!
$
ECG082/0898
What the Benchmarks Don’t Tell You
It is important to understand that benchmarks such as these are designed to give Exchange Server
implementation planners baseline references for understanding the capabilities of hardware
platforms from a single vendor such as Compaq or other competing hardware vendors. When
interpreting these benchmarks, two things should be kept in mind.
First, consider whether benchmarks are performed on what can be referred to as customer-
deployable configurations. A hardware vendor may publish a result that is based on a platform or
configuration that should not be deployed in a real-world Exchange Server deployment. For
example, many vendors (including Compaq) publish results using disk subsystems configured
with RAID0. While RAID0 does provide the highest levels of disk subsystem performance, it
fails to provide any protection against data loss. Compaq recommends deploying an Exchange
Server with disk fault tolerance such as RAID1 or RAID5 for the highest levels of data
protection.
Second, most vendors, including Compaq, conduct benchmarks for Exchange Server that are
single-server in nature. Also, keep in mind that benchmarks do not account for issues such as
backup and disaster recovery or information store maintenance sizing. Whatever the issue, care
must be taken when interpreting benchmarks to ensure that they represent useful information for
your Exchange Server deployment and are based on valid simulation methodologies.
While it is significant that the ProLiant 7000 server can successfully scale to 15,000 medium
MAPI e-mail users (MMB) in a single-server benchmark exercise, Compaq recommends careful
evaluation of all issues involved in real-world Exchange Server deployments – issues such as
management, administration, and disaster recovery.
MAPI Messaging Benchmark (MMB) - LoadSim Medium
User Redefined
To distinguish clearly between throughput benchmarks and capacity planning information for
Microsoft Exchange Server, Microsoft has established the MAPI Messaging Benchmark (MMB)
based on the workload from LoadSim Medium User profile. The MAPI Messaging Benchmark
representative workload focuses on the resulting throughput and clearly communicates the profile
under test.
The workload profile has not changed from the former LoadSim Medium User, but expressed in
clearer fashion. The intent is to make sure that customers can understand the MAPI Messaging
Benchmark workload and can compare it to other platforms. In addition, the renaming of the
benchmark reinforces that the test is a measurement of messaging throughput and that additional
considerations are required in capacity planning.
MMB Transaction Load
The transaction load created by the benchmark is equivalent to the user actions over an eight-hour
day as shown in Table 3.
