In the first part of this analysis () I outlined the massive impact of a series of technology-driven forces transforming the IT industry. These include Open Source, Big Data, hyperscale systems, flash storage, converged systems, software-defined infrastructure, public cloud services, and mobile computing. Together these are reshaping our lives and the IT industry. In this second piece in the series, I want to look at the probable impact of these forces on the vendors.
In any revolution those at the top have the most to lose. In this case those include the hardware giants -- IBM, HP, EMC, Fujitsu, Hitachi, & Dell -- and software giants Microsoft, Oracle & SAP. Of these, IBM and Microsoft come into this with the strongest positions, which is not to say that they are guaranteed to come out unscathed. All the major hardware vendors face a major threat in hyperscale architecture, which was developed by the big cloud service providers and built on cheap white-box hardware from Asia running in a virtualized environment in which all management and other features have migrated up to the software layer. The question is when will large traditional enterprises start adopting hyperscale and how deeply will that cut into the hardware market. But even without that, software-defined infrastructure and Moore's Law promise to reduce hardware at all levels to commodities within this decade, forcing hardware vendors to compete on price and survive on razor-thin margins. When that happened in the desktop market two decades ago, all but one of the independent PC makers -- Dell -- were either purchased or went out of business. Apple was on the way out when Steve Jobs revolutionized the company with the iPod, iPhone, and iPad. Even IBM eventually exited the market it had largely created.
Of the big hardware vendors, actually IBM is in the best position, mainly because in the early 2000s it made the wrenching transition to become in large part a software and services company. Today it has completely embraced Open Systems, providing Linux on servers and mainframes and Hadoop along with DB2. This is not to say that it will not be hurt if the bottom falls out of the hardware market, it still makes a handsome income from the sales of its mainframes and servers. But it is no longer dependent on those sales for the majority of its income. It will survive and even prosper.
HP is primarily a hardware company, and this makes it more vulnerable. While it has made no announcement, it appears to be following the path IBM pioneered, focusing more on software and services. It is extracting the management layer from its hardware and redesigning it to work across multiple vendors' boxes, just as IBM did, for instance with StorWize, several years ago. HP also is the only member of this group to have announced its own hyperscale products, which seems to be a white flag flying over its hardware business. Obviously HP management sees the tsunami, the question is can it get to high ground in time.
EMC has created its own survuval strategy through diversification. It also has a strong services arm, which always has been an important part of its market presence. But those are mainly low-level break/fix services, the kinds of things that are increasingly being automated. It appeared well positioned with its purchase of VMware, but this year we have seen that falter as its competitors, particularly Microsoft, have caught up in primary functionality and offer their hypervisors at a fraction of VMware's price. It is difficult to under-price Hyper-V. Part of the question here is whether Pat Gessinger, a very smart man, can find a new winning strategy for VMware. Personally I would not bet against him or the rest of EMC's leadership. I think EMC will survive, although itwill be battered and may eventually need to merge with Cisco to play in an increasingly converged hardware marketplace.
The other three in the hardware group -- Dell, Fujitsu, and Hitachi -- are the most vulnerable. The two Japanese vendors have the advantage of being parts of huge, highly diversified companies. Those parent organizations will probably survive. However their computer divisions may not, or they may turn themselves into commodity white box hardware providers and compete in the hyperscale market. Dell is in the midst of its own transition, and it may not make it. The mobile computing trend is hitting it particularly hard, and while it has made some excellent software acquisitions -- Toad comes to mind -- these alone are not enough to save the company as a whole.
THE SOFTWARE SCENE
On the software side, Microsoft has taken the biggest hit so far. The iPad and, to a lesser extent, Android have taken a big bite out of its consumer desktop/laptop business. It was two years late with Windows 8, giving Apple a free hand in the market for too long. Today it has lost a large part of the consumer market permanently. However, the business market is a different story. iOS and Android cannot compete here in terms of providing a robust enough basic package to meet business needs. And businesses know Windows. Almost all IT shops have Microsoft skills in house. All their end-user software, including front-ends for enterprise systems, are written for Windows. The next big thing here is virtual desktop, probably delivered as a service by. But because Microsoft licenses by the device and requires a separate instance for each user, this if anything will increase its sales since companies will need extra licenses to put their Windows-based images on those tablets. And Microsoft's secret weapon is the new generation if convertible laptops that combine laptop and tablet into one device. As these take over office desktops they will push competing tablets out as business devices. Few users are going to carry two tablets around the office, whatever they may do at home.
So far Oracle and SAP have been mostly untouched by the forces of change. That will not last. Both have built their technologies on RDBMS technology, making them vulnerable to Hadoop and the Big Data market. Oracle in particular faces a problem since its RDBMS is the heart of its business, and Hadoop, while not as feature rich, is much less expensive.They both also have built their applications on the old massive architecture model, which makes their systems inflexible and slower to adapt. Going forward their main competition will come from groups of vendors and service providers who will build on Hadoop and industry standards, allowing users to pick the features they need and move quickly as needs change. Longer term both are vulnerable, and Oracle is not helped by its hyper-controlling attitude and price gouging.
SINGLE-SILO VENDORS
Below these are a large group of single-silo vendors who sell just storage, networkswitches, or similar limited product sets, including established vendors like Broadcom and NetApp and the Flash startups of the last few years. Many of these are excellent companies with advanced products. But in a market that is moving away from best-of-breed to converged hardware and virtualized environments, they as a group are all very vulnerable. The flash storage vendors like Fusion-io are high growth startups today, but their future is not that bright. Already IBM, for instance, is selling its servers and mainframes with large amounts of flash built in, and that flash is read/write. All the big vendors now have flash-only arrays. Fusion-io itself has found a market in providing large amounts of flash on the white-box servers in hyperscale environments, but it is only a matter of time before the manufacturers of those boxes start adding their own flash storage to their products. The future for most of these companies is to sell themselves to a major vendor while they can.
All of the survivors among these vendors will look very different a decade from now, and the market itself will be restructured. White box vendors, which may include HP and possibly other U.S. and E.U. based companies, will dominate a large portion of the hardware market, almost all environments will be fully virtualized, public and private clouds will be so tightly integrated that they will have no clear boundary, and differentiation will be in the software layer, which will be totally separated from the underlying hardware.
In the third part of this series I will look at the impact of this technology revolution on the careers of the people in the industry and on the economy at large.
Full Post
No comments:
Post a Comment