Conclusions and future work - Reliability Assurance of Big Data in the Cloud - page 79

Conclusions and future work

8

In this chapter, we summarize the whole topic. The structure of this chapter is orga-
nized as follows. In Section 8.1 , we overview the contents of this topic. In Section 8.2 ,
we summarize the key contributions of this topic. In Section 8.3 , we present further
discussions to the research and our future work.

8.1 Summary of this topic

The content of this topic was presented in the following order.
• In Chapter 1 , we introduced the background knowledge for our research, which is the defi-
nition of data reliability, current development of data reliability in the Cloud, distinctive
features of Cloud storage, and the life cycle of Cloud data. Afterward, we outlined the key
issues of this research and presented a high-level overview of the whole topic.
• In Chapter 2 , we intensively reviewed the literature on existing technologies related to the
research. First, we presented our review for existing hardware reliability theories, in which
disk reliability theories were reviewed specifically. Second, we presented our review for ex-
isting software-based data reliability assurance approaches, in which reviews for replication-
based approaches and erasure coding-based data storage approaches were presented respec-
tively. Third, we presented our review for existing data transfer approaches in distributed
systems.
• In Chapter 3 , we presented a motivating example with the analysis of our major research
issues. First, we presented the details of the pulsar searching scientific application as the
motivating example of our research. Second, we presented our analysis on the problem of
cost-effective big data storage in the Cloud with data reliability assurance in detail, in which
major factors of Cloud storage cost, data storage devices and schemes, Cloud network for
data transfer during data creation and data recovery are addressed. Third, based on the analy-
sis, we presented the details of our research issues.
• In Chapter 4 , we presented our novel generic replication-based data reliability model in
detail for describing the Cloud data reliability. First, despite previous investigations, we
determined further details of the model, including reliability metrics, presentation type, and
failure rate pattern. Second, according to the reasoning process, we demonstrated the data
reliability model in detail.
• In Chapter 5 , based on the generic data reliability model, we presented our approach for
calculating the minimum replication for meeting the data reliability requirement. First, we
presented related formulas of the calculation approach as well as optimization of the for-
mulas for reducing the computation complexity. Second, we discussed the application of
the minimum replication as a benchmark for evaluation of replication-based data storage
approaches. Third, we briefly summarized the satisfactory outcomes of the evaluation for the
minimum replication calculation approach.
• In Chapter 6 , we presented our novel cost-effective data reliability assurance mechanism
Proactive Replica Checking for Reliability (PRCR) for maintaining big data in the Cloud