ITIL vs The Cloud: Pick one

Like Tweet Pin it Share Share Email

My major premise is that ITIL (according to its proponents, at least) defines professional standards for IT management. My minor premise is that ITIL and The Cloud are incompatible. My conclusion: Moving your IT infrastructure to The Cloud is unprofessional.

Make this conclusion a major premise. The minor premise: Gartner predicts that within 5 years, one out of five companies will have 100% Cloud-based IT infrastructures. The inescapable conclusion? Gartner is predicting that, in a demi-decade, 20 percent of all companies will abandon their professional standards for IT management.

“Wait wait wait!” I hear you protest through my tinnitus-crazed cochleae. “What do you mean, The Cloud isn’t compatible with ITIL?”

I don’t claim guru-level expertise on either ITIL or The Cloud. Based on what I do know, I’m pretty sure ITIL-approved Change Management can’t happen when IT has no say in whether an upgrade takes place or not. That’s quite important in ITIL-land (and in the land of your IT operations even if you aren’t an ITIL advocate).

Still, I’m working hard to avoid strongly held opinions in the absence of expertise. Which is why I called on my friend and ITIL guru Rick LiaBraaten.

KJR: Rick, if you look at The Cloud and compare it to what ITIL requires for IT best practices (and when will ITIL wake up and stop using that phrase?), how does it stack up?

Rick: I hate to break it to you, but the folks who manage ITIL have woken up. Version 3 talks about “good practices,” not “best practices.” Maybe they read the KJR Manifesto and paid attention.

KJR: Stranger things have happened. Not many, but a few. Anyway, what do you think — does The Cloud conform to ITIL’s good practices then?

Rick: Some of them, sure. Keep in mind that ITIL’s “good practices” are guidelines, not rules. Implementations should be tailored to each organization’s specific situation. Based on the companies I’ve worked with, I’d have to say The Cloud is going to fall short for a lot of your subscribers in several very serious ways.

KJR: Which are …

Rick: You already picked the worst offender — Change Management. Performance and Availability Management are also problematic. So is Problem Management.

KJR: How so?

Rick: Let’s start with Problem Management. When you’re working in The Cloud, unless you can get your entire technology portfolio from a single source you’re going to integrate business solutions from multiple vendors, just as you do now. That means when something goes wrong you’ll deal with multivendor finger-pointing, just like now.

This is a very old problem in IT. Only it’s much worse in The Cloud. In your own data center staff experts can get at the technology themselves to figure out what’s going on. With Cloud-based services they can’t, so if IT has assembled a solution from even three vendors, all of which verify their servers are up … well, the word “screwed” comes to mind.

KJR: And that’s the easy one? How about Performance and Availability Management?

Rick: It’s like this. These days, most IT shops figure if they don’t know an application is down until a user calls the Service Desk to complain, something is terribly wrong.

To be fair, we’re starting to see a few third-party tools that can monitor SaaS applications and alert IT when they aren’t available, but this isn’t what you’d call mature technology. So far as I can tell, the SaaS vendors themselves … and I’m including Salesforce.com, which is the one everyone points to as the shining light of Software as a Service … offer nothing to help IT manage the application.

In fact, it’s the opposite — part of the selling point is that you don’t need IT to manage it. That’s the SaaS vendor’s job. And as long as you consider “trust me” to be a good approach to supplier management, I guess that can work.

By the way, Supplier Management is an ITIL core process, and it doesn’t list “trust me” as an example of good practice.

KJR: Imagine that. And Performance Management?

Rick: It’s a good-news/bad-news situation. The good news is that elastic provisioning is one of The Cloud’s major selling points. If performance starts to suffer, it’s easy to add resources.

KJR: And the bad news?

Rick: The bad news is that the tools for monitoring performance are even more primitive than the tools for monitoring availability.

KJR: And we haven’t even started to talk about Change Management. That will have to wait until next week, though, because I keep KJR to 800 words or so.

It’s my version of good practice.

Comments (15)

  • So was the broken ITIL link an intentional error?



