|HOME||PUBLIC LIBRARY||ANOTHER PERSPECTIVE||INFOPERSPECTIVES||CONTACT|
In July, IBM announced its next generation of mainframes. At first, IBM will be able to sell every box it can build. This could continue for a year or more. During this sales surge, IBM's hardware revenue should rise. So, too, will intake from mainframe systems software and middleware. IBM's services operations should enjoy a nice lift as well. IBM's employees may feel more secure, its shareholders more confident, its customers pleased. Even curmudgeonly commentators will produce more prose of praise than pans.
In the past, IBM has stimulated forward migration of its mainframe base by pricing bundles in attractive ways while at the same time making standstills more costly. The method is pretty simple: IBM sells new systems or upgrades under terms that provide hardware, software, maintenance and services financed in a time-limited arrangement. When the deal reaches the end of its contractual term, some discounts, notably price breaks on vital software, end. Costs rise, and customers cannot easily get terms on installed computers that are as favorable as those on similar deals built around next generation systems. This marketing strategy has worked pretty well for quite a long time. It will probably work just as well for the next generation of IBM mainframes, and for subsequent ones, too.
On the other hand, this strategy is predicated on the belief that customers will continue to run their legacy workloads on IBM mainframes using a lot of IBM software. Users that replace legacy apps with new ones running on alternate platforms, however, may not stay on the course they had followed in the past. They may divert some of their resources away from mainframes to other platforms or to cloud-based services. Just as the key IBM mainframe strategy has persisted for quite some time, so, too, has the movement of users' workloads to other platforms, or to the mainframe-based Linux environments that may be more advantageous than traditional mainframe operations confined to IBM's proprietary operating systems.
The upshot, notably for the past several years, for the period during which IBM's boss has been Ginni Rometty, has been a mix of good and bad news of IBM's revenue, profit and mainframe installed base. On the positive side, loyal mainframe shops enjoy the use of systems they can trust, systems that can grow to accommodate vast workloads, and systems that are understood by corporate computing staff, by outside software providers and by top consulting groups including IBM's services teams. On the negative side, mainframe systems may be viewed as too costly, tempting customers to reduce if not eliminated their commitments to Big Blue's offerings.
The way this is likely to play out in hardware, if the past is prelude to the near future, is that most mainframe shops will continue to use IBM z systems. The most dedicated users will move from old to new platforms as their bundled deals expire, signing new deals that also typically run for three years. This will provide IBM with a significant volume of sales that include both while new machines an in-frame upgrades of installed processors. Any whole processors that are removed by IBM end up in a pool of equipment that IBM can remarket or, more likely, put to use in its services groups' data centers. But not all customers will renew or expand their installed clusters of big iron. Some will cut back, moving work to other platforms. In some cases these other platforms will be IBM Power boxes, but often work that is taken off mainframes will end up on X86 hardware. That hardware may come from HP, from Dell or from other suppliers of compatible equipment. If the customer moves from in-house machinery to hardware at a service bureau, the user may actually not know or care who makes the underlying processor hardware. Only one thing will be certain. If work moves to X86 equipment, it won't be on a box with an IBM label. But the customer's new cloud platforms could still belong to IBM; IBM has become an enthusiastic player in cloud computing.
For the past several years, the overall trend in IBM hardware sales has been down. The company's growing participation in cloud offerings has risen. But so far the rise of cloud hasn't offset the decline of hardware-based sales. This has not only been affecting shipments of computing equipment, it also is reflected in sales of IBM software.
When mainframe customers boost the power of their installed systems they generally must pay more for the software they continue to use. They may also add new products, such as middleware that helps customers manage workloads. Software is very profitable for IBM as it is for its competitors. But this cuts both ways. If the sales volume of IBM software declines, corporate income takes a licking. There is a lot of pressure on IBM to keep installed software alive at customers' sites, to find new customers for its various software products and to get users to add products and features to installed software products.
As a new generation of hardware gets deployed, IBM will focus on customers whose software needs are growing. It will invariably get additional customers to adopt its software products. But it cannot do this the day it installs new machines. For the most part, customers who upgrade mainframes start by running their existing workloads. Once they are satisfied the new machines are stable and after they have examined the way their new boxes run legacy workloads, they will move up and forward. New mainframe installations pretty much guarantee some new software sales (along with lots of renewals of existing software contracts) but the financial benefits may not appear for a while. Machines that will first be shipped late this year may not produce a software sales upturn until next year or later. The money is coming, but the users' checks will still be in the mail.
What this means for people watching IBM's top and bottom lines, and for those who depend on IBM's financial health for their own careers well-being, is that even a very well received new generation of mainframes cannot yield instant benefits when it comes to software revenue.
A similar picture is likely to emerge in services that are tied to software use. Customers who upgrade their mainframes, systems software, middleware and applications may need a bit of help moving from old to new platforms. But the big jobs for services providers will come later, as users, satisfied their new hardware is satisfactory, press to make the best use of it. Making this even more complicated for IBM, and for customers, too, is the rivalry among mainframe services companies. IBM's competitors are pretty good. They may work for less than IBM, they may do equal or better work, and they may respond even faster than IBM to situations in which quick reflexes win over customers. All of this takes a while to unfold after a customer moves to new hardware. In services more than other segments of the market, the process can be slowed by the presence of competition. It can take quite a while for a customer to make up its corporate mind. Even shops that are overwhelmingly predisposed to get services from IBM often feel obliged to listen to the sales pitches of other companies and to entertain bids that could result in lower costs and higher levels of service regardless of the ultimate outcome of negotiations. In other words, this year's mainframes sales, and next year's, too, may not show up in the revenue of services providers for quite a while, possibly not until 2019.
Notwithstanding various time delays that are inherent in the hardware, software and services upgrade processes, new mainframes are bound to be good for IBM. Conversely, weak hardware sales make it very hard for Big Blue to improve results from related products, which the possible exception of hardware maintenance. So, as IBM rolls out new big iron, the company's downturn in hardware sales will ease and perhaps reverse, the promise of future improvements in software sales will gain substance and the opportunities in services will begin to blossom.
It remains to be seen whether IBM as a whole can report an upturn in revenue, overcoming the persistent slippage in its top line that has gone on for more than five years. The company has a very complex mix of activities these days, some of them at times in conflict with others. Cloud services, for example, may compete with systems sales, with software sales and with glass house related services. Watson-inspired AI technologies may compete with legacy computing cultures when it comes to getting top management attention and, consequently, resources. As always in technology, the new can be an adversary of the established and a deadly enemy of the traditional.
In the end, Big Blue's big iron boost may please those who are tied to the mainframe computing culture, but nonetheless frustrate those who would prefer IBM to pour all its energy into alternatives, even if that redirection of resources comes at the expense of venerable mainframe business and its software and services siblings.
— Hesh Wiener July 2017