Article Preview
TopIntroduction
Cloud Computing paradigm has emerged as a technology that has provided various computing services to its users. These computing services can be terms of hardware as well as software resources which can be availed in virtual manner. These resources have really helped big as well as small enterprise to migrate their data over the cloud. The migration of data over the cloud has leads to high level of productivity (Chen et. al, 2016) and also the issues related to any of the resource can be reported to cloud expert and can be resolved in stipulated time. Also, the migration of data over the cloud has also leads to on demand approval of cloud services and scalability has also been provided to cloud service consumers in which consumer can expand or shrink his requirement of cloud service as per his usage. Enterprises or companies can take computing resources on lease and can use them and have to pay for it as per the usage only. By this, the investment of buying those resources physically has been reduced incredibly. These Cloud services can be deployed on private, public, hybrid or community cloud as per the requirement of the service (Singh et. al, 2016), (Pei et. al, 2015), (Jararweh et. al 2012).
In the case of large-scale enterprises, where high speed performances take place, sometimes, a system faces a failure which can be due to some execution constraints, addition or removal of multiple resources in the same cloud environment or while increasing the load of the server due to heavy workload (Bilal et.al, 2015), (Malik et. al, 2011). If number of failures occur in cloud environment, then it will affect the reputation value of a particular cloud service provider. Undoubtedly, the cloud services are offered at its best, still they are not highly reliable and can be suffered from multiple failures (Gomeza et. al, 2011). A failure is a state of the system where it stops working efficiently. A system is fault tolerant until and unless it can bear a fault. It will be in inactive stage as soon as it becomes completed faulty. In case of occurrence of a fault, the request of the customer will not be fulfilled in given stipulated time. There can be multiple reason behind the occurrence of the fault such as increasing workload, hacking over the network, network failures, non-availability of the resources, fluctuation problems in power etc.
The tendency of occurrence of fault is more in public cloud as comparable to private cloud. In past times, number of faults have been noted in various public cloud such as Amazon and Google. In the year 2013, Amazon services became inactive for almost one hour which costed around five million US dollars. In the next consecutive year, various services offered by Google such as Calendar, Google Docs, Gmail became stagnant and which lead to huge loss of millions. Next, in 2015, various services of Microsoft Azure such as virtual server became unavailable for different regions over the globe. Each time, a fault is recognised in cloud environment, it costs more than 50,000 US dollars (Infoworld (2016)). Keeping all these stats in mind, a framework or technique is required for reliable and high available cloud services which can avoid fault occurrence in the cloud environment, and even if occurred, can be removed in short span of time.