According to some, software as a service and cloud computing are nowhere near as cool as everyone thinks.
A viewpoint in BusinessWeek argues that most of the good points about SaaS are actually just hype. Author Gene Marks argues that SaaS is not cheaper, does not really reduce hardware investment, is not quicker to set up, is not secure, and has not been in use for years. He then concludes that SaaS is "just another option for getting stuff done," which I agree with, but the rest of the article seems to counter sweeping generalizations with new ones. It would be better to caution against irrational excitement and remind readers that SaaS isn't the panacea for all the world's problems and that you do still have to do your research and be careful. Saying that SaaS as a whole isn't easier to set up or cheaper because of a couple of isolated examples seems erroneous when you could quite easily cough up examples of where SaaS can be cheaper and easier to set up. It goes both ways.
Then, an article by James Maguire in Datamation discusses the "many dangers of cloud computing." Maguire doesn't call the benefits of cloud computing "myths" as Marks probably would, but he does point out some important security concerns to keep in mind before choosing a cloud solution, such as ensuring privileged user access, considering data location, having a disaster recovery plan, and looking into the viability of the cloud computing provider. All are good points.
And then, on a very loosely related note, another article in BusinessWeek caught my eye with the headline It's Too Darn Hot. The article discusses the tremendous cost of powering and cooling data centers and how some companies are investigating placing data centers in naturally cold and hard-to-inhabit areas such as parts of Iceland. Apparently data centers spend half their energy usage or more on air conditioning. I suppose this would be one indisputable way in which cloud computing is not "cool."
Here's my comment on the Business Week article:
This article should have been written in 1999 when ASPs were trying (and failing) to reduce costs and complexity by hosting licensed software. Or maybe 1996, before every business PC was connected to the Internet, and before the Internet was securly processing billions of financial transactions each year.
Today, SaaS solutions are written from the ground up for fast implementation, low cost of ownership, security, continual upgrades, and usability. Apparently Mr. Marks hasn't actually worked with any of them, or maybe his consulting practice relies on big implementation projects of old-style software.
Un-debunking Mr. Mark's "myths" is pointless with his lack of facts. Here's just one un-debunk: has there been a SaaS vendor security breach? No.
Mr. Marks can go back to the 90's and implement Siebel for tens of millions of dollars and zero payback because the end-users just won't use it, but the rest of us are in this millenium where SaaS solutions deliver a far better ownership experience for customers.
I am a huge proponent of SaaS (that's what our company specializes in) but I do think the dangers of underestimating the complexity of a solution just because it is a SaaS deployment are real.
Many enterprises consider SaaS because of the promise of shared innovation and a lower TCO. Unfortunately that sometimes translates into the line managers underestimating the cost and complexity of planning the implementation correctly.
The requirements planning and design process are even more critical in a SaaS implementation because if you try to implement a requirement in a way not supported by the SaaS vendor well, costs and timelines can skyrocket.
I think the big problem is when companies try to implement a SaaS solution with the same methodology they used to implement premise based solutions where they had control over every aspect of the development process. We have found that to realize the benefits of SaaS, you need a SaaS specific methodology. Said another way, if you try to implement a SaaS solution with a non-SaaS methodology you will get the same results you got before.