  • My initial thought when I read about cloud computing was “this will never work”. However, hardware, software and service vendors are pushing this hard because this is the only new thing that can generate the kind of revenue that a complete change in topology and methodology will do. I have made a lot of money in the past writing software to convert to the next greatest thing even though I advised clients against making changes I didn’t think were viable. I’ve made a lot of money re-writing code to remove the things that were judged by the industry to be poor practices after they were implemented for a few years. My prediction is that a lot of people will get a paycheck thanks to cloud computing regardless of its viability.

  • Short answer – you get what you pay for… Down sides that you speak of are all true(I’m living them now) That said, I paid pennies on the dollar for the service that was configured, up and running in 8 weeks from contract, for 10K users…

    Same arguments can be made for outsourcing (why it isn’t a good idea) – but the bottom line always wins.

    In 5 yrs these issues will have answers.

  • Bob. I disagree. Let’s see…

    Your change management argument makes sense. In the cloud, you have one or two more levels of indirection before you can get to the root of any problem. However, your argument discounts one very important fact. Large cloud providers such as Amazon and Google have built their infrastructures (parts, architectures, implementations, operations, and support) to be hyper-efficient and hyper-redundant. They have done this so that they don’t have to spend their (very expensive) time dealing with their customers’ “problem management” teams. The result of this efficiency is that, their customers don’t have to have large “problem management” teams. i.e. their hyper-redundant hyper-efficient infrastructures make change problem management a non-issue. Win-win.

    For performance and availability of the infrastructure, I’ll go back to the hyper-efficient and hyper-redundant infrastructure of the large cloud providers. Because they have standard parts, standard architectures, standard implementations, standard operations, standard support, and economies of scale, they run a very low-cost operation. Some of that cost advantage is used to make the system redundant. Some is used to over-provision infrastructure (e.g. computing, storage, network) to give elasticity (i.e. on-demand provisioning), and standard operations include good monitoring. For performance, it becomes easy and cheap to quickly provision extra on-demand resources to address performance and availability problems. Win-win.

    There is an age-old argument of whether to get good apparent performance by using high-quality, well-monitored systems, or to take lesser systems with less monitoring and just over-provision them. I remember 15 years ago, Nortel had the excellent networking equipment with high availability, excellent real-time traffic management, and fantastic management. Cisco’s equipment was not so good. Who ultimately won? Cisco! Why? Because for a CIO, it was easier to buy, manage, and over-provision inexpensive Cisco equipment than it was to buy relatively expensive Nortel equipment and then pay an expensive operations team to manage it perfectly. The same analogy will hold for cloud computing. It is cheaper for a CIO to just throw cheap on-demand servers at performance and availability problems than to have an expensive, in-house, well monitored and well managed ITIL-good-practices-based system.

    ITIL has done a lot to advance the state of IT management. Cloud computing will lower the cost IT management, and it will significantly lower the cost of IT. At the end of the day, the executive team of an organization does not care about ITIL or any other formal processes. They want a reliable job done at the lowest possible cost. Cloud computing offers this, and if cloud computing offers this at a much lower cost and manageable risk, then cloud computing will be used. The ITIL standards will adapt or die. Look what happened to CCITT/ITU (International Telecommunications Union) in communications. They have been steadily eclipsed by IETF (Internet Engineering Task Force) because IP networking (specified in IETF drafts, proposals, and standards) can now be used for most communication-related services at a fraction of the cost of doing it the CCITT/ITU way.

    The professionals will be the ones left standing at the end. I predict that these folks will be the ones who learn to leverage cloud computing to solve their IT problems.

    • Your argument makes complete sense in a single-vendor environment. If a business of any size and scope can operate through only a single Cloud provider, or through multiple stove-piped providers without the collection of integration points you find in most internally managed architectures.

      I’m skeptical that this scenario will work for a business of any complexity. As I’d hoped I’d made clear, the issue isn’t a single Cloud vendor failing. The simple stuff is always easy.

  • Since many organizations fail to implement ITIL correctly (though they tell their stakeholders otherwise) I’m not sure the loss of ITIL will make a difference.

