What will power IBM Power servers?

There were few clear answers to IBM's Power server problems in CEO Ginni Rometty's letter to shareholders, and they need to come soon.

In her annual letter to shareholders this week, IBM CEO Ginni Rometty certainly accentuated the positives of 2013.

She spent ample time citing the progress made in the company's cognitive and cloud computing initiatives, even managing to sprinkle in a little positive financial news about the diluted operating earnings per share reaching a new record last year.

What she spent much less time on were the disappointing aspects, most notably the spiraling fortunes of IBM's Power series of mid-range servers and its storage products.

Rometty is justified in touting the advances made by Watson and the new groundwork laid down for its cloud computing strategy now anchored by SoftLayer and what both mean for the future. But IT shops, particularly those looking to launch new cloud environments this year, need more concrete details sooner rather than later about what the company plans to do now about its flailing Power series.

Many IT shops have sunk significant investments in these systems and data and must know if it is worth their financial while to stay with the Power series or migrate to less expensive Intel-based servers. And the fact IBM will deliver the Power 8 chip next month puts more pressure on their decisions.

IBM could end up where it started in the hardware business almost exactly 50 years ago, with just its mainframes.

Unfortunately, Rometty didn't offer much guidance.

Rather vaguely, she said IBM hopes to shift its hardware business toward "new realities and opportunities," as well as toward Linux. This, she says, follows the lead of the company's "successful" mainframe business.

It is true that IBM's mainframe business throughout 2013 saw a remarkable resurgence thanks to new systems, but I am not sure how much the Power series can ape the success of IBM's Series z. While both systems can attribute some success to bundling Linux, the mainframe is much less dependent on the open source operating system. Mainframes and Power systems also play very different roles in most corporations.

Underlining its commitment to the Power series, Rometty points to the sale of its Intel-based server business to Lenovo. Last year's sale of the lower-end systems is consistent with the company's long-term strategy to abandon lower-margin businesses, as it did with its desktop PCs, hard disks and retail products. This makes sense given IBM's corporate overhead, but IBM users who migrate off the Power platform and onto an Intel platform will be going to an IBM competitor.

If IBM puts a lot of eggs in its Power series basket, and if it stays true to its mission of "remixing to higher value," as Rometty wrote, that eliminates the possibility of significantly lowering prices to increase sales.

Maybe she figures the ever growing demand for big data, cloud and mobility solutions will provide ample opportunity for both the Power and System z servers without one trampling the market opportunities of the other. The Power ship has yet to rise with that tide.

In conversations with IBM executives at the company's Pulse conference last month, it became clear the company bets that SoftLayer can not only lift the fortunes of its cloud business but it's hardware, too. By tailoring a range of key open source software, as well as migrating IBM’s most popular applications to SoftLayer, which will also be optimized to take best advantage of the Power systems architecture, IBM can deliver compelling tools for large companies.

That made sense five, even 10 years ago. But in this overly obsessed price-sensitive world of IT, cost savings often trump a system that represents a good bang for the buck.

So despite Rometty's assurance that IBM is "not exiting hardware, and will remain a leader in high-performance and high-end systems," IBM must communicate a compelling set of reasons that bolsters users' faith in the future of the Power series. Otherwise, IBM could end up where it started in the hardware business almost exactly 50 years ago, with just its mainframes.

Dig Deeper on IBM system z and mainframe systems

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Will you continue to invest in Power servers?
It is one of the MOST dependable systems on the planet. It is a CONSISTENT performer, especially if you are running Oracle's JDEdwards ERP software. We have had I-series systems for the last 12+ years and they keep going, and going, and going..... Keep the I-series platform humming. Over the course of time, you will definitely spend less $$$ compared to a Windows server.
because they are just GOOOOD!!
Have signifigant investment is software that is constantly being upgraded to accomidate changing business requirements. Secondly, there is not a Intel system that can compete with the i on many metrics.
There are always two kinds of market. One set of customer look for consistent , reliable , secure setup of system and other set looks for cost effective solutions. May be because the hype created by cloud is pulling every one's interest towards that , but my gut feeling is stand alone servers has got its own market and it will continue , probably the share might get reduce but not its reliability. Sooner or later , it will be understood by many customers and there are very good chances of stand alone servers picking up its market again.

Good luck for Power8 , I hope it rocks !
I will invest on POWER system as they are cream in the market ..
It is just about technology... you can find servers everywhere, but not technology on every server even pushing IBM team members.
Lack of clear future roadmap by IBM, cost
In the brazilian market IBM still presents an competitive advantage.
Power is not going anywhere, but up... Like IBM, both have a very long and rich legacy; and will only keep getting better with the planned release of the new Power8.

In fact, at Servergy, we believe so much in Power and Linux that we have bet our company's future on it - by creating a whole new class of clean and green "I/O Intensive" PowerLinux server for scale-out, that complement IBMs "Compute Intensive" scale-up servers.

IBM is a great company with great people, great technology and lots of resources. The company has clearly shown its strong long-term commitment to Power and Linux, as well as reinventing itself, as it has so many times before. The future looks very bright, indeed, for both Power and IBM...

