Don’t look now, but the recently moribund server market rebounded sharply this year. According to Ben Worthen’s recent blog on WSJ.com,
server shipments jumped 15% during Q3. This increase continues the year’s
higher shipment rates, and reverses the previous declines of 2009. Of
course, some of this growth reflects the economic improvement vis a vis the investment clampdown resulting from the global recession of 2008-2009.
However, the blog also contains some facts that point to a rapid growth of public cloud computing. “Skinless,” or servers shipped sans case, increased
(by revenue) 22.5% and 25.6% for HP and Dell, respectively. Skinless servers are primarily purchased by cloud service providers (CSP) rather than
corporate customers, and are often custom or semi-custom manufactured to CSP specs. In particular, components unnecessary for serving as cloud
processing devices are left off of CSP design specs — think USP ports, sound chips, etc. Every avoided component reduces the total server price
and, crucially, trims operational cost by reducing power use.
Looking at the growth in the skinless server market makes one thing clear: a lot of cloud computing capacity is coming on stream — and that
holds some intriguing implications for cloud users:
Prepare to be bombarded by CSPs encouraging you to use their service. Many of the new providers have to establish a name and a track record, and
they’ll be looking for reference customers. Look for incentive deals designed to get you to sign up.
Understand that many newly-launched cloud providers are learning about what it means to be a CSP. Implementing cloud computing technology is not
the same as having streamlined operational practices characteristic of efficient CSPs. Many newly-minted cloud providers are coming from the colocation
or managed service world, and need to learn how to operate in an automated fashion at scale. If you’re a customer for a just-launched CSP, expect
challenges and glitches. Drive a hard bargain in return for being an early customer, because you deserve to be compensated for providing a learning
opportunity for your CSP partner.
Begin your internal skill-building for employees. Architecting applications for cloud environments requires new design patterns, which most application
groups do not yet understand. Likewise, operating scalable, elastic applications requires new system administration and management skills. The burgeoning
field of “devops,” which I discussed earlier this year,
reflects the changes in IT operations imposed by cloud computing. Expect that your staff and colleagues will experience a learning curve as the organization
grows new skills.
In this respect, cloud computing is no different than previous technology platform shifts, each of which marched IT organizations through a painful
developmental “opportunity.” The skill upgrade is common, but the specific skills to be upgraded change, so recognize that staff will need to be educated
and allowed to learn on real-world applications.
Look forward to significant CSP price competition. One thing I remember from my MBA microeconomics class is that capital-intensive businesses are
motivated to accept any business as long as it covers its variable costs and makes some contribution, no matter how small, toward capital expense. One
has only to look at the airline business to get a sense of what this will look like for the CSP industry. A lot of cloud providers that have sunk a bunch of
capital into their offerings are going to be fighting for users, and yield management will become a necessary CSP talent.
This price competition will expose those CSPs who have high variable costs (i.e., operate less efficiently) and suffer high cost of capital (i.e., have high
fixed costs). As Warren Buffett says, “You only find out who is swimming naked when the tide goes out.” In other words, only in adversity do you see who
can be successful. Expect to see a number of newly-launched CSPs hastily retreat from the market once they understand the just how costly and
challenging running a profitable CSP can be.
Get ready for your scale challenges. I don’t mean the challenge of architecting scalable applications, addressed above. I mean the challenges of scale in
terms of an explosion of the numbers of applications you will be expected to build, obtain, operate, and manage. This week I’m in the U.K. running a cloud
computing workshop, and took the opportunity to attend a London CloudCamp. Simon
Wardley of the Leading Edge Forum kicked off the event with an overview of cloud computing. He
has an historical bent, and explored the rise of other commoditization (or, as he spells it, commoditisation) events in history. Interesting and entertaining.
Vitally important, however, is Simon’s discussion of Jevon’s Paradox,
named after a 19th century British economist. Jevon, in examining the coal industry, observed that (as summarized by Wikipedia) “technological progress
that increases the efficiency with which a resource is used tends to increase (rather than decrease) the rate of consumption of that resource.” Put more
bluntly, this means that as resources (i.e., coal or computer processing via cloud computing) get cheaper, we use more of them.
This implies that we will do much more computing (implement many more applications) because the cost of applications, thanks to cloud computing,
will plummet. The increased number of applications (I expect the number to explode) will cause many dislocations in IT organizations, which have
processes in place appropriate to much lower numbers of applications. Start thinking now how you will respond to a world of 10 times the number of
applications in your portfolio — and don’t imagine that the explosion in numbers won’t happen. It will.
Being on the cusp of a revolution is exciting and scary, all at the same time. Behavioral change is forced on us, which is uncomfortable in the best of
circumstances, and painful when the change is unbidden. Reading the tea leaves of server shipments gives some forewarning of what the IT landscape will
look like in the next few years, and gives one the opportunity to plan for a new world rather than find oneself dumped in a new setting. Take advantage of
this portent and exercise prescience. You’ll be glad you did.
Bernard Golden is CEO of consulting firm HyperStratus, which specializes in
virtualization, cloud computing and related issues. He is also the author of “Virtualization for Dummies,” the best-selling book on virtualization to
Follow Bernard Golden on Twitter @bernardgolden. Follow everything
from CIO.com on Twitter @CIOonline