IBM lifts the veil on Power7 chips
Promises upgrades from Power6 boxes
With every new server processor generation, there is always the possibility that the chip and its associated chipsets and system components are so different that both the chip and the server platform have to change at the same time. This will not be the case for the Power7 processors due next year, says IBM.
IBM is in no hurry to ship Power7 chips and the systems that use them. Not with so many other products being delayed. Intel's "Tukwila" quad-core Itaniums have been pushed back until the first quarter of 2010 and probably will not show up in systems until the middle of 2010. While its eight-core "Nehalem EX" (sometimes called "Beckton") Xeon-class processors are not going to be ready until late this year and probably not into systems until the first quarter of 2010.
Advanced Micro Devices is expected to get its 12-core "Magny-Cours" Opterons and their G34 chipset out the door in the first quarter of next year, but it is unclear who will be building big bad boxes out of them. And the fate of the 16-core "Rock" UltraSparc-RK chip was far from certain even before Sun Microsystems was slated to be eaten by software giant Oracle, and the rumor back in mid-June was that Sun had already killed off the Rock chip ahead of the Oracle deal. Fujitsu may have some speed bumps for its quad-core Sparc64-VII processors, which are sold in Solaris systems by Sun and Fujitsu, but it is not expected to get its own eight-core "Venus" Sparc64-VIII into the field until 2010 or 2011.
So IBM is not feeling the pressure from its competition in the midrange and high-end of the server racket to say squat about Power7. But its Power Systems customers, having seen so many other chips in the market delayed in the past year or more, are looking for a little confirmation that the Power7 chip is on track.
They also want to know whether or not the Power7 chips will be available to them through an upgrade process or if they will have to do a push-pull box swap to get to the new machine. In this economic environment, when Power server customers are not sure when the next thing will be available, what kind of performance it will offer, and what kinds of upgrade paths will be available, customers are not inclined to spend now. Since a new generation of chips and servers almost always means a boost in bang for the buck, it is smarter to wait it out.
IBM, of course, doesn't want customers to wait it out. It wants shops to buy Power6 or Power6+ systems today because it needs to make its numbers this quarter and next. And so Scott Handy, vice president of marketing and strategy for the Power Systems division, says that customers who buy Power 570 machines (which have up to 16 Power6 cores or up to 32 cores Power6+ cores) and Power 595 machines (which have up to 64 of the Power6 cores) will be able to upgrade to Power7 machines in such a way that they can preserve their serial numbers on their boxes.
Such an upgrade path is important because accounting rules require that an upgrade that preserves serial numbers, and therefore does not subject the box to immediate depreciation of any extant value in the box, keep a fairly large amount of the original box as part of the upgrade. According to Handy, the Power 595 machine, which is a big SMP box with NUMA-ish cache coherency, will be upgradeable by swapping out processor motherboards (what IBM calls books) with Power6 chips and swapping in Power7 boards.
The Power 570 has a slightly different architecture, with processor cards plugging into a chassis and several chassis being linked together with NUMA-ish chipsets and presenting a single system image, but Handy says that the way IBM is upgrading Power 570s from Power6 and Power6+ chips to Power7 chips will allow it to legally preserve serial numbers.
This means customers who are buying Power 570s and 595s today know they won't have to ditch these boxes entirely if they need to move to Power7 machines next year or the year after, which means they don't have to write off their investments ahead of their normal depreciation schedule. The guarantee of upgrade paths also means that customers had better take a stab at what they think pricing will be on Power7 boxes and make sure they get steep discounts on the current Power Systems machines, because they can bet IBM isn't going to give away upgrades to Power7 chips.
The neat trick with upgrades this time around is that the PowerVM hypervisor that logically partitions Power Systems machines has a feature called Live Partition Mobility, which will enable customers with two physical servers to move workloads off of one box, upgrade it to Power7 chips, and then move workloads back to partitions running on the new Power7, all without shutting down the running workloads. Customers with only one machine can do the upgrade as part of planned downtime (or they could rent a secondary box from IBM or a reseller just to do the Aztec two-step).
Next page: Price cut coming?
"Though, I have a feeling you're exaggerating quite a bit here since a "major UK bank" experiencing the frequent outages you describe would switch to something else ASAP."
Oh, there are politics involved. The best tech doesnt necessarily wins. How many uses sucky Windows, and never switch to Linux? In my company, large fortune 500, Ive recommended new technology that would make enormous savings and better up time. But to no avail. The management has decided on some technology and that's it. I can do nothing, despite showing better calculations.
@AC Not Found
The link worked for me, and just tried again and it works
" IBM Power 595 and 570 Servers
CEC Concurrent Maintenance
April 14, 2009 "
"I can say that popping ram chips, MCM's and RIO books happens so frequently that I'd love it if I could hotswap with the system on line"
Something must be wrong with your environment, we have a dozen of p570s here untouched for months and a couple with 1 year+ of uptime.
"overcommitted as per IBM's own best practices please explain how I remove 4/8 cores from a system already running at capacity?"
Another IBM best practice in these cases is to have Live Partition Mobility-ready servers, so you can move partitions easily without disrupting your apps.
Best regards! Cheers!
@AC 'Please, update yourself before posting. '
Same shit, different badge - you did notice I said 'rs6000/system P/pSeries' didn't you?
I use p570's on a daily basis, and I can say that popping ram chips, MCM's and RIO books happens so frequently that I'd love it if I could hotswap with the system on line. Yeah I know the new flexy interconnect helps coldswapping a node out, but with a virtualised load which is overcommitted as per IBM's own best practices please explain how I remove 4/8 cores from a system already running at capacity?
I followed your link in case I missed something from last weeks IBM sales propaganda sessions... I guess not...
The requested URL /common/ssi/sa/wh/n/pow03023usen/POW03023USEN.PDF was not found on this server.
IBM_HTTP_Server/18.104.22.168 Apache/2.0.47 (Unix) Server at ftp.software.ibm.com Port 80
@AC RS/6000 misinformed dude
Dude, RS/6000 brand is gone not yesterday but 8-10 years ago.
After that, it was pSeries and then POWER servers.
Please, update yourself before posting.
Try starting reading, this:
That is related to concurrent maintenance of high-end servers that ONLY 'mainframe' class systems are enabled to..
Best regards! Cheers!
'That means, you can repair live machines without powering it off.'
I nearly wet myself.
I can't recall the last time I did any system maintenance at a major UK bank that didn't require an outage on rs6000/system P/pSeries (still sounds like 'pessary' to me). pSeries stuff is so far behind 'mainframe' class systems for in-situ repairs. You still can't replace memory or cpu while the box is up which is such a PITA since the parts IBM sources fail constantly.
Oh you meant disk swaps? Yeah, you can do those without shutting an rs6000 down... usually.