Power 550 i Edition Versus Windows: Get Big Discounts on i Software
September 15, 2008 Timothy Prickett Morgan
In last week’s issue of The Four Hundred, I did the first in a series of articles comparing the Power Systems i Edition platform to other competitive platforms. In that case, I compared entry Power 520 rack servers and JS12 blade servers in the new BladeCenter S chassis to ProLiant alternatives running Windows from Hewlett-Packard. I was gratified to see that the i platforms were absolutely competitive with the Windows iron, at least when comparing machines with system software suitable for online transaction processing.
This week, I am not as happy, since the Power 550 M50 machines, thanks in large part to the excessive fees IBM is charging for software, is not really all that competitive with the equivalent Windows machinery. Hardware is not the problem. Software is. The good news, however, is that the four-core JS22 blade server does a better job competing with HP’s ProLiant rack and blade servers, at least as far as I can tell from my comparisons. If you haven’t gotten the message yet, let me repeat it again: If you are an SMB shop that does not need the computing capacity provided by the Power 570, then Big Blue really, really, really wants you to consider using the blade form factor. The economics are far more favorable for configured systems, particularly when comparing the Power 550 M50 and the JS22 blade. The gap was not nearly as large when comparing the JS12 blade and the entry Power 520 servers with i Edition software.
As you can see from the configurations I have come up with this week’s OLTP performance and price/performance table, and from the prior System i and Power Systems i comparisons I made earlier this summer, IBM gave customers a pretty good jump in bang for the buck based on per transaction costs when moving from the System i5 550 to the Power 550–roughly 30 percent to 40 percent by my math. But on a per user basis–assuming the number of users remains constant across the two families of machines–the improvement was considerably less, and the 3.5 GHz Power6 configurations in the Power 550 lineup actually offered worse bang for the buck than the Power5+ System i5 550s using 1.9 GHz cores from nearly two years ago. Comparisons with similarly power HP ProLiant machines running Windows Server 2008 Enterprise Edition, SQL Server 2008 Enterprise Edition, and VMware‘s Infrastructure Standard Edition hypervisor and related tools are even less appealing, and as users are added to configured systems, the gap is even wider.
IBM’s per user charges for i5/OS V5R4 and i 6.1 seats after the basic operating system license are not a problem. It costs just under $40 per seat for a Windows CAL (short for Client Access License) and $162 for a SQL Server CAL if you use the per system plus CAL licensing option from Microsoft that is most like the packaging IBM uses for the i platform entry and midrange machines. IBM is charging $250 a seat for users, and is even giving customers a break and calling a user a concurrent user, not a named user. This is reasonable. But Windows Server 2008 and SQL Server 2008 are priced based on a system–you can do per core licensing for SQL Server, but on entry and midrange boxes with relatively few users, this is a stupid option–while IBM is charging per processor core for i5/OS and i, which of course package the operating system and database together. And because most i shops need the 5250 protocol for their applications, this is added in as well. And that, my friends, jacks the price up through the roof compared to Windows boxes.
As X64 iron running Windows gets more and more cores, the amount of water that will open up between these two platforms will only get wider. And that is a problem for an IBM that should be more determined to protect its base and less worried about Software Group’s already high profit margins. If Software Group loses large numbers of AS/400, iSeries, and System i customers–trust me, they are going all Microsoft on their stack, these companies like an integrated stack from one vendor, after all–there is not going to be any DB2 for Windows or Linux action for Software Group to take as consolation. So maybe that $40,000 per core for i5/OS V5R4 and i 6.1 and that $50,000 per core for 5250 enablement is a great business, and maybe it isn’t. The money is only good if you get it and you can spend it. Otherwise, it is a disincentive for customers to buy, and they jump to Windows when the CFO or owner puts pressure on the IT manager.
Take a look and see for yourself in the charts below. The first one compares IBM’s Power Systems i and HP’s ProLiant Windows boxes with 40 and 150 users with roughly equivalent performance and software stacks. This is cost per user for a configured system:
Because IBM has priced the i software stack very aggressively on the JS12 and JS22 blades, the i platform does pretty well against the HP blade alternatives on a per-user basis. The chart above compares JS22 blades to the HP ProLiant BL460c blades, both of which are two-socket machines; the IBM blade has a maximum of four cores, but the HP blade can do up to eight cores if quad-core Xeons are used. As you can see, at least for OLTP workloads, HP is charging a premium for its X64 blades compared to the X64 rack servers, and the difference on this is entirely hardware. And absolutely expected. But IBM’s pricing for the Power 550 M50 server is way out of whack–and even if you knock off $200,000 and completely rip 5250 enablement off the machine (which AS/400 shops can’t do, of course), it is still overpriced compared to a similarly powerful Windows box. I am talking about a factor of four to six more expensive with 5250 enablement on all cores, and a factor of two when you remove it from the equation.