  • Hi Bob,

    I think the issue here is that the cloud does not yet have a mature set of tools. But integration tools, monitoring tools, and the like will get better. Cloud vendors will give customers more control over when to implement new features. All of this will push cloud computing into the mission critical space. It’s not there yet, but it will get there.

    IT history teaches us that technologies change. We have gone from mainframes to mini-computers to client/server to the web. Full blown cloud computing is the next step.

    — Mike

  • It’s not ITIL (IT process) or cloud – I think it’s IT process (ITIL/others) and cloud (one way to provision).

    Whether the process model it ITIL or/and something else, large companies will still have a mishmash of in-house IT and service providers of many flavors. And they will need some type of process framework to manage it all until IT service (at the application level – not network/ping dial tone) becomes so consistently available that no one worries about it. I doubt that will be in anyone’s lifetime who’s working in the industry today – even the newbies.

  • Bob.
    Interesting premise, but you are mixing apples and oranges. ITIL provides a framework of processes for “Best/Good Practices” in computer service operations. It tells an organization that here are the processes you should apply to ensure that you are providing the most efficient and effective computer service operations. It doesn’t tell you how to implement these processes or if you must do all the processes. It is a “buffet” of service related processes to assist an organization in improving the computer service operations. It has nothing to do with professional standards and in fact if you want to be certified as following professional standards, I’m sure ISO has many standards in which you can work to be certified. Of course being certified in a standard doesn’t mean that you are doing anything useful for the organization, but gosh you are certified!

    Now cloud computing is just taking out-sourcing to the final step (Well, I guess you would also need to completely out-source your entire business process to be completely out-sourced. Only the CEO left as an employee. Now there is an out-sourced model that hasn’t been tried.)
    Cloud computing just takes overhead of your IT assets (HW, SW, Infrastructure and People) and transfers them to a supplier, who then provides computer services to the company for a contracted price. This doesn’t mean that all the processes that you had in place are thrown out. There still needs to be Asset Management, Change Management, Incident Management processes being executed, to ensure that the IT assets are delivering services in an efficient and effective manner. It is just that these processes are shifted to the vendor for most of the roles and your organization plays a different role. You still have “IT Process Owners” along with their business counterparts who need to ensure that the IT service is being provided and that the vendor is meeting the SLAs that have been agreed to in the contract. (I guess you never get rid of SLA Management or Vendor Management)

    Cloud computing is just a way for an organization transfer IT risks to a vendor using SLAs to manage the contract. It doesn’t mean that you wouldn’t have a say in how your cloud is managed

    Finally, the cloud computing contract is just that, a contract. When the contract expires, the organization may decide to in-source the IT operations or move them to another vendor. The ITIL processes being in place will make that a smoother transition.

    Am I on-target?
    Thanks for the chance to respond to your thought provoking article!

  • Interesting perspectives on this. I have to admit when I first read Bob’ article and he said he was not an expert in either ITIL or the Cloud, I was like ” then shut up, we have enough unexperienced bloggers out there.” So glad you brought in an expert. But before I get too far, can you please give me a link to Rick’s profile. I’d like to know what makes him in expert in either area. Because honestly, if we are talking about swapping a hard drive, or installing a server as the underlying reasons for initiating a change management process, then I’m really reading the wrong stuff. An under-taking to cloud services at any level of the infrastructure requires excellence in configuration and change management. The forklift plug and play mentality is what cowboys do, and offer now business value or relevance in the professional space. The planned and well executed reallocation of highly scalable resources to a team of focused and dedicated functional experts on the other hand is professional. I really think Jay hits it on the head about getting stuff done. A CIO’s lifespan is short these days. We don’t have time to sit around and wait about institutional cultural development. Give me an OS, then give me 20. So I can show some momentum and get some traction on the new social media technology, or CRM platform, or analytics engine. If people are still taking 2 weeks to do change reviews, or holding weekly meetings then shame on them, and stop BLAMING ITIL.

Comments are closed.