Table of Contents
Advertisement
Applications on
distributed servers
10% average CPU
App
1
74 tps
App
40
74 tps
Figure 4. Server consolidation example
When you have established the baseline of 74 transactions per second for the
distributed servers, define an equal number of z/VM virtual servers in which to
run the applications.
To assess the system capacity required to support the same volume of work, tune
the workload driver so that each instance of the application running in a virtual
server produces the same transaction rate as its distributed counterpart.
The previous example showed an even distribution of work activity. However, the
vast majority of real-world workloads skew the distribution of work. At any given
moment, some applications are active while others are less active or idle. Unless
your workload is evenly distributed, consider skewing the workload distribution
as part of your capacity assessment.
Same applications running in virtual servers on z/VM
App
App
App
App
1
2
3
4
74
74
74
74
tps
tps
tps
tps
App
App
5
36
74
74
tps
tps
z/VM
LPAR
Chapter 2. Planning for Linux virtual servers
App
App
App
App
37
38
39
40
74
74
74
74
tps
tps
tps
tps
23
Advertisement
Table of Contents
