With the bulk of stories on cloud computing having taken a negative tone, we have one that talks about the top ten benefits of “cloud transformation”. The question is whether enterprises really accept some, or even any, of these. Here, I’ll look at them all and recount what I hear from the 314 companies that shared what they thought about the benefits being claimed. I want to point out that I don’t survey enterprises, but rather collect their spontaneous views, so the data here is assembled from the comments they’ve made that address each of those claims.
The first benefit the piece cites is enhanced security. Only 8 of the companies said they believed the cloud enhanced security, while 47 said that cloud usage reduced it, and the rest said there was little or no impact. Of the 47, 36 said that the cloud increased the attack surface, 65 said that the point the article makes about cloud provider security teams wasn’t their own experience. The main point is that virtually none believed security to be a benefit to cloud transformation.
The second benefit is increased flexibility, and here all 314 enterprises agreed, but suggested that the point in the piece should really be renamed “increased scalability”. They said that this was the benefit that was actually critical to justifying the cloud in the first place. Data center infrastructure has to be sized to peak load, which means a lot of resources are wasted on the average. The enterprises did point out that if there was little variability in the load levels of the applications, then cloud usage was unlikely to be cost-effective.
The third benefit, increased integration, was accepted by only 13 of the enterprises. They said the type of integration cited in the piece, which is the integration with third-party tools and elements, doesn’t figure into their cloud justification, and integration among their own applications was no easier in the cloud than in the data center. The 13 who accepted the benefit said that their cloud adoption made it easier to customize a particular user’s view of their applications and data because the cloud component of the applications could pull from any useful application or database resource, from any of their own applications.
Benefit four, enhanced data analytics, was sort-of-validated by 99 of the enterprises, but not the commentary on the available business intelligence tools. Enterprises say that they run their BI applications with the data, and all but that 99 say their critical data isn’t in the cloud. Even where it is, the majority said that the availability of analytics in the cloud was not the compelling benefit that justified their cloud transformation.
The fifth benefit, improved performance, was rejected by all the enterprises. Nobody felt that the cloud outperformed data center resources unless you figured in the scalability point, which the article did not. The majority did agree that the cloud allowed them to site application elements near the user even if users were widely geographically distributed, but only 50 said that this had a meaningful impact on performance, and 23 of that group said they could extend or had extended their corporate VPN to the full user geography.
Benefit six, enhanced disaster recovery, was accepted by all users, but not exactly in the way the article suggests. The enterprises saw this as an attribute of cloud scalability, noting that replication and synchronization are neither automatic nor free, and only 39 users said they could justify those features. The majority of users also noted that managing availability in the cloud wasn’t much different from doing that in the data center and that most of their core applications and critical databases remained in the data center anyway.
For benefit seven, increased cost savings, user experience is decisively not in favor of the way the article presents things. If you have stable workloads, enterprises say that the cloud is usually more expensive than the data center. Moving to the cloud strands data center costs in nearly all cases, and larger enterprises, about half my group, say their economy of scale in their data centers is nearly as good as that of the cloud (remember that economy of scale follows an Erlang curve, asymptotic to a limit) and that cloud provider profit margins eat more than the limited cloud provider advantage in economy of scale.
Benefit eight, increased business continuity, seems to be largely a part of the disaster recovery benefit already cited as number six.
To users, benefit nine (increased competitive advantage) is a toss-up. Of the 314 enterprises, 174 said that this was illusory simply because the cloud portion of their applications (remember, few enterprises move everything or even most things to the cloud) is only a piece of functionality, not all of it. The remainder agree that cloud front-end technology is more agile and that customer-facing applications can benefit at least somewhat from the cloud transformation.
Benefit ten, improved agility and innovation, seems to be largely a repeat of other benefit claims, especially the two previous ones on the list. In addition, some of the points seem to users as a point that could be claimed as an example of good corporate policy, but not as an actual project justification. Others that deal with facile cloud development cite positives without citing cloud negatives, and so overstate their value.
Overall, enterprises haven’t expressed much enthusiasm for the great majority of these claimed benefits of cloud transformation. I can’t express much myself, because I think the majority are repetitive, platitudes, or both. The fact is that nearly all of them are not any more attainable by simply adopting cloud computing than they would be by simply hosting in the data center. The specifics here are critical, and that perhaps is a justification for taking a superficial approach; there simply is no way that a topic as complex as “cloud benefits” can be usefully covered in a thousand words or so.
Or, of course, this may be another example of the fact that what gets published online is what gets clicked on. If so, that’s too bad, because you can’t make good business technology decisions by counting the number of clicks your references get.
