Sizing; Sizing When Using Dedicated Processors - IBM 170 Servers Manual

Intel power systems operating systems
Hide thumbs Also See for 170 Servers:
Table of Contents

Advertisement

14.6.3 Sizing

Sizing methodology is based on the observation that processor time required to perform an I/O
on the IBM i operating system Virtual SCSI server is fairly constant for a given I/O size. The
I/O devices supported by the Virtual SCSI server are sufficiently similar to provide good
recommendations. These numbers are measured at the physical processor.
There are considerations to address when designing and implementing a Virtual SCSI
environment. The primary considerations are:
• Dedicated processor server partitions or Micro-Partitioning
• Server partition memory requirements
• One thing that does not have to be factored into sizing is the processor impact of using
Virtual I/O on the client. The processor cycles executed on the client to perform a Virtual
SCSI I/O are comparable to that of a locally attached I/O. Thus, there is no increase or
decrease in sizing on the client partition for a known task.

14.6.3.1 Sizing when using Dedicated Processors

One sizing method is to size the Virtual SCSI server to the maximum I/O rate of the attached
storage subsystem. The sizing could be biased to small I/Os or large I/Os. Sizing to maximum
capacity for large I/Os balances the processor capacity of the Virtual SCSI server to the potential
I/O bandwidth of the attached I/O. The negative facet of this sizing methodology is that, in
nearly every case, we will assign more processor entitlement to the Virtual SCSI server than it
typically consumes.
Consider a case where an I/O server manages 15 physical SCSI disks. We can arrive at an upper
bound of processors required based on assumptions about the I/O rates that the disks can
achieve. If it is known that the workload is dominated by 16 KB operations, we could assume
that the 15 disks are capable of 1 read transaction every 36 milliseconds. An IBM i operating
system Virtual SCSI server could support around 30,000 read transactions per second on a single
processor provided enough disk were present.
IBM i 6.1 Performance Capabilities Reference - January/April/October 2008
©
Copyright IBM Corp. 2008
Chapter 14 DASD Performance
238

Hide quick links:

Advertisement

Table of Contents
loading

This manual is also suitable for:

7xx serversAs/400 risc server

Table of Contents