“If we are victorious in one more battle with the Romans, we shall be utterly ruined.” – King Pyrrhus (apocryphal): “A Pyrrhic victory is a victory with such a devastating cost that it is tantamount to defeat. Someone who wins a Pyrrhic victory has been victorious in some way; however, the heavy toll negates any sense of achievement or profit.” http://en.wikipedia.org/wiki/Pyrrhic_victory
There is no question that cloud computing is going to be a growing market. It will also be a huge market. But it will probably also be a lousy, low-profit, high anxiety market for the infrastructure suppliers. A loose analogy; Wal-Mart was also a fast-growing, huge market, but selling to Wal-Mart sucked. It probably sucked even more as Wal-Mart got more dominant.
The major cloud providers customers are going to be awful to work for (and suppliers will really be working “for” more than “with” them).
- They are going to be very few customers: Lets face it, these days “the cloud” is basically Amazon and a bunch of VERY distant number 2’s. We’ll assume that a few other players catch up (with even that far from certain). That is a painfully concentrated market to sell into.
- They aren’t dumb. The big cloud guys can afford to hire people who are as smart (or smarter) than their suppliers. The dirty little secret of enterprise technology is that its profits depend largely on most customers being the sort of gutless time-servers you would expect to gravitate toward an internal support function. That wont be the case here.
- They are going to manage (brutally) to a lowest total cost model. For a cloud provider, the infrastructure is a direct cost that sets gross margin, not some vague “administrative overhead” cost allocation like in traditional enterprise IT. That gross margin number will attract brutal internal pressure from every executive in the company. It will attract even more brutal external pressure from competition in what will be a fairly transparently priced market. That sums to a whole lot of pent-up brutality – which will be vented to foul the air of the cloud infrastructure suppliers.
- They will have HUGE leverage. Obviously, the cloud vendors will be writing make-or-break-the-quarter checks almost every quarter.
- They will have weak vendor lock-in. Cloud builds are done (and re-done) on a per-data-center-basis. With demand increasing fast, there will be frequent new builds. That means frequent “clean-sheet-of-paper” competitions as the buyer tries to figure out a lower-total-cost equation using all possible vendors. A dominant role in data center “A” doesn’t mean much if you are starting from scratch designing data center B, C, D, and so on.
- They will try to build the good stuff in-house. The equation infrastructure = gross margin = competitive position resolves in favor of in-house, proprietary, “secret sauce” developer teams. If Amazon can figure out a way to run their infrastructure a few points cheaper, why in God’s name would they develop that with a vendor who’ll just sell it to everyone else? So the vendors are going to be stuck selling semi-finished, programmable gear that the cloud players will finish tweaking themselves.
None of the above sounds particularly appealing. I’m trying to approach the whole “cloud” phenomena as a series of vignette posts vs one big dissertation, so more to follow.
