Filter
Top Products
14.1.10 Direct Attach 571E/574F and 571F/575B Observations
We did some simple comparison measurements to provide graphical examples for customers to
observe characteristics of new hardware. We collected performance data using Collection
Services and Performance Explorer to create our graphs after running our DASD IO workload
(small block reads and writes).
IOP vs IOPLess: no measurable difference in CPU or throughput.
Newer models of DASD are U320 capable and with the new IOAs can improve workload
throughput with the same number of DASD or even less in some workload situations.
IOA’s 571E/574F and the 571F/575B achieved up to 25% better throughput at the 40% DASD
Subsystem Utilization point than the 2780/574F IOA. The 571E/574F and 2780/574F were
measured with 15 DASD units in a 5094 enclosure. The 571F/575B IOAs attached to #5786
EXP24 Disk Drawers.
System Models and Enclosures: Although an enclosure supports the new DASD or new IOA,
you must ensure the system is configured optimally to achieve the increased performance
documented above. This is because some card slots or backplanes may only support the PCI
protocol versus the PCI-X protocol. Your performance can vary significantly from our
documentation depending upon device placement. For more information on card placement
rules see the following link:
IBM i operating system iV5R2: http://www.redbooks.ibm.com/redpapers/pdfs/redp3638.pdf
IBM i operating system iV5R3, iV5R4:
http://www.redbooks.ibm.com/redpapers/pdfs/redp4011.pdf
Conclusions: There can be great benefits in updating to new hardware depending upon the
system workload. Most DASD intense workloads should benefit from the new IOAs available.
Large block operations will greatly benefit from the 5094/5294 feature code #6417/9517
enclosures in combination with the new IOA’s and DASD units.
Note: The #6417/9517 provides a faster HSL-2 interface compared to the #2887/9877 and is
available for I/O attached to POWER-5 based systems
IBM i 6.1 Performance Capabilities Reference - January/April/October 2008
© Copyright IBM Corp. 2008 Chapter 14 DASD Performance 204