According to Buyya et al. (2008), “A Cloud is a type of parallel and distributed system consisting of a collection of inter-connected and virtualized computers that are dynamically provisioned and presented as one or more unified computing resources based on service-level agreements established through negotiation between the service provider and consumers.” The enormous advantages of cloud computing in improving businesses have generated unprecedented interest in its adoption. However, customers face numerous cloud computing obstacles such as service requirements; unexpected outages; invalid assumptions about the operating environment; poor isolation between users, hardware degradation, misconfiguration of software; cost implications of failure and uncertainty about cloud providers’ ability to meet service level agreements (SLA). Cloud providers are usually responsible for problems related to their own infrastructure. Although cloud providers monitor their physical resources such as servers, storage and network systems to provide a highly stable infrastructure, they usually do not guarantee individual instance availability.
In addition, Benson and his colleagues analysed the problems users faced and categorized them into five categories: Application-related (e.g. Email server setup, Windows Licensing, LAMP setup, Linux), Virtual Infrastructure-related ( e.g. Virtualized Storage, Attach/Detach, Virtualized load balancer, DNS & Virtualized IP) , Image Management-related ( Image bundling issues, storage and migration of image between buckets, Update/kernel install issues),Performance-related ( Instance not responding, Instance stuck in terminating, EBS performance), and Connectivity-related (General connectivity, Firewall, Connection performance, Connecting to app).
These risks hinder the adoption of cloud. Zardri et al. (2013) believe that “evaluating pre adoption choices at early stages is cost-effective strategy to mitigate risks for probable losses due to wrong or uninformed selection decisions”. They suggested that companies should identify obstacles and their importance through understanding their consequences on the adoption process (obstacle prioritization). Then, appropriate tactics should be used to handle, manage, and solve the problems. Furthermore, analysing Service Level Agreements (SLA) of cloud providers and matching them against users’ requirements can be useful in revealing potential SLA violations, conflicts and probable risks.
Benson (2013) also introduced several strategies that cloud providers can apply to help users.
“Best Effort Support Model: “user forum” is considered as one of the most common version of support models. Unfortunately, cloud providers do not make guarantee on the response time of the operators and the resolution time of the problem.
Premium Support Model: In this model, cloud providers guarantee the user that problems will be resolved within a certain period of time. The SLA provided to a user is inversely proportional to the price paid by the users; a higher price demands a guaranteed for a shorter resolution time. Certain providers also ensure that users have access to dedicated operators who are familiar with the user’s environment and needs.”
As discussed above, it would seem that companies should identify and understand early the properties of the problems that they will face in using a cloud before migrating any servers, databases, applications or data to the cloud. Data security and privacy breaches, as well as regulatory and legal compliance are significant issues when moving from in-house IT infrastructure to cloud services. Therefore, specific standards, education, and appropriate support mechanisms should be designed to solve the problems.
By Mojgan Afshari