Cloud definitions and economics

Like Phil Wainewright, I attended the excellent CloudCamp in London several weeks back. Unlike Phil, I walked away with more questions than answers but with my enterprisey antenna on high alert.

Two things struck me. Rhys Jones of Royal Bank of Scotland's assertion that while cloud computing sounds like a good idea: "We're not going to do it - at least for the time being." That despite Phil's angle:

Rhys Jones of Royal Bank of Scotland, said that providing aggregation of multiple clouds was one of the key things that "cloud vendors have to get together and do for us," adding that "a big provider could really clean up" by becoming the linchpin of cloud federation. Oh, he also foresaw increased use of SaaS by enterprises for non-critical apps, which also warmed my heart.

For Rhys, issues around risk, compliance and control figured largely in his thinking. That was echoed by a representative from Credit Suisse who added further color by quizzing Microsoft on how it plans to manage patch releases and regression testing in cloud environments. There are no clear answers today.

This will be music to the big on-premise application vendors that are wedded to a business model that demands up front licenses which aaS does not, relying instead upon an annuity model in the form of monthly fees.

The second thing that struck me is that all things aaS seem now to be associated with cloud computing. That must be horribly confusing for any decision taker. It was for me until I read today's post by James Governor entitled:  Three Better Ways To Tell Its not Cloud Computing?

James has been running a series on this topic, attempting to put different angles (mostly technically focused) in an effort to distill what it means to be a cloud computing environment. He captures it well when he quotes:

Staffing Software Talk takes a very hard-boiled and boiled down approach:

For those of you not already plugged into this latest addition to tech jargon, you can read more about cloud computing here. But actually I wouldn’t waste your time. If you’re over 50 just think “service bureau”. 30 to 49, think “Application Service”, and 20 to 30 “SAAS (Software As A Service)”. If under 20 then cloud is everything you need in your online life - amazon, ebay, facebook, myspace, gmail.

For this simple soul, the above quote is a great way to characterize positioning which I can put another way - cloud computing is anything you want it to be as long as it's over the Internet. That of course is facile because as speakers at CloudCamp pointed out, clouds can be internal to the enterprise. But it is indicative of the variations on a theme that seem to litter the current literature on the topic.

While definitions are always useful, I sense that if we're not careful as an industry, we run the risk of plunging the topic into the trough of disillusionment before its time.

Putting my buyer's hat on, I want my computing infrastructure to be available on demand at the lowest possible cost. I want to drive those baseline cost efficiencies into as many of my applications as possible, but not at the expense of sacrificing or endangering security, or my ability to run a compliant set of applications. In many scenarios and especially those that are regulated, operational code 'ownership'  is important.  Overall then, that should have me firmly leaning towards the promise of cloud computing but holding something in reserve as suggested by Rhys Jones.

My concern is that with the siren call of all things aaS rendered in the cloud, I may become unwittingly dragged into believing the 'Cloud/s' do/es represent a computing Utopia. That is not a proven statement by any stretch. Even Microsoft, which is pitching Azure would not go that far (as if they would anyway) because as was said at CloudCamp by Microsoft and others, these are very early days.

However, I don't think it is too early to start articulating the business value and how that will play for both buyers and sellers. In all the discussion on this topic I don't see that emerging clearly.

There is good reason for that because in order to get to a reasonably well understood economic description, a lot of things have to fall into place. James emphasizes standards and he is right to do so. The trouble is that history tells me standardsd is something of a sinkhole. I sincerely hope I am wrong on this occasion but I do wonder whether Simon Wardley's vision of open standards as the essential pillar for portability and interoperability will carry the day. Back to the business.

When I read Larry Dignan's analysis of Salesforce.com's results I had one of those occasional financial itches that makes me want to go back in time. The comparison is with the PeopleSoft of 1997-98. That was the time it hit the $1 billion sales barrier and needed to move forward under fresh leadership. The difference was that at that time, PeopleSoft was reporting profits far healthier than Salesforce.com at about the same age.  There are two crucial differences between the PeopleSoft of yore and Salesforce.com today.

PeopleSoft was riding the client/server mega trend accompanied by large up front license revenues. As an aside, it was also engaged in some creative accounting that were subsequently shown to represent cost switching. Today however, we see that Salesforce.com with its $1 billion in annual sales is attempting to build out a platform for other cloud based services. Many will argue that aaS economics are very different and that any comparison is spurious. I get that but I'm not sure Wall Street would be so sentimental.

Even so, it is far from clear whether Salesforce.com will be able to hit mega profits anytime soon with  the accompanying free cashflow. (Check the persistent questioning on this point.) As Phil Wainewright points out in a discussion about spending money by saas companies:

The key to understanding the formula is to recognize that SaaS companies bleed cash with every new customer they acquire — the complete opposite of what happens when a conventional software company lands a new account and pockets a huge upfront license fee. Especially if, like Omniture, the application requires significant infrastructure investment but the subscription is billed monthly.

Gulp! Just to add more confusion, Zoho is yet to become cash positive but expects to reach that milestone in the next year. It however has the backing of a cash machine. Smaller players I know are doing well enough and generating free cash flow plus profit but will not command the attention of enterprise buyers anytime soon. Perhaps it is the case that to play in the enterprise space means huge uncertainty over long periods of time. That's hardly a recipe for confidence in a down economy, or any economy for that matter. But then Salesforce.com could throttle back on sales and marketing costs with an immediate impact on the bottom line but almost certainly at the cost of reduced top line numbers.

If cloud computing is to be the business success so many in the industry anticipate, then as a buyer, my primary concern must be to know that my providers will be there in 5, 10 and 15 years' time. All the evidence I've seen so far suggests it requires companies capable of making massive investments to develop a viable cloud based ecosystem. I'd rather pay a wee bit more now to ensure vendor survival than be lured into what often seems like a race to the bottom on pricing in the deluded expectation that somehow it will all work out.

As a final note, I recall that 1997-98 period very well. PeopleSoft was riding high and looking good. Today, it is a footnote in computing history although its legacy lives on through Oracle. The parallels may not be quite the same but there are lessons for us all.