared by multiplegroups or organizations (hybrid clouds).A cloud comprises processing, network, andstorage elements, and cloud architecture consistsof three abstract layers. Infrastructure isthe lowest layer and is a means of deliveringbasic storage and compute capabilities as standardizedservices over the network. Servers,storage systems, switches, routers, and othersystems handle specific types of workloads,from batch processing to server or storageaugmentation during peak loads. The middleplatform layer provides higher abstractionsand services to develop, test, deploy, host, andmaintain applications in the same integrateddevelopment environment. The applicationlayer is the highest layer and features a completeapplication offered as a service.