Clouds used to be cute puffy white things in the sky. These days when you talk about clouds, you are more often talking about Internet-based clouds. Even recently just a domain for geeks and techies, knowledge of Internet-based clouds is penetrating down to the rest of us. It may be that iCloud icon on your smartphone or iMac, or the convenience of a Gmail account that you access from a hotel business center. It’s starting to register with us that we are using clouds. We no longer store data in our own personal devices. We don’t know or care where it is stored, just as long as it is. Clouds are here to stay.
The cloud is just the latest manifestation of a trend that has been emerging for some time. With the cloud we no longer worry about whether data like our email or digital pictures are archived and backed up. We assume that if we have a connection to the Internet it is all available instantly. The promise is that at some point it will also somehow done transparently and with no hassle.
All of this is also something of an illusion because in reality this level of cloud computing is really, really hard. Google and Amazon are pioneers in the cloud computing business, but pretty much all the major IT providers are lining up to provide cloud services too, from the mighty Microsoft with its online Office cloud service to the lowly web host (who will probably use someone else’s cloud and make it look like their own). Google has had some infamous cloud outages over the years, most frequently affecting Gmail. More recently Amazon has experienced some embarrassing cloud failures too. It’s nice to know that your stuff is out there somewhere, but making it always instantly available is quite a trick. 99.999% uptime is pretty darn good and most of us would not notice minor outages. The same is not true for businesses that depend on continuous uptime, like United Airlines. Since they can’t take any chances, they are sticking with their own data centers, at least for now. In general, cloud computing tends to be a lot cheaper than doing your own hosting. You just don’t want to jump into the cloud computing arena unless you are really, really sure you can trust your cloud vendor.
The Department of Interior, where I work, is taking the plunge. It recently signed a contract with a company that resells Google’s infrastructure to provide the whole department’s email, calendaring, instant messaging and various other services. In doing so it will save heaps of money, unless the vendor’s claims don’t match actual experience. In that case, lots of highly paid people will be twiddling their thumbs until service is restored because the cloud will be like a light switch: it will either be on or off, and bad things will happen if it goes off even for a little while. At least right now when there are problems they will tend to be localized instead of enterprise-wide.
In any event, we are going into the cloud. To affect change you have to break a few eggs, and in this case a lot more than a few eggs are being broken. Every employee in the department has to be retrained. Microsoft Exchange and Lotus Domino servers are retiring to greener pastures (well, more likely landfills). Our comfy though somewhat weird email clients are being traded for doing email in a browser. Everyone has to adapt, including our director and Ken Salazar himself. I doubt that even the department will know where the heck their servers are. It’s someone else’s problem, specifically Onix’s, which got the contract. Email alone is mission critical for our department. It’s got to work and work reliably, and the transition has to be smooth. Everything ties into email in some fashion. About 80% of my work day is spent reading and responding to email. My team depends on instant messaging as well, as we are spread across four time zones. We run a mission critical system, but this new cloud-based system is even more mission critical. If we cannot communicate to fix our mission critical system, then it can go down. The nation’s motto is “In God we trust” but perhaps it should become “In Google we trust” here at the Department of the Interior.
There are so many wrinkles to this cloud computing stuff that I will be taking a course to understand most of it. For those of us managing information systems, one of the compelling features of the cloud is the promise that it can scale up automatically to meet higher demand. If true this is quite a feature. Automatic and transparent failover to redundant systems is also available. Obviously, vendors charge more if you need these features, but the promise is that overall it will be less expensive to use the cloud than it will be to have your own hosting center.
This may mean unemployment for many technicians now keeping servers running. Those who physically touch these machines are most in jeopardy. At least in the short term, those who configure these machines in the cloud to do unique stuff probably have secure jobs. With a few clicks you may be able to have your cloud provider install an operating system or a web server. (In reality these machines are already likely provisioned, and are just sitting idle.) If your needs are modest, then you may need fewer system administrators. Integrating a server and the applications that run on it are not necessarily simpler because they are in the cloud. It just becomes more abstract. In some ways, administering cloud servers applications may be more complicated, since the whole cloud architecture needs to be well understood, and things don’t work quite the way they used to.
I learned recently of a revolution that happened around 1820 in nearby Harpers Ferry, West Virginia. The town is known for many historical events, and some innovation. One of its innovations was the invention of a breech-loaded rifle. It was constructed from completely interchangeable parts. This revolutionary idea first perfected in a gun was extended to all sorts of items. It made possible the Model-T and many other inventions.
Cloud computing is the latest refinement of this idea born in Harpers Ferry by a man named John Hall. The management, storage and configuration of systems used to store data and information is becoming virtualized and commoditized as well. The interchangeable parts are not so much hardware but the software that runs on the hardware. They are becoming so excellent and interoperate so well with other standard software parts that reside on these servers that new levels of performance and cost savings can be achieved.
Cloud computing is the latest and if it works as advertised will arguably one of the most important revolutions in information technology. We out here in the business world will fuss over it for a while, and there will be more growing pains, but like the breech-loaded rifle cloud computing is a fundamental invention made up of lots of other clever inventions, many of them abstract and conceptual and modeled in software. These will become savings in time and benefits of convenience that we will soon take for granted but which could be as fundamentally transformative to mankind as the Internet itself.