Just What The Heck Is An MSU??
Published:
March 24 2009, 10:15 AM
by
Reg Harbeck
G'day Mainframers,
I would like to welcome both zBert and Bramie to me blog. Down Under, all are very welcome… as long as you have a slab on its way over - first class of course! So I am eagerly awaiting the two packages of VB in the mail. On a serious note I say a very warm (and thirsty) G’Day to both zBert and Brammie and wish them all the best. Hoping that you share your worldly (and occasionally funny) anecdotes from experiences that relate to innovations and inventive solutions to Mainframe issues of today...Now for the MSUs...
I used to sit at my desk and hope that today the laws of physics would alter - and that empty vessels would not make so much NOISE (by continually voicing their utter lack of technical knowledge). I guess I was getting a little fed up with all the negativity from those others working (for want of a better term) for that outsourcer. I gleaned one truth – hope does not turn the crashing tide of incoherent babble - to silence. Finally the torrent subsided as the sirens call for nicotine and caffeine was answered by the pack (Mustelus Antarcticus). The lack of senseless noise was almost deafening - as I was at last, left alone to concentrate on my customers frustrations over MSUs. They wanted some simple way of visualizing it and the impact of Soft Capping – and the outsourcer’s boys would have none of this. They liked to keep the customer in the dark – I wished only to be left in silence. So have you ever wondered – ‘Just What The Heck Is An MSU’. What the dickens is an MSU anyway? Where can I find it? What do I do with it – hit it with a stick, eat it or drink it? The good news is that MSUs are simply - Millions of CPU Service Units and is the measure of the capacity of your IBM Server. What is it for? Software Charges/Pricing for products (IBM and other ISVs) based NOT on total capacity of the Server – but on the utilization within a Logical Partition/s (Sub Capacity). If I were a customer – what would I want? I would have this presented to me in a simple and concise manner with a discussion on what is - (1) Sub Capacity Planning and what the (2) Reporting Tool is all about. Then I would want to know how to (3) Monitor MSUs in real-time. Not wait till the end of the month for a nasty shock. This as opposed what we had delivered - a post-it note on the bottom of the console … ‘Call me if you have any probs D.H.’ Is this appropriate – I bloody well think NOT. The best place to start is … at the beginning. The (1) Sub Capacity Planning Tool is used to analyze your Central Processing Complex (that’s the big box from IBM in the middle of the floor) and all Logical Partitions for their 4-hour rolling average to determine if you qualify for and will benefit from - Sub Capacity Charging. This will require the SMF type 70 (subtype 1) and 89 (subtype 1 & 2) records on ALL LPARS. If the outcome with respect to your software charges from IBM is favorable – check that your other software vendors also support this method of billing – before you leap. Only when all software charges are taken into consideration should you progress and decide that it is appropriate to move to Sub Capacity charging and that the IBM (and others) terms and conditions are able to be met. Also note that the tool uses a different algorithm than the Sub Capacity Reporting Tool (+/- 3-5% variance) and can also analyse servers in basic mode.
Now that we not only qualify – but glean some financial incentive to do this - it is required that the (2) Sub Capacity Reporting Tool (SCRT) be run against SMF data to produce the monthly report that is sent to IBM. This is used as input to determine your monthly charges - other vendors will have the same or similar stipulations. Also the use of the Soft Cap - sometimes referred to as the beanie (Aussie for woolen cap used at the footy) is a mechanism to limit the Non-Production capacity through WLM prioritization. The throttle of lower important workload to reduce throughput till the extrapolated 4 hour average meets that stipulated in the Beanie.
Last but certainly not least is - How to (3) Monitor MSU’s which is the major focus of this note. Why wait for the end of the month or batch reports when I can view the MSU consumption online (almost real-time or at least within the RMF sample time). How do I determine if delays in NON-Production are related to Soft Capping (yes the beanie) or some other issue. The answer is a non-mainframe (Windows based) client that will graphically display the MSU over time and WLM Capping%. Thus I have attached a simple – How to for RMF PM (Java) on the PC.
RMF PM (Java) GUI for Windows
Stay tuned for a simple 'How to' for RMF PM for Windows – it is intended to assist with monitoring of MSU usage on your Lpars....
Cheers Now,
Bruce
By: Reg Harbeck
Reg Harbeck is CA's Product Management Director for Mainframe Strategy. In the more than two decades since he received his Bachelor's Degree in Computer Science he has worked with operating systems, networks, security and applications on mainframes, UNIX, Linux, Windows and other platforms. Reg...
Read More..