Since we are currently trying to get multipathing working with FCP and SLES9 (SLES8 was easier but different) it does make us well aware of how we take it for granted when we use good old easy ECKD (not that ours is very old- ECKD is still always improving). I would add to the list of ECKD benefits- simpler admin and simpler,more automated DR. On the other hand, SAN/FCP dasd is bought in bulk and has cheaper chargeback. ECKD is viewed as expensive. We've seemed to come up with arbitrary standards - like if you need more than 25GIG- you get SAN, or keep software product filesystems on ECKD for easy cloning, etc. Basically the ECKD dasd we have is much more limited (and more expensive) and SAN seems unlimited. So for all larger dasd requests you end up giving the customer SAN lun's. Then you get the questions 'Since the dasd is the same as a SUN Solaris server the I/O performance must be the same, right?' Or 'How does linux on the mainframe I/O performance compare to linux on intel? It's the same dasd. Must be the same, right?'
Are there published benchmarks for (1) mainframe linux ECKD vs. FCP or (2) mainframe linux FCP vs. intel linux FCP? What are the sources you refer to? -----Original Message----- From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Pieter Harder Sent: Sunday, September 10, 2006 1:55 PM To: LINUX-390@VM.MARIST.EDU Subject: FCP over ECKD performance advantage - why? Hello list, there are a number of sources that indicate that FCP attached DASD performs better than classic ECKD DASD. My own numbers seem to confirm this. But I am wondering what exactly is the advantage that FCP has over ECKD? I can't be the physical storage box, that is the same for most people, something like a DS8000/DS6000 or a Symmetrix. I can also hardly believe it is the software layer within the storage box, as both FCP and ECKD are emulated/simulated on top of a native storage structure. It could be that the software/hardware interface provided by QDIO is so much better than the old Start-I/O model. But is it? Is it not just a matter of FCP not spending the cycles to provide stuff that ECKD users take for granted, like: - multipathing - performance instrumentation - device isolation for security reasons - error handling - and more Is it possible that when all the above is added to FCP there is no performance advantage at all? I am sure there are knowledgeable people on the list who have something to comment on this. Thanks for any insights. Best regards, Pieter Harder [EMAIL PROTECTED] tel +31-73-6837133 / +31-6-47272537 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ************************************************************************* This communication, including attachments, is for the exclusive use of addressee and may contain proprietary, confidential and/or privileged information. If you are not the intended recipient, any use, copying, disclosure, dissemination or distribution is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return e-mail, delete this communication and destroy all copies. ************************************************************************* ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
