Introduction:
Cloud computing is the latest buzzword in the I.T. industry. In concept, it is a paradigm shift whereby details are abstracted from the users who no longer need knowledge of, expertise in, or control over the technology infrastructure in the cloud that supports them.
It typically involves the provision of dynamically scalable and often virtualized resources as a service over the internet. These services are broadly divided into three categories: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service
(SaaS).
Fault tolerance:
Fault tolerance is an ability of system to continue operation in occurrence of failure of part of the system rather than failing completely. Occurrence of failure might restart the whole system. This partial failure may result in reduction in throughput or increase in response time and leaves the system barren for considerable amount of time. The major aspect of fault tolerance is to maintain the integrity of the system in any damage.
A system can be made fault tolerant by:
Replication: providing identical instances of same systems or subsystems. Thus by directing task to all of them in parallel a correct result can be chosen from the quorum.
Redundancy: Providing identical instances of same system or subsystems and then switching over to other in case of failure of one system.
Diversity: Providing multiple different implementations of the same specification, and using them like replicated systems to cope with errors in a specific implementation.
Load balancing:
Load balancing is an act of distributing work so as to achieve optimal resource utilization, increase throughput, reduce response time and avoid overload. It is important for networks where it is difficult to predict the number of request that will be issued to the system. If one system starts to get swamped, requests are forwarded to another system with more capacity. Because load balancing distributes the request based on the actual load at each node, it is excellent for ensuring availability and defending against denial of service attacks.