An ongoing discussion about SAP infrastructure

HoP keeps Hopping Forward – GA Announcement and New IBM Solution Editions for SAP HANA

Almost two years ago, I speculated about the potential value of a HANA on Power solution.  In June, 2014, SAP announced a Test and Evaluation program for Scale-up BW HANA on Power.  That program shifted into high gear in October, 2014 and roughly 10 customers got to start kicking the tires on this solution.  Those customers had the opportunity to push HANA to its very limits.  Remember, where Intel systems have 2 threads per core, POWER8 has up to 8 threads per core.  Where the maximum size of most conventional Intel systems can scale to 240 threads, the IBM POWER E870 can scale to an impressive 640 threads and the E880 system can scale to 1536 threads.  This means that IBM is able to provide an invaluable test bed for system scalability to SAP.  As SAP’s largest customers move toward Suite “4” HANA (S4HANA), they need to have confidence in the scalability of HANA and IBM is leading the way in proving this capability.

A Ramp-up program began in March with approximately 25 customers around the world being given the opportunity to have access to GA level code and start to build out BW POC and production environments.  This brings us forward to the announcement by SAP this week @ SapphireNow in Orlando of the GA of HANA on Power.  SAP announced that customers will have the option of choosing Power for their BW HANA platform, initially to be used in a scale-up mode and plans to support scale-out BW, Suite on HANA and the full complement of side-car applications over the next 12 to 18 months.

Even the most loyal IBM customer knows the comparative value of other BW HANA solutions already available on the market.  To this end, IBM announced new “solution editions”.  A solution edition is simply a packaging of components, often with special pricing, to match expectations of the industry for a specific type of solution.  “Sounds like an appliance to me” says the guy with a Monty Python type of accent and intonation (no, I am not making fun of the English and am, in fact, a huge fan of Cleese and company).  True, if one were to look only at the headline and ignore the details.  In reality, IBM is looking toward these as starting points, not end points and most certainly not as any sort of implied limitation.  Remember, IBM Power Systems are based on the concept of Logical Partitions using Power Virtualization Manager (PVM).  As a result, a Power “box” is simply that, a physical container within which one or multiple logical systems reside and the size of each “system” is completely arbitrary based on customer requirements.

So, a “solution edition” simply defines a base configuration designed to be price competitive with the industry while allowing customers to flexibly define “systems” within it to meet their specific requirements and add incremental capability above that minimum as is appropriate for their business needs.  While a conventional x86 system might have 1TB of memory to support a system that requires 768GB, leaving the rest unutilized, a Power System provides for that 768GB system and allows the rest of the memory to be allocated to other virtual machines.   Likewise, HANA is often characterized by periods of 100% utilization, in support of instantaneous response time demanded of ad-hoc queries, followed by unfathomably long periods (in computer terms) of little to no activity.  Many customers might consider this to be a waste of valuable computing resource and look forward to being able to harness this for the myriad of other business purposes that their businesses actually depend on.  This is the promise of Power.  Put another way, the appliance model results in islands of automation like we saw in the 1990s where Power continues the model of server consolidation and virtualization that has become the modus operandi of the 2000s.

But, says the pitchman for a made for TV product, if you call right now, we will double the offer.  If you believe that, then you are probably not reading my blog.  If a product was that good, they would not have to give you more for the same price.  Power, on the other hand, takes a different approach.  Where conventional BW HANA systems offer a maximum size of 2TB for a single node, Power has no such inherent limitations.  To handle larger sizes, conventional systems must “scale-out” with a variety of techniques, potentially significantly increased costs and complexity.  Power offers the potential to simply “scale-up”.  Future IBM Power solutions may be able to scale-up to 4TB, 8TB or even 16TB.   In a recent post to this blog, I explained that to match the built in redundancy for mission critical reliability of memory in Power, x86 systems would require memory mirroring at twice the amount of memory with an associated increase in CPU and reduction in memory bandwidth for conventional x86 systems.  SAP is pushing the concepts of MCOS, MCOD and multi-tenancy, meaning that customers are likely to have even more of their workloads consolidated on fewer systems in the future.  This will result in demand for very large scaling systems with unprecedented levels of availability.  Only IBM is in position to deliver systems that meet this requirement in the near future.

