Abstract:
As computer systems are taking more and more
responsibilities in critical processes, the demand for
storage is increasing due to widespread applications.
Saving the digital information in a large disk is
expensive and unreliable. As a result, if the disk fails
all the data is lost. Therefore, the yearning for a
better understanding of the system’s reliability is ever
increasing. In greatest hit storage environments,
deduplication is applied as an effective technique to
optimize the storage space utilization. Usually, the
data deduplication impacts the bad result for the
reliability of the storage system because of the
information sharing.
In this paper, reliability guaranteed deduplication
algorithm is proposed by considering reliability
during the deduplication process. The deduplicated
data are distributed to the storage pool by applying
the consistent hash ring as a replicas placement
strategy. The proposed mechanism is evaluated and
the result is compared with pure replication and
erasure coded replication. The proposed mechanism
can provide the better storage utilization and the one
hundred percent of assurance for demanded
reliability level in compared with the existing systems.
