Cost-effective data transfer for data creation and data recovery - Reliability Assurance of Big Data in the Cloud - page 67

Figure 7.1 Overall network model
model describes the connecting status between two network devices over the link;
the pipeline agenda model describes the detailed bandwidth usage and schedule of a
pipeline; and the overall agenda model describes the detailed bandwidth usage and
schedule of the entire data transfer link.
7.2.1 Overall network model
Figure 7.1 shows an example of the overall network model. There is one source and
one target at each end of the link, which indicates the start and end of the data transfer.
Each of the two ends of the data transfer link can be a storage device or a subset of the
dedicated network. Between the source and target , there are several routing devices over
the data transfer link. Without losing generality, these routing devices are abstracted as
“router” throughout the topic for ease of description. We assume that these routers have
the capacity of changing link rates. Each router has one input port and one output port
connected to the link. The connection between the output port of a router (or source ) and
the input port of the subsequent router (or target ) forms a “pipeline” in which the link
rate and available bandwidth vary over time. As shown in Figure 7.1 , the boxes indicate
the link rate of the pipeline while the dark parts indicate the available bandwidth.
7.2.2 Pipeline model
Figure 7.2 shows the pipeline model between two routers. The status of a pipeline can
be described as a set ( LR, availableBW, t ), in which LR is the link rate, availableBW is
the available bandwidth, and t is the time. For bandwidth reservation purposes, routers
at both ends of the pipeline record and maintain the pipeline status. Each of the records
is called a “pipeline agenda.”
Figure 7.2 Pipeline model