A colleague of mine passed on a link to an article: “ITIL vs. The Cloud: Pick one”. The main premise of the article is that ITIL (which is an IT focused control framework for managing IT systems) and Cloud (dare I define parenthetically as outsourced IT services typically providing software as a service capabilities).
While I think the article is provocative enough, I’d like to pile on – the people who are often pushing the most for cloud technology are often the least empathetic to what IT actually does. When you see IT as slow and expensive, you can easily be seduced by the idea of completely outsourcing business supporting systems to cut IT out, with the immediate reward of “it just works” type of thinking. (Everyone seems to believe that the other guy’s IT is better than my own. Having been in many organizations, be careful what you assume.). And, frankly, it’s often hard to argue that there aren’t some really great opportunities for SaaS to have major break-throughs. For example, I’m a big proponent of Concur, which is a travel and expense SaaS tool that I’ve been using for a long time. Equally, Salesforce.com has shown how CRM can be extremely effective in a cloud-based configuration.
But, as I often point out, it’s rarely the big applications themselves that are the hard part. What’s hard in IT is system integration, monitoring/operating things in connection with each other, managing data and dealing with unexpected situations. Cloud doesn’t solve the integration problem unless you completely migrate all systems onto a single platform. Cloud doesn’t solve monitoring/operating because it doesn’t get rid of all systems – and it perhaps makes it worse because you now have off-premise systems that you aren’t able to understand more fully what may be going wrong (i.e., is it a network failure, too much data, etc.) While there is some movement here, it’s still pretty much a black box. Managing data isn’t any better – it still exists in spreadsheet, databases, enterprise systems and cloud systems. As a matter of fact, cloud also often makes this worse as limited data format customizations within the cloud systems means that you have to implement additional ETL capabilities to make it fit the platform. And, finally, it doesn’t address unexpected errors – that can still happen across the organization and even in the cloud (or connecting to the cloud) as well. And, as the article points out, unexpected errors could now be coming from your SaaS provider who makes updates to their software without any typical production communications. The mantra “there are no migrations or versions” is sexy to the end user until all the data connections feeding the system breaks. (In computer science, we call this a “side effect” – which like the military equivalent “collateral damage” – sounds really innocuous until you realize that there is literally nothing you can do about it until you resolve the root cause. And, do you suspect it will be your cloud vendor stepping up to admit that it was their change that caused your problem?)
As a good example of this, I was at Dreamforce (the Salesforce.com conference) earlier this year where I listened to a team discuss a “success story” of building an application on Force.com. They called it a success because the team was able to build a ground up application in 9 weeks that pulled in data from numerous sources to provide visibility to multi-level-marketing commission values. As the topic went on, though, it was noted that the data was actually fed from a corporate data warehouse that integrated all the systems into a single view which was then forwarded to the Force.com platform. Essentially, they had created a fairly straight forward non-transactional visibility system that relied on heavy duty IT support and integration to make function. If the underlying data warehouse went down or the integration between the systems broke, the application itself would go stale with data and be unusable. So, of course, this solution still requires IT support. (Perhaps that’s why salesforce.com changed its motto from “No IT” to “No software” – or perhas they were picking fights where they should have been building bridges.)
From my perspective, cloud should be seen not as everything but as another tool in the belt of solving business problems. There are a number of situations where it’s a great fit today. And, I expect that to conitnue to expand as more investments are made in the space. But, equally, it should not be seen as the solution to all problems, and, IT especially, needs to continue to deal with governance and processes like ITIL.
And, as a proponent of cloud concepts, I’d also like to warn IT that this point of view should not be used as a weapon to defend the status quo. Cloud is upon us, and we need to take advantage of it however we can – just like the introduction of any new technology that can lower costs, make us faster and deliver more value. Don’t be afraid to roll with the change…