Introduction - Reliability Assurance of Big Data in the Cloud - page 4

huge amount of resources consumed by these data, efficient data management could
generate huge value. For managing the massive amounts of Cloud data, the Cloud
storage system needs to be powerful enough and able to meet the diverse needs of the
data of different usages at different stages.
1.2.2 The Cloud data life cycle
The Cloud data life cycle refers to the period of time starting from the data being cre-
ated (generated or uploaded) in the Cloud to the data being deleted when the storage
space is reclaimed by the Cloud storage system. The life cycle of each piece of Cloud
data consists of four stages, which are the data creation stage, the data maintenance
stage, the data recovery stage, and the data deletion stage, as depicted in Figure 1.1 .
1.2.2.1 Data creation
The life cycle of Cloud data starts from the creation of the data in the Cloud storage
system. When the original piece of Cloud data (the original replica for short) is cre-
ated, certain numbers of additional replicas of the Cloud data also need to be created
according to the specific reliability requirement of each piece of data and the storage
policy [8,9]. All these replicas are transferred and stored on specific storage devices
in a distributed fashion.
1.2.2.2 Data maintenance
After the data are created and stored, the data maintenance stage commences, which
occupies the majority of the Cloud data life. At this stage, Cloud data are processed
within applications to achieve different goals. However, for most of the time these
data are just stored in storage devices waiting for later use. Certain mechanisms can
be conducted to maintain all the replicas so that the service quality is not jeopardized.
From the data reliability aspect, the redundancy of Cloud data is maintained at a cer-
tain level, so that sufficient data reliability assurance can be offered to meet the stor-
age user's data reliability requirement.
Figure 1.1 Cloud data life cycle