Now, let’s double up the processing and memory capacity of the servers and also double up the users. Look at the gap grow for the Power 550:
No matter how much IBM wants to call the Power 550 an enterprise-class box that should not have user-based pricing and low software costs, the market is saying otherwise. And that has been my contention for years. So I will say it again: If IBM really wants to price to value, it needs to have user-based pricing across the entire Power System i lineup. Period. Charge a reasonable fee per system or per core (let customers choose) and then a reasonable fee per user above that–just like the Windows stack.
This is the methodology I used to make these comparisons. First, I took the Power Systems i configurations I put together earlier in the summer as part of a comparison to the two prior generations of iSeries and System i machines (see The Power Systems M15 and M25 Versus Their Predecessors and The Power Systems JS12 and JS22 Blades Versus Other i Boxes) and consolidated these down into one sheet with just the Power 520 and BladeCenter S plus JS12 blade servers. I did a lot of different mixes of processor performance, memory and disk storage, and user counts to give you a sense of how these factors affect the overall price of a system and the price per transaction or user. Each configuration included i 6.1 and the PowerVM Standard Edition hypervisor, which allows basic logical partitioning. The i tower machines have the number of disk drives shown and RAID 5 data protection, while the i blades have two on-blade disks for local storage as well as access to in-chassis storage blades, which are configured with RAID 1 mirroring protection. I also threw in an enterprise-class SAS LTO-3 tape drive for all of the machines, which is perfectly reasonable. In the comparisons, I also removed Software Maintenance fees from all of the systems, even though IBM requires customers to buy it, because I think customers should have the choice of maintenance services or not (or IBM or otherwise) and because I want to isolate system costs from support costs.
On the Windows side, Hewlett-Packard is the undisputed dominant shipper of X64 ProLiant tower and blade servers in the world, and comparing Windows running on ProLiants is what matters at this point in the server market. IBM, Dell, Sun Microsystems, Fujitsu-Siemens, and others all have to compare their speeds, feeds, and prices against the ProLiant line, which is the touchstone for workhorse X64 computing. Dell has its fans, and Sun has nifty designs, but HP is the market leader. And so my comparison, this time around, is limited to HP ProLiant rack servers of equivalent processing capacity to the Power 520 i Edition and BladeSystem c3000 blade boxes with ProLiant BL class blades that have roughly the same oomph as the IBM BladeCenter S chassis using JS12 entry blades. The HP boxes are all configured with Windows Server 2008 Enterprise Edition, the roughly analogous version of the Windows platform that has the features of i 6.1, plus the new SQL Server 2008 Enterprise Edition database, the reasonable choice for midrange-sized OLTP workloads. (I used the cheaper SQL Server Standard Edition in the Power 520 comparisons last week.)
I also threw VMware‘s Infrastructure Standard edition hypervisor stack, which includes the ESX Server 3.5 hypervisor plus the VirtualCenter management, the Virtual Machine File System (VMFS), Symmetric Multi Processing (vSMP, which allows a virtual machine to span multiple cores), and tools for updating, backing up, and clustering virtual machines. This is not a cheap software stack, even on a small box. And there are cheaper Windows editions and freebie hypervisors, but these are not the ones that real companies will deploy for their core ERP systems. Finally, I removed HP’s various tech support options from the box, and ditto for the Windows software stack. There is no way to separate the first year of support from VMware’s stack, however.
To help make your capacity planning a bit easier, I calculated the cost per transaction per minute for the machines based on IBM’s Commercial Performance Workload benchmark and my conversion of that rating to estimated TPC-C online transaction processing performance. For the HP machines, I worked backward from real TPC-C tests on X64 iron and estimated what each HP setup would do. I would be the first to agree that there is a lot of estimating there.
As I said last week, you could argue that a database engine and ERP box doesn’t need to be virtualized, and therefore adding the extra costs of the virtualization hypervisor and associated management tools that make it useful is not a fair comparison at all. But in 2008, I think this is becoming an increasingly invalid argument. More and more companies are deploying virtualization, and I think that many companies are going to be deploying their n-tier application architectures on smaller numbers of virtualized boxes–the kinds of things that AS/400, iSeries, and System i customers have been doing for years with database and application servers using subsystems (a precursor to logical partitions and akin, in some ways, to virtual private servers for Windows, Linux, and Solaris today). I think data center managers are being increasingly pressured to simplify their infrastructure, and not for green reasons–unless you are talking about money. The green benefits, in terms of energy efficiency, are only a concern for the very largest data centers with exploding workloads, and maybe a bunch of midrange shops in urban areas where energy and office space are priced at a premium. And while I am thinking about it, the OLTP performance estimates in these comparisons do not take into account any processing overhead that the hypervisors impose. I have no idea how much capacity gets chewed up by the current crop of hypervisors. That is a story for a different day.
With Intel announcing its new six-core “Dunnington” Xeon processor for four-socket servers this week, I am keen on seeing how Power 570s rate against big Xeon iron. I am working on that now, and hopefully I can have that done next week.