Above the Clouds: A Berkeley View of Cloud Computing
Michael Armbrust, Armando Fox, Rean Griffith, Anthony D. Joseph, Randy H. Katz, Andrew Konwinski, Gunho Lee, David A. Patterson, Ariel Rabkin, Ion Stoica and Matei Zaharia. Above the Clouds: A Berkeley View of Cloud Computing. EECS Department University of California, Berkeley Technical Report No. UCB/EECS-2009-28 February 10, 2009
No review due
Comments
The paper explains cloud computing in detail from its origin to its present status and gives its visualization from as any aspects as possible.
The concept of cloud computing by its nomenclature got a mixture of arguments. Most of scholars agree that there is not a single definition of cloud. The term itself received severe critical remarks from FSF and other non profit software promoters. Some of them with a hardware image says it’s a rejection of mainframe and grid computing. Most constructively, it is an enthusiastic effort to make resource elasticity in distributed systems. Earlier developments led applications as services over network(SaaS). It was an overcome from software as Bits with advancement of technology. Though SaaS had its own benefits like ease in maintenance and controlled usage, but recent developments experimented on combining Hardware, Infrastructure and Platform as a Service. But the latest event of cloud computing as a utility computing tells us to pay in proportion to our growth with illusion of infinite resources, with no upfront costs and fin grained billings. The main motivational force for drift towards cloud may be mitigating the risks of over-provisioning and under provisioning, less maintenance and operational costs, saving time as well as funds. The illusion of infinite computing resources avoids far head provisioning. Up front commitment elimination helps us to start small and grow with requirements. The pay by demand helps to reduce costs as we utilize resources when needed and pay accordingly. A comprehensive cloud computing solution provides mobile and web applications, extensibility of desktop software, batch processing and many others. Public and private cloud may differ on the ground of usability as public cloud is more subject to utility computing while private might do it on a lesser extent and may restrict is services for rental purpose. Cloud computing has overcome the limitations of P2P as peers do not get paid and have lower reliability. Cloud is also not subject to as tight commitment as is required by grid computing.
The dominant advantages of cloud are relaxation from under and over prediction for scalability and cost reduction. Cloud also provides abstraction of SaaS and utility computing. It has led to the development of trend of processing by hours and storage by days. The authors supports philosophically that challenges faced by modern day cloud can be treated as opportunities for the future.
The universal adoption of cloud is subject to various restrictions as are addressed as challenges by authors. There are still questions on the inclusion of internal datacenters in the cloud. Also it’s too broad to include everything we already have in cloud, which ultimately means we have everything as part of cloud else nothing is cloud and the term is vague here. The hardware requirement for cloud too may be different and needs extensive exploration. United States legal system may also act as a hindrance to cloud computing too. Security is a major issue. The primary security mechanism of virtualization adopted by cloud is too subject to attacks. Besides it cloud’s user protection against cloud provider is an important issue. Before making any mind for cloud, the security issued should be analyzed on the basis of individual service’s privacy too. User level encryption need to be promoted. Also, the question “Who should adopt cloud computing?” is to be answered at least with some acceptable benchmarks.
Posted by: y,Kv | April 14, 2010 02:50 PM