HP Mainframe Connectivity Design Guide

Determining bandwidth requirements
To determine bandwidth requirements in a mainframe environment, use one of the following:
• Vision tool (formerly RMF Magic)—Determines the amount of data that must be replicated
between sites. You can use the results to determine the bandwidth required to support a specific
application. For more information about Vision, contact an HP storage representative.
• Amount of new or changed data—Use the following procedure.
To measure the amount of new or changed data:
1. Determine the peak read and write workloads for a specific time period (use RMF reports).
2. Determine the average write data block size for each application or CHPID. Peak workloads
are typically at the end of a period, such as a week, month, or quarter.
3. Use the number of writes per second plus the write data block size to calculate the write
throughput for specific applications. Do this for each application or CHPID.
Writes per second can be a number or a percentage of the total number of DASD I/O
operations. For example, if you know there are 100 read/write I/O operations per second,
and that 30% of them are write operations, you can calculate that there are 30 writes per
second.
4. Determine the peak write activity (in Mb/s or Gb/s), the average write activity, and how often
the data rate bursts above the average.
• 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 data collection process become your design goal.
5. 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 application you use.
• RTO—Indicates when to start using the recovery site, and provides data about application
failover and restart.
For example, an RPO of 30 seconds and an RTO of 1 hour means that if a disaster occurs at
the primary site, you can lose up to 30 seconds of data that was written, and the applications
will be up and running at the recovery site within 1 hour.
HP P9000/XP Continuous Access Synchronous, HP P9000/XP Continuous Access Journal,
and HP P9000/XP for Compatible Extended Remote Copy (XRC) support an RPO from near
zero to several hours.
6. After you collect the data:
• If the RPO is near zero, use the peak write rate and throughput to estimate the required
bandwidth.
• If the RPO is greater than zero, determine the average change rate over the RPO interval.
You may need to increase or decrease the bandwidth, depending on the environment
and the amount of time required to complete the last write of the day.
NOTE: It is difficult to predict data compression. Therefore, these calculations do not account
for compression. If you determine that you can compress all data at a constant rate, then you
can reduce the required bandwidth accordingly.
142 FICON and FICON SAN extension