Mad Dog 21/21: Server Makers Shift Shape, Stressing Legacy Customers
April 1, 2013 Hesh Wiener
Server vendors are doing a lot these days, but one thing they are doing less of is making servers. IBM, the biggest, richest, and smartest of the bunch has long since become a services and software supplier; hardware is a diminishing part of its trade. Dell is on a similar path, and trying to extract itself from public ownership to duck the requisite transparency and restrictive rules of conduct. Hewlett-Packard? Oracle? Who can say what’s going on in those madhouses? Still, it is safe to guess change is underway. Where does all this leave customers? No place comfortable.
One company that seems to have a pretty accurate understanding of the information technology market is SAP, and it shows no inclination whatsoever to get into the hardware business, at least as far as equipment for its customers is concerned. Even if SAP becomes a hardware company to build its own cloud computing centers, the way Amazon and Facebook and Google, and other lively enterprises have, it won’t build anything it would offer to end users.
Of course, as popular as do-it-yourself data center development might be, it is not the only viable option. Some big cloud app companies prefer to run their services on hardware, middleware, and other stuff they buy rather than build. There are examples of such companies doing well, pleasing customers and still remaining buyers of servers and software from the computer industry. They haven’t suffered because of their decision to build on top of hardware and software available on the open market. Salesforce.com is an example, and one that seems to be doing quite well, its shares up 50 percent from a low hit last summer. The superbanks are all pretty much buyers of servers, systems software, middleware, and applications, too. Even PayPal, a pseudobank (at least in the USA, but technically a bank in parts of Europe) lives on X86 boxes that are standard enough to run some kind of Red Hat Linux that lives inside virtualization code provided by outside suppliers (currently VMware but in the future it is expected to be stuff from OpenStack).
The thing that many buy-not-build user organizations have in common is reliance on big ticket application software, such as SAP. In addition, companies that continue to invest in legacy systems particularly those wedded to proprietary environments from IBM, HP, or Oracle (and we realize that Solaris fans, like AIX shops, imagine that their favorite Unix variants are standard) may be thinning out, or growing slowly as the computing world around them grows more rapidly. These user shops may have a long history of developing their own apps, customizing and purchased apps or linking a collection of off-the-shelf apps in special ways.
Still, as we reported, the market for servers sold to end user organizations is shrinking. Server vendors may eke out some growth this year and next, but only because growth among computing firms provides a big boost to total server sales. The cloud factor mainly contributes to shipments of X86 servers. There isn’t a big and booming market in other kinds of servers, whether based on IBM mainframe or Power, Oracle Sparc, or Intel Itanium chips, among the cloud companies. If the cloud computing demand pattern changes to include more servers in addition to X86 types, it seems likely that those other servers would be based on ARM chips or, in the relatively esoteric supercomputing realm, machines that use graphics chips, such as those from Nvidia, to perform scientific calculations.
Server vendors study market trends very carefully and try, like duck hunters, to aim their product development strategies ahead of the market, or a least point their plans at places they think the market will soon be most lively. These vendors can to a considerable extent count on the existing customers to keep buying more of the same . . . as long as they retain their established applications and their businesses grow.
At times it may seem as if the server vendors take customer loyalty or lock-in for granted, but the vendors are actually not that cynical. The vendors know that they cannot push on a string. They cannot get their customers to build up legacy systems faster than their revenue grows. And when it comes to really new applications, the legacy vendors have found that the allure of classic computers may seem dim when compared to the economic charm of X86 systems and, these days, cloud services.
The server vendors all offer cloud services, and IBM in particular has been making quite a lot of noise about its offerings. But IBM has a very large vision of computing and appears to be interested in managing electric meters in a megalopolis but not too keen on fighting it out in the low end where cloud LAMP systems (and to a lesser extent Windows-based systems) can be economical, flexible and satisfactorily secure.
