.SAN design reference guide Vol. 1-5 785350-001

Determining the required bandwidth
You can determine the required bandwidth for any application. This example explains how to
measure the amount of new or changed data:
1. Collect the peak read and write workloads for a given period of time. For Windows operating
systems, use a tool such as PERFMON to capture the current performance requirements while
HP P6000 Continuous Access is not running. Similar tools exist for other operating systems.
• At each sample interval, capture reads per second (I/Os per second), read throughput
per second (Mb/s), writes per second (I/Os per second), and write throughput per second
(Mb/s).
• If possible, collect read and write latency data.
• Perform the collection by application, capturing the data for each logical unit (device)
used by that application.
2. Create a graph of each data set that shows where the peaks occur during the day.
• Determine whether the daily average change rate is level or bursty.
• Consider how these numbers will increase over the next 12 to 18 months.
The results of this scaling process become your design goal.
• Determine the values for RPO and RTO:
◦ RPO measures how much data is lost due to a problem at the source site. By definition,
an RPO of zero (no data can be lost) requires synchronous replication, regardless
of which data replication product you use.
◦ RTO indicates when to start using the recovery site. This measurement includes data
about application failover and restart.
◦ For asynchronous HP P6000 Continuous Access, the RPO design space is near zero.
◦ HP P9000 (XP) Continuous Access Asynchronous supports an RPO from near zero
to many hours.
◦ HP P6000 Continuous Access and HP P9000 (XP) Continuous Access all synchronous
replication with an RTO equal to zero.
3. After the data has been collected:
• If the RPO is near zero, use the peak write rate and throughput to estimate the bandwidth
you need. For some real-time applications (such as Microsoft Exchange), increase the
bandwidth between 2 to 10 times this initial estimate due to wait time for link access.
• If the RPO is greater than zero, then average the change rate over the RPO interval and
use this value as an estimate of the inter-site bandwidth. You might need to increase or
decrease this bandwidth, depending on the environment and the amount of time needed
to complete the last write of the day before starting the next day's work.
NOTE: Because it is difficult to predict data compression before you begin, these
calculations do not account for the impact of compression. If you determine that you can
compress all data at a constant rate, then use that ratio to reduce the effective throughput
required from the link.
FCIP gateways
Table 136 (page 270) lists the HP FCIP gateways and supported switches.
Multi-protocol long-distance technology 269