1. SMF Type 113 Records
2. SHARE in Anaheim
3. zPrime Update
1. SMF Type 113 Records
In Cheryl’s List #147, we described two new tools that we created to show you your Relative Nest Intensity (RNI) and workload. These are available on our website (www.watsonwalker.com and select Free Tools on the left). But we forgot to mention that the MXG 29.02 level is needed in order to provide correct information. Remember that we’d love it if you would share your results with us.
Here is an updated link to John Burg’s presentation –http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TC000066.
2. SHARE in Anaheim
What a great week I had in Anaheim. Attendance is picking up now that the economy is coming back, and the energy at the conference was terrific! There were a lot of sessions on the z196 (which has gotten rave reviews) and z/OS 1.12 (which has also gotten rave reviews). And, of course, there were many sessions that included the things we can expect from z/OS 1.13.
I gave two presentations during the week, and you can find them on the SHARE website (www.share.org) or on our website (www.watsonwalker.com) under Presentations. They are session 8797 – Hot Flashes #25 and session 8697 – Influencing IBM’s Development (e.g. SHARE requirements). I try to include something for everyone in the Hot Flashes, so I hope you find it interesting. If you’re a member of SHARE and want to have some influence on what IBM includes in the next release, then please consider joining the MVSE requirements process. See the presentation to find out how to join and participate.
3. zPrime Update
As our Tuning Letter readers may remember, we wrote about zPrime from Neon Enterprise Software in our 2009 No. 4 issue (pages 27-32), and added more about it in our 2009 No. 5 issue (pages 34-35). zPrime is the software product that moves more work to zIIPs and zAAPs, more than IBM feels is valid. I suggested that users might want to wait to use the product until the disagreement was resolved, and that it might end up in court.
As I predicted, the lawsuit between Neon and IBM happened and is still ongoing. But there was an interesting Summary Judgment on one aspect filed recently that will be of interest to those following this story. The decision was filed on February 17, 2011 and was in response to Neon’s Partial Motion for Summary Judgment and IBM’s Cross-Motion for Partial Summary Judgment. Both motions were denied, but the finding might give an indication of what will happen during the actual trial. I thought that the finding was fascinating reading. Here are three important and interesting sections from the 33-page finding (https://ecf.txwd.uscourts.gov/doc1/18117680118, PACER website, requires registration and payment of $0.08/page):
IBM argues that a customer cannot run workloads on processors unless they are authorized to do so. Neon argues that once a processor is activated, IBM mainframe customers can run any workload on the processor, unless the IBM customer contracts contain a specific workload restriction. The Court agrees with IBM and disagrees with Neon.
The record contains volumes of marketing information and official IBM Redbook publications discussing what zIIPs and zAAPs can do and why they may be an attractive option for customers. At no point in the record is there an instance where IBM indicates that zIIPs and zAAPs may process all types of workloads. Rather, zIIPs and zAAPs have always been intended to process a limited class of workloads.
IBM’s argument assumes that for each authorized zIIP or zAAP there are a set of clearly defined authorizations as to their “workload-specific capabilities”. This assumption is incorrect. As mentioned in Part I, the Court finds that the extent of authorizations for the “workload-specific capabilities” of zIIPs and zAAPs are a question for the jury-at this moment, they are unclear. Because of this undefined variable, there is no way to determine whether zPrime allows IBM customers to circumvent technological measures in the Machine Code that limit authorized “workload-specific capabilities” (permissible), or whether zPrime allows mainframe customers to circumvent technological measures in the Machine Code that limit unauthorized “workload-specific capabilities” (a violation of section 3). Thus, because circumvention alone is not enough, IBM’s Motion must be denied.
There is a footnote to the next to the last sentence in the previous paragraph:
In other words, it is currently impossible to tell whether technological measures in the Machine Code interact with z/OS to impermissibly restrict workloads that have been contractually authorized.
I still stand by my Tuning Letter recommendation, which is that customers should probably wait to see what’s going to happen. It’s also a possibility that other Independent Software Vendors (ISVs) may jump into the scene in some manner, simply because their bottom line will be impacted if a lot of work is moved from the general processors (since most software vendors charge on MIPS or MSUs).
For one more reference, our Tuning Letter 2008 No. 4 (pages 28-30) listed over 30 products that I could find that run on zIIPs and zAAPs. There have been several more added since then, but all are generally either monitor tools or new workloads.
This is a good time to repeat my recommendation for software pricing. The current problem with z/OS pricing is that most software is charged on the size of the machine, not the amount of usage of the software. So if you want to move a new workload onto z/OS that requires more MIPS, the cost of some of the other work on the system will also increase. This has been a problem ever since I was first involved in chargeback in the late ’60s. You might be running an IMS system and want to add a WebSphere application to the machine and double the number of MIPS. The IMS system, which is serving the same number of customers as before, and using the same amount of CPU as before, will now cost more because the machine is bigger.
IBM attempted to resolve part of this problem by introducing usage-based pricing (Measured Usage License Charges, MULC), where you pay for the highest use of the software during the day/month. But that was only for a few products. IBM and other vendors have also tried user-based pricing where you pay for each user. With either of these types of pricing, you wouldn’t pay more just because the size of the machine increased. I’d like to see more products available as usage-based or user-based pricing. Until such a solution, however, every z/OS site will need to deal with the cost of existing software whenever the platform capacity is increased. MULC was described in our Tuning Letter 1998 No. 2 (pages 21-22).
At present, the best chance you have of lowering your software bill is to use sub-capacity pricing. Workload License Charges (WLC) and variable WLC (VWLC) were described in our Tuning Letter 2001 No. 4 (pages 4-26).
I’ll keep you posted as I hear news of the trial.