Bill Mapp,
Chairman and CEO
Brick solid reliability and performance. The thing just works. Period. Almost never goes down, user-friendly, virtually hacker-proof. Why would we want another lesser box? The price of not having the above exponentially outweighs the "cheaper price" of non i5 operating systems and platforms. It is the '57 Chevy of the computing world; a '57 Chevy with a turbo! Even the low-end boxes are scary fast. What I cannot fathom is why IBM is not advertising and singing the praises of this box to the skies. I've been involved with the IBM Midrange for nearly 30 years and there is not a more stable reliable platform for business out there. During all that time, IBM has stupidly failed to advertise this box and its predecessors. Its OS is so far advanced than any other on the market that they could dominate the computing industry if they'd wake up and understand how superior it is. Why does CEO after CEO ignore this jewel? Please, IBM, don't be stupid. Please IBM, don't kill the golden goose! Advertise this box. Tell the world what it can really do and watch the cash roll in. In my opinion, IBM should fire any CEO that doesn't leverage this technology to the max
It is definitely not value for money anymore. Further, I believe the IBM has lost its core values. One has only got to visit a university (as I did last week to present my company's credentials) and listen to the undergraduates who think that IBM does not relate to the world anymore.
IBM has consistently delivered on their Power road map and should continue to do so. If your applications require serious RAS that only high-end system can deliver, why look anywhere else?
yes, but in smaller scale, meaning only as platdorm for Oracle db servers. app servers like websphere will be migrated to x86 and Linux
Gone are the days for proprietary hardware and software. We are moving to x86/Linux as replacement for POWER/AIX.
Because they are still the most advanced servers available to mission critical applications.
Power systems provide the ideal platform and capabilities for big data and advanced analytics and when coupled with linux, provide an extremely performant cost effective platform for the future
No. Power servers are just too expensive and now significantly worse price/performance compared to SPARC/Solaris and with Oracle now investing massively in hardware, running Oracle SW on IBM Power no longer makes sense. Oracle SW ROI will clearly be best running on Oracle HW.
It depend on is IBM committed to Power and how it's position against the up and coming Intel base server
No faith in IBM as a hardware company now.
Depends where IBM goes with it. A clear path is needed.
The provide functionality and security I cannot get with other systems. We have an installed base of critical applications here that get more power out of power than out of intel.
it is the most stable compute platform in the industry today and runs most of the mission critical applications inside most companies today.
Current leadership at IBM does not understand the HW business. It has been this way since Palmisano took over. Looking at other companies (Apple, Lenovo etc), it is clear that there are ways to be successful at all levels of the HW business. Innovative, leadership products bring sales. IBM has no clue. Further, the relentless pursuit of EPS is reducing the overall experience of GBS as IBM lays off older more experienced consultants to cut costs and attempts to replace them with offshore resources. More agile consulting companies will hire these people and eat IBM's lunch in this space over time. Even IBM's software has very few leadership areas. As much of it is tied to HW sales it will eventually lose market share as well. In the cloud space they will easily be out priced by Amazon, Google and others. I do not see a bright future for IBM.
We are very closely tied to the i operating system. In addition you have to look at cost of ownership, staffing, cost of running, and other factors. Reliability compared to a Windows based system is not even in the same realm. By far, a much better system
Confusing day one cost with TCO is one of the things driving the server market towards ever "cheaper" hardware (seemingly ignoring the advice of Dr. Deming). IBM has told a very compelling story of consolidation, virtualization, and TCO gains ever since they debuted POWER7. Running AIX on PowerVM on POWER hardware drives a large but largely intangible benefit of having one vendor for the stack from processor chips through O/S. Even their fiercest competitors tend to respect their hardware and virtualization technology from a licensing perspective.

The pricing around the low end of the POWER line and especially the Linux-only POWER systems is pretty attractive for people who know about it. The ability to manage POWER and Intel hardware and virtualization technologies from a single pane of glass in the PureSystems arena is also a powerful draw.

Maybe IBM just needs to acquire some of their competitors' marketing people to get their message across...
What is clear is that you don't really paint a full picture of IBM's HW business situation.

First, IBM's HW business has not grown in the last 13 years, looking at IBM's latest annual report, page 7 top left chart.

While it has grown significantly in SW and Services, IBM Hardware business has been flat to declining.

Secondly, you don't even mention Oracle? IBM's #1 competitor in the hardware space, besides the myriads of x86 resellers, is Oracle with the acquisition of Sun almost 4 years ago. Oracle has also taken the #2 spot in worldwide software sales, from IBM.

If you recall back when Oracle acquired Sun, Oracle made a public statement "We're in it to win it. IBM, we're looking forward to competing with you in the hardware business"-Larry Ellison http://www.oracle.com/us/features/suncustomers.html and since then, Oracle has invested $billions in advancing its hardware products, and especially its SPARC and Solaris technologies.

In 3 years, Oracle has released 5 generations of SPARC processors and systems, and has increased the price/performance of SPARC by roughly 40x in that timeframe. Yes, that’s a significant jump, where SPARC/Solaris was practically last place back then, and today, continues to hold over 20 world record benchmarks in both performance (throughput and response times), performance/CPU and especially Price/Performance.

Today, Oracle is #1 marketshare leader in "Integrated/Engineered Systems" according to both IDC and Gartner (Oracle was the largest supplier of integrated platform systems with $181.7 million in sales, or 45.6 percent share of the market segment, according to IDC’s latest findings.

And in the UNIX space, where IBM was leading for last several years, SPARC has shown amazing growth and is clearly impacting IBM's Power as well as Mainframe business. In its Q2FY14 earnings call, Oracle reported double-digit growth in bookings for its Engineered systems’ division, with triple-digit growth in revenues from the high-end SPARC SuperCluster line of servers.

So up until 3 years ago, IBM didn't really have any significant competition in UNIX or high end (including Mainframe) systems, with Itanium dying rapidly and Sun financially hemorraghing. IBM was able to charge what it wanted.

Today however, Oracle has shown that it can deliver superior system performance, at a 2x-3x or higher cost savings, forcing IBM to offer 80%+ discounts, which has rapidly destroyed its revenues as seen from the last 5 quarters of double-digit revenue declines. If you don't believe me, just search google for the keyword "SPARC T5 world record" and you'll see why. Heres other references.