The reality of the cloud is not much different than what most MSP offer today. However, the physical box and fail-overs become someone else's responsibility. The content is another person's responsibility. Upgrades to continue future-proofing the hardware and environmental concerns may be another person's responsibility. The bottom line is the expectation of the user and cost. Generally, higher expectation will require higher cost.

With any first wave (or 1.0) product, the vendor "fishes." They push half-baked applications on a half-baked server in a half-baked rack space. Suddenly the users start pushing the system in ways the vendor hadn't considered. Then the vendor plays catch-up to finish baking the areas of importance. We've all seen this.

As the users consider, "what can the cloud do for me?", the vendor continues to fish and provide a bit more, trying to lay the path as the user is walking down it. Finally, reliability is directly proportional to the distance traveled thus far. The vendor often forgets this since the return on investment hasn't quite came about. However, I suggest this is crucial to the vendor's investment and strengthening of the product offering. The factors of physical location, redundancy and fail-overs play a huge part in the success of the offering and must be kept on a steady incline with the popularity. Not always popular is the cost, however. But no one said popular is always cheap. Too many failures will bring you the wrong type of notoriety.

What's this boil down to? The right software developed for the cloud in conjunction with the right infrastructure/MSP is a good formula. I believe there will continue to be a shift in how software is developed first, aimed for the cloud and then a good infrastructure company can put it to its intended use--with the fail-safes. The cloud growth will occur as verticals are developed for the cloud and API's interlink the cloud applications. In order to stick around, the MSP will need to evolve to a MCSP--a managed cloud service provider.