Cost-effective data transfer for data creation and data recovery - Reliability Assurance of Big Data in the Cloud - page 75

Table 7.1 Simulation parameters

startTime 0

deadline 80, 800, 8000 seconds

dataSize 1 MB-10 GB (8 Mbits-80 Gbits)

maximumBW 10 Mbps, 100 Mbps, 1 Gbps

routers 10

events 500

linkRates 10 Mbps, 100 Mbps, 1 Gbps

routerPower 0.1, 0.4, 3.6 W

deadline ) pair, dataSize , maximumBW , routers , events , and linkRates . Among these
parameters, maximumBW is set to different values representing different data trans-
fer speeds between different devices. routers indicates the number of routers on the
link, events indicates the accumulated number of events of all the pipeline agendas,
linkRates indicates the link rate that the router can be working at, and routerPower is
set according to the link rate of the router. The routerPower value is obtained based
on the research presented by Zhang et al. [80] . Table 7.1 shows the range of these
parameters in the simulation.
In this simulation, we do not involve the shutdown period as a parameter. By in-
volving it, the result is obvious: The minimum-speed strategy and maximum-speed
strategy would consume more energy by allocating bandwidth during the shutdown
period while LRCDT does not. This only magnifies the proportion of energy con-
sumption reduced by LRCDT.

7.5.2 Energy consumption comparison
Figure 7.9 shows the average energy consumptions of data transfer by applying
LRCDT and the other two strategies.
In Figure 7.9 , each subgraph shows the energy consumptions for transferring data
files ranging from 1 MB to 10 GB with different maximumBW transfer speed upper
boundaries. It can be seen that LRCDT is able to transfer data files with the least en-
ergy consumption for all sizes of data files under all three maximumBW values. Com-
pared to the other two data transfer strategies, LRCDT reduces at least 27.6% of the
energy consumption. Under different maximumBW, however, the energy-saving effect
of LRCDT is different. Specifically, compared to the maximum-speed strategy when
maximumBW is higher, LRCDT is able to reduce more energy. It consumes on aver-
age 37.8% less energy when maximumBW is 10 Mbps compared to 63% less energy
when maximumBW is 1 Gbps. In contrast, compared to the minimum-speed strategy,
LRCDT is able to reduce more energy consumption when maximumBW is lower. On
average, 27.6% energy consumption of data transfer can be saved when maximumBW
is 1 Gbps whereas 33.7% can be saved when maximumBW is 10 Mbps. In addition
to the preceding discussions, Figure 7.9 also shows that the energy-saving effect of
LRCDT gradually decreases when the data size becomes too high. This is because the
maximum data size that can be transferred within the transfer period is limited. When