Review on Reliability and Energy-Efficiency Issues in Cloud Computing

Classification Of Failures

Figure 1.

Classification of failures

The occurrence-based classification is related to the sequence among the failures i.e whether the occurrence of one failure leads to the occurrence of another or not in the system. Occurrence-based failures are again categorized into independent and correlated failures. Independent failures occur discretely. This type of occurrence is hypothetical because the literature has demonstrated that there is a correlation between failures (Fu & Xu, 2007) [4] (Yigitbasi et al., 2010) [6]. In correlated failures, the occurrence of a subsequent failure is correlated to each other. The correlated failures could be a spatial correlation and temporal correlation.

The failures are further separated into two types in the architecture-based classification: Resource Failure and Service Failure. Resource failure, as the name implies, is caused by the loss of certain physical resources, such as a system failure, a network or power outage, or a software error. The majority of the literature on failure tolerance has focused on resource failures (Fu, 2010) [8] (Vishwanath & Nagappan, 2010).

Causes Of Failures

It is important to identify the causes of failures in Cloud Computing Services (CCS) to make them more reliable and available at all times. Various causes of cloud computing problems are depicted in Fig. 2.

Figure 2.

Classification of causes of failure