Details on these solution editions can be found at
In the last few days, IBM and other organizations have published information about the solution editions and the value of HANA on Power.  Here are some sites worth visiting:

Press Release: IBM Unveils Power Systems Solutions to Support SAP HANA
Video: The Next Chapter in IBM and SAP Innovation: Doug Balog announces SAP HANA on POWER8
Case study: Technische Universität München offers fast, simple and smart hosting services with SAP and IBM 
Video: Technische Universität München meet customer expectations with SAP HANA on IBM POWER8 
Analyst paper: IBM: Empowering SAP HANA Customers and Use Cases 
Article: HANA On Power Marches Toward GA

Selected SAP Press
ComputerWorld: IBM’s new Power Systems servers are just made for SAP Hana
eWEEK, IBM Launches Power Systems for SAP HANA
ExecutiveBiz, IBM Launches Power Systems Servers for SAP Hana Database System; Doug Balog Comments
TechEYE.netIBM and SAP work together again
ZDNet: IBM challenges Intel for space on SAP HANA
Data Center Knowledge: IBM Stakes POWER8 Claim to SAP Hana Hardware Market
Enterprise Times: IBM gives SAP HANA a POWER8 boost
The Platform: IBM Scales Up Power8 Iron, Targets In-Memory

Also a planning guide for HANA on Power has been published at .


May 7, 2015 - Posted by | Uncategorized | , , , , , , , , , ,


  1. Hello Alfred,

    Thank you for your detailed explanation on HANA on Power. I would like to add some thoughts:

    * System scalability is important and Power leads here in threads per core. But – you need to make the workload scale. This is usually the most difficult part. So the superior Power scalability does not necessarily translate into faster HANA performance – at least until it is proved by benchmarks for example.

    * Making the most of the capacity of your system: LPARs can provide this feature for sure, they are a mature technology. Nevertheless, I am not sure if SAP allows you to run HANA and non-HANA workloads on the same physical server. I’d bet they don’t. Besides, this can be also achieved by VMWare on x86 HANA platforms. Hypervisors always add complexity and overhead, so perhaps the best way of making the most of your server is to leverage native HANA features, like MCOS, MCOD or multitenancy.

    * About RAS, what you state is true for classical, basic x86. Intel E7 v2 x86 processors incorporate many RAS features, like memory quarantine in HP servers. This means that you don’t necessarily need memory mirroring. And high end x86 servers, like HP Superdome-X, go well beyond. Superdome-X incorporates the Error Analysis Engine for predictive error correction, all intel E7 RAS features, DDDC + 1 for memory (no mirroring required), PCI online error recovery, redundant crossbar fabric and many more. On the other hand, it appears that there is no hot repair capability for the E870. IBM’s solution is to use Live Partition Mobility to evacuate the box that needs repair. This means you need another server to put the running workloads on or take down the whole system for maintenance/upgrades.



    Comment by Isabel Martin | June 29, 2015 | Reply

    • Isabel, Thank you for HP’s your thoughtful comments. You are absolutely correct, benchmarks tell the story better than speculation and the latest BW-EML benchmark shows the IBM POWER8 result delivered 12% more performance than Haswell-EX with 45% fewer cores resulting in slightly over twice the performance per core than Intel.

