March 16, 2010

Some Thoughts on the Cloud Connect Conference

I ventured out to the Cloud Connect conference and expo at the Santa Clara Convention Center this morning. Unlike most trips to industry events where I take part in the transportation cloud, usually of the rail nature, today I was firmly self reliant with a set of four tires on the parking lot known as CA-237. Part of my reasoning was to go get out of office on a sunny day, but more importantly was my intrigue on this “market segment” known more or less as Cloud Computing. Readers know that I have a heavy dose of skepticism about the marketing fluff related to Clouds, or the often ill defined opportunity that purports to be the next greatest thing in IT. With few exceptions, vendors and the industry as a whole have historically done a poor job of defining this market opportunity. Is Cloud a market segment? Is Cloud a delivery model? Is it both? So far, this morning’s keynotes have only reinforced my skepticism, which is unfortunate. For all of its technical promise, the continued lack of industry definition clarity remains deeply troubling.

By mid morning, we had heard from a Deutsche Bank Securities analyst, a market researcher, a couple of systems vendors, a startup pro, and some ancillary folks. The analyst talked about equipment, vague notions of markets, and then hardware sales he projected were related to cloud sales. While his sales projection of $20 billion is non-trivial, it utterly lacked any definition by which to differentiate these sales from plain old enterprise hardware sales. Just what is the uniquely Cloud stuff that accounts for this expenditure? From this presenter, it seemed that Cloud meant nothing more than enterprise IT with virtualization of servers and network switches. This conveniently left out perhaps the fastest growing segment of IT, i.e. storage. This Cloud discussion was underwhelming at best; an example of a 2002 mindset fixated on server virtualization with lip service to linking those servers together. This doesn’t sound like a game changer to me and begs the question of why would VCs plow money into such an ill defined opportunity?

A speaker from Sand Hill Group presented some recent research of CIOs about their Cloud plans. It was interesting to note from this non representative sample that about 60% of CIOs are buying SaaS today, and about 30% are buying IaaS, and 10% are buying PaaS. While there is some expected overlap, at first blush it aligns SaaS with the notion of the Cloud. From the same survey 49% of CIOs said they were using the Cloud to achieve business agility, and 46% said they used the Cloud to reduce costs. This was the first rational thought I had heard this morning; businesses want to be more competitive, but saving money is cool too. In addition, 53% indicated that their organization would be deploying Private Clouds, with 43% planning to use a Hybrid solution. (I guess they will be powering their equipment from a Toyota Prius, like I often do with my laptop.)

Although the detail of depth of the methodology was not discussed in the ten minutes allotted, it strikes me odd that 60% are buying SaaS, yet 53% plan to be Private Cloud within three years. Even considering the Hybrid strategy, on the surface it would seem that CIOs are currently going outside for an app or two, perhaps more, through SaaS, but are planning to come back inside as they can. If true, this would seem to counter the assumed predominance of SaaS, let alone Public Cloud providers. But then again, without any real definitions, there is enough wiggle room to drive a Mack truck through this Cloud marketing and positioning. Realistically, aren’t most organizations, even those with a Public Cloud only strategy going be Hybrids for a really long time? Rip and replace simply doesn’t happen. At a minimum, it impacts business agility and costs money.

Another speaker dazzled the audience with some more numbers, such as the serial numbers of his Amazon virtual machine IDs. However, perhaps most interesting was the statement that Amazon spins up 83k virtual servers a day compared with 488 by Rackspace, and 181 by GoGrid. Further, it was stated that Amazon EC2 has spun up 23 million servers since its incarnation. So without much thought one could pronounce that Amazon is the clear Cloud leader; but what is really being said here? Are these servers being used for anything? Spinning up a virtual server is like starting a truck at the freight yard. You can easily start the truck, drive it around the yard, and then shut it off. But have you really delivered anything let alone created a transportation industry? Clearly not. Just how many EC2 servers have remained up to do something useful in either the short or long term? It’s hard to tell. Are Clouds simply an exercising in turning on and then off because you can? Argh!

In an entertaining illustration of my concerns about the lack of definitions, the speaker from Neustar led the audience through a sequence of assumptions that culminated in him declaring the Conflicker botnet as the largest Cloud provider on the Internet today. High bandwidth, widely distributed, available on an as needed basis, remotely located, etc. all the buzzwords of Cloud Computing combined with all of the ethical shortcomings of criminal activity. Now that’s a platform that investors will want to get onto, especially if the investors are named Guido and are after a high rate of service for their services rendered.

OK, I am cynical and remain skeptical. Although this conference was subtitled “The Future of Business Computing is Here,” there was little discussion of business, unless you counted the vendor pitches from MS and McAfee during the first morning keynotes. When I have heard well reasoned pitches from vendors such as IBM, EMC, and a few others, they have well defined scope, goals, and boundaries in what they attribute to Clouds. They more or less define what constitutes each variety of Cloud, and most importantly offer an overarching architecture that is unique and differentiated to the endeavor. From this, it is possible to have rational discussion of the advantages of the technologies, their financial and human cost impact, and the potential for enhanced IT efficiency. If only the industry would demand this from all players.

It becomes more difficult when speaking multi vendor, as the industry does not embrace a uniform set of Cloud definitions. Perhaps, even more cynically, this is by and large what the industry wants. If sufficient levels of FUD can be maintained, each vendor can continue to blindly feel up the elephant while smugly articulating their discovery of just what an elephant is. As long as everyone keeps up the charade, vendors can continue to hype and position themselves with magic Cloud dust. Unfortunately, this short term thinking does little to focus the market, and is an impediment to fostering long term growth and opportunity.

Ultimately, the right answer may be to stop looking for the Cloud market altogether. Perhaps Cloud is really just an intelligent delivery model that addresses the state of art in IT. Maybe Cloud is a process, not a product. As such, things would make a whole lot more sense than the confusing overlap of jargon and techno-obfuscation that so many undertake in the name of the Cloud. This would be a welcome improvement not only in nomenclature, but perhaps in market clarity, which would then help drive market adoption. Money tends to follow well defined paths to ROI. Why should Clouds be any different?