End users who take the time to read about the accomplishments and intentions of IBM, Dell, HP, Oracle and others in their annual reports can see that none of these companies believe their future lies with the business of supplying and supporting servers in end user glass houses. More and more the focus of these server vendors has shifted to resemble the cloudy offerings of Rackspace Hosting, Amazon Web Services, Google Compute Engine, Microsoft Azure, and others. To further complicate matters, companies that currently offer payment processing services, such as PayPal, Square, and VeriFone and suppliers of basic bookkeeping software like Intuit are beginning to see a very large opportunity in the space now occupied by the software companies that sell costly and comprehensive accounting and business management systems.
As end users digest the unavoidable notion that they are no longer the only target of their server vendors and in fact might not even be the primary target they understandably feel hurt. It is very difficult for computing professionals whose lifelong careers were shaped by vendors who wooed them to adjust. Each individual user may feel his company is just as interested in and devoted to the platforms in that corporate glass house as ever. And that may indeed be the case. But the user base as an aggregate and the larger market that includes not only traditional users but cloud companies and myriad new (and often tiny) businesses that buys computing treats the legacy portion of the business as only one slice of a larger pie . . . and a shrinking slice at that.
The more intrepid suppliers of business applications are still trying, wherever possible, to make their offerings richer, better, and more user friendly, just as they always have. In addition, many vendors with roots in the end user glass house market are trying to create cloudy versions of their software that will reach new customers and also enable existing customers whose computing plans are under considerable pressure to segue into serverless (or less server-dependent) computing. It is not easy, because the skills suppliers of traditional applications have may not be the same skills that yield irresistible products in the new world of mixed client computing. (By mixed client computing I mean computing that works on PCs, Macs, smartphones, tablets, and basically anything that can run a browser possibly include a snazzy television set.)
But whether end users or application vendors or server suppliers like it or not, change seems to be coming. The information technology teams inside businesses are starting to discover, by trial and error, how to mix legacy applications with emerging ones. The IT specialists are discovering (and rediscovering) which in-house (or bought and customized) applications are indispensable and which ones are merely familiar. They area also learning to adjust to conditions under which the way a particular computing capability is evaluated might change, possibly because customers and end users expect something different, possibly because new alternatives have emerged.
It seems that a big change in the computing climate in which small and medium businesses now live is coming, perhaps even more surely and in a more unsettling way than global warming.
If the nation of app developers that today seems to be concentrating on the fun and social smartphone and tablet applications markets takes an interest in more mundane bookkeeping and business support, a number of killer applications are likely to emerge. What they will kill is the dominance and prominence of legacy business software. What has happened to an impressive extent in banking applications for mobile clients and airline applications, and hotel chain applications and shopping applications is going to take happen in the midrange. Consumers, at least the ones with smartphones and tablets, are already in a world of apps, apps, apps. The largest institutions, from the banks too big to fail and the airlines that are serial failers on down to premium grocers serve these individuals and small business. The corner shop taking payments with Square and making deposits by photographing checks has leapt ahead of the small manufacturer or service firm that is clinging to decade-old payment processing terminals (and their associated policy nightmares and incomprehensible fee structures).
Any minute now some cloudy business services company, maybe even IBM although we think Big Blue will stick to Smart Planet size projects, is going to show you (and your competitors) order entry apps or bookkeeping apps or billing apps or payables apps that just plain work, seem friendly, come with good security and credible safety guarantees or insurance (or an owner such as Amazon or eBay that has pockets so deep even lawyers can’t reach the bottom). Then, with due reluctance and trepidation, you will re-examine how your company’s business is conducted. If you can do it better yourself, as you have been, you’ll see why. If you can’t, you’ll see that, too.
This probably won’t be fun. Transitions rarely are. Usually they are threatening. But quite a few corporate IT folk will discover that their software vendors and possibly even their server vendors want to be part of the solution. They know that if they aren’t part of the solution they’ll be part of the precipitate.