>We have a parallel sysplex environment running on a z/800. We've used >all 15 LPARs supported by the Hardware, CPU and memory are max'ed out >and we need more physical resources. If we buy another (used) z/800 (we >are also considering an upgrade to a z9 but budget is an issue), can we >have the same sysplex spread across both z/800's sharing the CPUs and >memory and use the LPARs on the new z/800 to add more systems to the >same existing sysplex? If so, what needs to happen? Can anyone point me >to any literature that I could use as a guide? Currently we have ICF >setup in 2 LPARs and two processors shared by all 15 LPARs.
Yes, you can. In fact, you can add 31 more systems if you wish. :-) It's really not much different from what you're doing now, but of course you'd have two frames. Start here for details: http://www.ibm.com/servers/eserver/zseries/pso/ However, with respect to the System z9 BC and your budget, I would strongly recommend taking a very careful look at that. In particular, you should factor in your ongoing software costs which will be lower for the same LSPR capacity as your z800. I would recommend including speciality engines (zAAP, zIIP, and IFL) in your calculations. These processors can offload Java, certain DB2 work, and Linux, respectively, from the main processors and offer a lower cost profile. Encryption is much faster, too. The z9 BC capacity settings are more precise -- the system size can be exactly what you need at a moment in time, not too much. And I would look at the ongoing maintenance costs. The z9 BC supports up to 30 LPARs, depending on the model, so you should be fine there. (You have a decent number of LPARs, though. Have you considered z/VM for some of your work?) If you want to take a step forward and have a "true" Parallel Sysplex so that you have extra redundancy (and higher potential availability and reliability for your users) compared to what you have now, then you could keep the z800 as your secondary system. There are a variety of options there, but you might want to shift everything over to the new system and keep the z800 on cold or warm standby. That'd probably be the most economical two-system approach. Better yet, you could configure the z800 at or near its lowest capacity setting and place it into the Sysplex with Capacity Backup (CBU). In the unlikely event the System z9 BC failed (in a disaster, for example), you could fire up the CBU capacity on the z800 to handle your "essential workload." And that could substantially reduce your DR budget since you can bring some of it in-house. You may also wish to consider whether the z800 and z9 BC would be located in different buildings for extra protection, along with storage. The bottom line, though, is that it's not a good idea to look at the hardware costs alone -- for mainframes or for any other computing system. The hardware is cheap compared to all the other costs. I know it's hard to break that thinking since a lot of people remember when hardware really was expensive. Hope that helps. - - - - - Timothy F. Sipples Consulting Enterprise Software Architect, z9/zSeries IBM Japan, Ltd. E-Mail: [EMAIL PROTECTED] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