      VMware is most certainly a supported technology, but with HUGE restrictions. 1995460 – Single SAP HANA VM on VMware vSphere in production states that only one VM is supported for production which defeats the whole idea of virtualization other than enabling vMotion and DRS. It goes on to state that only one HANA instance may be run under VMware, than each VM for HANA can only utilize 64 vprocs and 1TB of memory which is interesting since 64 vprocs translates to 32 cores unless you disable hyperthreading and the amount of memory supported by 32 cores without VMware is 512GB, so this “supported” config will likely perform terribly compared to a native HANA implementation. By the way, 2133369 – SAP HANA on IBM Power Systems: Central Release Note has no such restrictions and your speculation about running HANA and non-HANA workloads on the same physical server is not supported by any SAP Note. You are also correct and incorrect on the comment that “Hypervisors always add complexity and overhead”. Correct when it is an “add-on” and layered hypervisor as you find with VMware and other x86 hypervisors, incorrect when it is built it as with PowerVM. Technically, PowerVM does have overhead, but since it is part of every benchmark, the overhead is already included in every benchmark result. MCOS, MCOD and multitenancy are options that customers should consider, however they are mostly an accommodation to x86 based customers that don’t have a viable virtualization option.

      Regarding your RAS statements, any vendor can add RAS to those components that they provide, but you can’t add anything to those that you don’t supply. Unless HP has found the magic back door to access and modify Intel chips, any RAS features or deficiencies are part of the package. As to your suggestion that their is no hot repair capability for the E870, you might want to do some research before you make statements like this as it is simply inaccurate. An E870 may be configured with two drawers with either one being able to be deconfigured, removed, fixed and replaced without causing an outage to the applications running on the system, assuming sufficient resources remain on the other drawer to support those running workloads, similar to any nPAR or vPAR based deallocation option. But why go through all of that headache when the system is smart enough to circumvent the problem in the first place. Much as many high end auto companies supply “run-flat” tires and have computers that automatically adjust to a wide variety of error and environmental issues, POWER8 systems sense and respond, usually before a problem is even noticeable, deallocating components when necessary, allocating other spare components to replace failing ones, all on the fly and without application outage or awareness. POWER8 systems can even handle many events, including a core, L1, or L2 cache failure, a memory lane failure and many others, that can’t be predicted, on the fly and, once again, without application failure. So, sorry, not only do you not have to take down a whole system to support maintenance/upgrades, but you can avoid many of the reasons for why you would need many these in the first place by using POWER8.

      Comment by Alfred Freudenberger | June 29, 2015 | Reply

      • Alfred,

        Thanks for the reply and the healthy discussion.

        Yes, there is a BW-EML benchmark for P870 4p/40c 1TB than shows 12% performance increase over an x86 server with 1.5TB. Less memory and more performance for this 1-2 TB size range. However, I don’t consider these configurations exactly large, scalable HANA implementations. So to me the current benchmark does not quite prove that Power *for HANA* scalability trumps Intel *for HANA* when the number of processors and RAM are increased. I am not saying they won’t scale, I am saying that it is not accurate to extrapolate a 1TB HANA benchmark to a 6TB HANA benchmark, so to speak. This was my point. Also, the BW-EML benchmark shows that the bigger the DB (more initial records), the lower the performance. This is another thing to take into account – for every vendor.

        VMWare has limitations, right. Personally, I don’t see the point of using virtualization in production for big HANA instances, I think those run better on bare metal. Simpler is better. For small HANA instances virtualization is fine and provides benefits. VMWare limitations and overhead become more irrelevant on these small configurations. Said that, SAP is allowing some customers to run several VMWare VMs in production on the same server under controlled availability ( This is usually the precursor of general availability. Power was under controlled availability for some time, as well as CS900 12TB, and now both are general available.

        Thanks again for the clarification about E870 RAS. About the HP “magic”: HP has develop the code to implement all the E7 RAS features on Superdome-X, while other vendors use the default Intel code that does not provide the same level of protection. Not exactly magic but something differential vs. other mainstream x86 systems. Also, RAS features and how systems handle errors are very dependent on the OS. There may be differences between AIX and Linux on Power.

        I am not questioning Power8 systems, but at the moment I am reluctant about HANA. HANA is new for everybody, including SAP. Everybody is learning and polishing their solutions. HANA on Power is even newer, GA since May 15. It is too early to build an opinion. Time (and market) will tell. In any case, variety and freedom is good for customers, so they can pick their preferred platform.

        Comment by Isabel Martin | June 30, 2015

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: