Cloud computing

Cloud computing is one of the most overhyped phenomena to have hit the IT industry in a long time.

Cath Everett, ZDNet.co.uk

There seem to be three main reasons that are commonly used to justify cloud computing. The first is that it provides a cheap and easy way to create IT infrastructure. This makes it appealing to smaller businesses, which seems to be the set of customers that like cloud computing the most. Maybe the fact that they might only need the equivalent of half of a server for something makes cloud computing appealing to them when they don't want to pay for the entire server. In any event, this claim seems like a reasonable one.

Another reason that is used to justify cloud computing is that can provide "utility computing" that's always available and can easily scale. Cloud computing proponents claim that it's easier to just add additional cloud computing resources than to go through the hassles of getting additional budget approved, ordering more equipment and dealing with the overhead that operating the additional equipment would cause.

I have to say that I find this argument unconvincing. These sound more like management problems than technical problems, which means that trying to solve them with technology is probably doomed to fail. If your organization doesn't want to pay for additional computing resources that they run themselves, they're probably not going to want to pay for additional resources that someone else provides either.

Cloud computing is also supposed to give you the ability to react quickly to changing business requirements. It's supposed to let you bypass your corporate IT department that may be unresponsive and overstretched. This also sounds like a problem with management instead of with technology. Your IT department exists to support other business units, and if they can't react quickly enough to changing requirements, this may be more a reflection of the management of the IT department instead of limitations of the technology that they use. Because of this, I don't find this argument convincing either.

This leaves two of the three reasons for cloud computing in doubt. The way in which cloud computing has experienced success seems to support doubting the weak claims. After all, most of the success that cloud computing has experienced has been with start-ups and other small businesses. These are the very businesses that benefit from its ability to cheaply and easily create an IT infrastructure.

Enterprises, which are the ones that would tend to benefit from the two weaker claims, are also the ones that haven't been as interested in cloud computing. If the only reason to use cloud computing that can withstand much scrutiny is that it provides a cheap and easy way to create IT infrastructure, it may actually never end up experiencing much success in the enterprise market.

Comments

[Rick Sanchez]

Luther,
While on the surface the unconvincing nature of the arguments is valid, I don't think you scratched the surface here.
1. The net spending on "additional" utility IT resources is far less in the cloud vs. doing it in house because you stop paying when you scale back.
2. The ability to "bypass" the IT department for quick reation to changing business needs is VERY valid. If my company got a HUGE contract today and needed additional resources tomorrow, my IT dept. would not be able to buy, instal, and configure the hardware needed in a timely fashion.

To me, it seems like your objections are a bit shallow.

[Kurt Cagle]

Luther,

I admit that I share your reservations about enterprise level cloud computing hype. I see three scenarios where cloud computing actually makes that much sense:

1) You are a new company that is not interested in investing heavily in a physical plant to support computing resources. In this case, hosted services make a great deal of sense, both from a development standpoint and from a SAAS foundation.

2) You are a mid-sized company with extensive expansion plans, putting you in much the same scenario as one; in this case, it may make more sense to move your existing IT into a common hosted environment as your new divisions, and even there its likely that you'd still maintain your existing your infrastructure.

3) As a company you are being forced to downsize - either to move out of a large facility that has physical plant support to one that has much poorer capabilities, or going entirely virtual. Of course, people don't like to talk about this scenario much.

If you're already an enterprise-size company, the cost benefits involved in making a wholesale jump to distributed services and hosted servers are probably minimal if you already have an existing infrastructure, and you lose the reliability and security benefits of maintaining a private network.

-- Kurt Cagle