Blue Sky Computing - Dissipating the Cloud
An Analyst Insight document published by Nick Castellina of the Aberdeen Group in March 2013 prompted me to speak up for the silent cloudless majority.
An Analyst Insight document published by Nick Castellina of the Aberdeen Group in March 2013 prompted me to speak up for the silent cloudless majority. Although Mr. Castellina is careful not to make bold affirmations supporting Cloud computing (through the frequent use of many non-committal wiggle words and free associations), he does offer up many gratuitous enunciations that warrant comment.
The first of these is the fact that the author offers no definition of cloud computing, but instead presents two definitions that he (loosely and wrongfully) associates with cloud computing. Paraphrasing the study:
"SaaS or On Demand: The software itself is not ...owned... (is) provided as a service...on a subscription basis. SaaS avoids a large upfront capital investment. This deployment method is one form of the Public Cloud"
"Hosted: Also considered part of the cloud, licensed applications are hosted by a third party."
You can see the use of wiggle expressions "one form of" and "considered part of" the cloud does nothing to persuade anyone that the cloud is anything more than a loose virtual construct. Fact is hosted applications where around in the 80's, where many business applications resided on mainframes, and applications (e.g. payroll) were deployed to guest users through telephone lines, modems and dumb terminals. With the client/server technologies that followed, nothing prevented any organization from having even legacy applications hosted off site at a third party's location, and deployed through Citrix or terminal server technology, to computers or dumb terminals, using the internet or T1 lines as a conduit. My point is that the hosted part of what is associated with the cloud is not limited to applications with web interfaces.
As for SaaS, it has to do with finance. You don't buy the software up-front, you just rent it. Similar to a car lease, except that the car lease has the advantage of having an end of payment date. It's actually more like renting a water heater. You pay monthly, forever. Again this has absolutely nothing to do with web interface applications.
To strengthen this case, many supposed cloud applications can be purchased in the traditional way, and deployed on premise if that's what the customer wants. Hosting and method of payment have very little, if anything at all, to do with the cloud.
Another train of thought worthy of comment is the author's (weak) attempt to explain manufacturer's resistance to Cloud computing by offering the following three possible explanations:
1. Misinformation to the effect that cloud computing is not as secure: To manufacturers, this is not misinformation. Where there is smoke...there is fire. Cloud computing has that stigma because most applications offered in cloud deployment are indeed deployed on server farms owned by very large data centres. Most of these companies have redundancy capabilities (an equally large server farm capable of hosting applications in the event of failure at the first location) located offshore. To a convenience store or small law firm, no big deal. To manufacturers who compete with Asia (and have seen production off-shored for many years), different story. The idea that, one day, their whole system and data could reside in Asia, is a risk. Many of these countries have completely different sets of rules than North Americans have when it comes to intellectual property. To North American manufacturers, this is a needless risk.
2. Manufacturers may not be aware of cloud computing's benefits: Owners of small and medium sized manufacturers all have a solid understanding of concepts such as capital investments and ROI. All these laggards the study refers to, who have owned systems for 7 to 15 years, may not all be on the latest tech, but they finished paying for their system in the same year they bought them a decade ago. Presenting rentals to them will hardly be perceived as a benefit. Quite the contrary, manufacturers quickly come to the conclusion that the only group benefiting from SaaS are the software companies, who get a steady, never ending revenue stream.
3. Manufacturers may have a change-resistant culture: I will concede this. Many do. And it has served them well. But when there is value, they also recognize it. Improvements in robotics and other advanced manufacturing technologies that have value are adopted by manufacturers large and small when the ROI is real.
There are other advantages of on premise computing that are not discussed here, including the flexibility to decide on the timing of an upgrade; another is the greater freedom to customize; last but not least, robustness and reliability (you cannot be both new, and have withstood the test of time).
Web based applications have a lot of promise. Greater ease of use, navigation ability, and multiple device deployment can all facilitate business system adoption rates and user productivity. If it is serious about penetrating the manufacturing segment, the cloud community would do well to focus on these areas, rather than continue its futile attempts at clouding the computing issue.
