Vendor Comparisons: The Cloud Wars
I believe that vendor comparisons at this level are almost meaningless. There are differing cloud markets in play, namely, consumers, Small-Medium-Business (SMBs), and Enterprises (I could, but won't, bother to break down Enterprises into multiple sizes). Most of the vendors mentioned play in one or more of these markets, and their success and credibility is different in each. Some of these vendors are struggling to define where they want to concentrate their marketing efforts. Plus the vendor landscape is evolving through acquisitions, mergers and partnerships. Verizon/Terremark, NTT/Dimension Data/OpSource and CenturyLink/Savvis are good examples of Telco's trying to expand their horizons into the Cloud world, while you have AT&T trying to do it internally. Don't forget Deutsche Telecom and T-Systems who were focused more on the full outsourcing spectrum, but have developed a fairly powerful cloud offering with their "Twin-Core" data center concept. Some of these players are struggling to define what they want to be when they grow up. Integration of the telco and cloud businesses has proven to be elusive.
And, of course, you can't completely ignore the traditional outsourcers who are struggling to get into the cloud game (CSC, Dell/Perot, IBM, HP, XEROX/ACS, etc.). These providers are used to long-term, fixed-price contracts. Pay-for-usage models are counter-intuitive to them.
It will be fascinating to see the market battles in the next few years, as the core cloud players (Amazon, Google, Microsoft, Rackspace, etc.)try to move up the enterprise food chain, and the above vendors try to compete. The core players have established a powerful base that provides them with scale, and therefore, pricing advantages. But that consistently have problems, as @accident rightly pointed to in his/her comments. Have the biggies outgrown their ability to manage their environments? Many of the outages are actually caused by changes being made during mid-week, prime hours. The impact of such changes often start off minimal, but the entire environment then degrades into network traffic storms trying to "catch-up", which, in turn, affect many clients. And, when they have problems, their client communications are horrific, if not, non-existent, usually, a blog site that is updated every 3-12 hours. Enterprise-class clients will not stand for that kind of service. The traditional outsourcers and telco's know how to deal with client communications, primarily through dedicated Account Executives and Service Delivery Executives. I always love to see the defenders say "enterprises have outages, too". The difference is that when an enterprise has an outage, all hands are dedicated to recovering that enterprise's most critical systems. The cloud providers, by definition, can't dedicate their resources to one large client. There are no "dedicated" resources. The outsourcers have a middle ground, namely, having resources working on the general problem, plus dedicated resources, focusing on the individual clients.
The cloud industry, for some time, has discussed the concepts of Public, Private, and Hybrid Cloud models. Large enterprises need (at least, they believe) all three. Few of the above players have successfully incorporated all three into their quivers under a single management umbrella. Some are getting very close through acquisition of "cloud management" capabilities.
I have rambled all over the place (I apologize). I think, someday, we will look back on these coming years as the "Cloud Wars". Who survives, how long it takes, will customers be the winners, are all unanswered questions. My bet is, this will be analogous to the battle between mainframes and distributed systems, with the declared death of the mainframe back in 1990. We will continue to see remnants of all models for a very long time. Should be fun.