Database Reference
In-Depth Information
IOMeter
—
http://www.iometer.org/
.
Dell DVD Store
—
https://github.com/dvdstore
.
I/O Analyzer
—
http://labs.vmware.com/flings/io-analyzer
. I/O Analyzer is a
great fling. It's a vApp, has multiple distribution points, uses IOMeter, and
integrates with ESXTOP metrics. (See
vmware-io-analyzer-video/
for a video of how it works.)
Comparing Baselines of Different Processor Types and Generations
The infrastructure metrics covered in
Table 10.1
are standard metrics that are
collectable across all systems, even if they are a completely different system
architecture. If your source system is Itanium based, although the metrics themselves are
the same, they will not all directly translate to your destination x86 systems. The same is
true if you are comparing across system generations. The CPU performance in particular
is one main area where you will have to do some adjustments when baselining and
comparing systems based on different CPUs, including Intel Itanium-based Windows
and SQL Server environments to x86-based SQL Server virtual machines.
As described briefly in the previous section on vendor benchmarks, the SAP SD 2 Tier
Benchmark is a standard, repeatable, publicly published benchmark that can be used for
comparing different systems quickly. It is an OLTP-type benchmark, and the results are
available to view and download at
http://www.sap.com/solutions/benchmark/sd2tier.epx
. The results available include
systems covering Intel Itanium, Intel Xeon, and AMD processor types, across multiple
CPU generations and multiple versions of SQL Server.
Using the SAP SD Benchmark allows you to figure out what system processor
utilization on your source system during baselining is likely to equate to in terms of
processor utilization on your destination system during your performance testing,
baselining, or benchmarking. This is very useful when comparing dissimilar source and
destination processor generations or processor types. So if you're going from Intel
Itanium on your physical system to Intel Xeon on your destination virtual machine, or
switching from AMD Opteron to Intel Xeon, or vice versa, this gives you a quick and
easy way of comparing CPU utilization.
The main unit of measure provided by the SD Benchmark is SAPS. By comparing the
SAPS between different processor types, you can get an idea of the relative
performance differences between those processors. The next section gives you some
examples.
Comparing Different System Processor Types
