Information Technology Reference
In-Depth Information
14
11.765
12
on Hadoop
10.703
9.75
on MPI
10
8
5.37
5.365
5.36
5.34
5.5
6
4.968
4
1.828
2
0
4 MB
8 MB
12 MB
16 MB
20 MB
FIGURE 8.19
CPU time spent on Hadoop and MPI.
292
300
on Handoop
250
on single
machine
200
146
150
100
74
50
23
25
24
0
8 hours
16 hours
32 hours
FIGURE 8.20
Running time on Hadoop and a single machine.
framework message passing interface (MPI) program model, CPU time spent
filtering baseline wander is compared with that of Hadoop. In comparison,
a single machine computing overhead is depicted in Figure 8.19.
On Hadoop, with the increase of data, the running time is essentially the
same. On a single machine, the time changes linearly with the amount of data.
For example, with 32 hours of data, the performance is improved 11 times. For
three nodes, the performance is improved by 82%. So, this platform achieves
higher performance than ordinary solutions, as shown in Figure 8.20.
8.5 Conclusion
The HCloud system incorporates multiple advanced technologies, such
as a precise and convenient data acquisition solution and a high-efficiency
data storage and analysis method, for monitoring the health status of users
Search WWH ::
Custom Search