If ITIL didn’t exist, would DevOps need to invent it?
At conferences, my introductory slide often says that I’m interested in ‘anything that helps IT work better’. As part of this, you’ll find me at a broad range of events, from ITSM to Agile to DevOps, learning more about what's going on in the world of IT.
Jan 24th I was at Leeds DevOps (organized by the indefatigable Andy Burgin and sponsored by Infinity Works) watching two talks on Site Reliability Engineering (Sam Smith, Sky Betting and Gaming) and The Unexpected (Infrastructure Management) Journey (Mihaela 'Micky' Drumeva). Quite often the talks at DevOps meetups are a bit over my head in terms of technical content, but both of these sessions were right in my comfort zone – people, process, value, objectives – classic service management?
Sam spent a lot of time talking about Google’s SRE approach including SLAs, including SLIs and SLOs (you can read more here). Micky focused on her story of rescuing a failed project, including the importance of monitoring, rollback plans, documentation, and making sure you keep the support team up to date about what’s happening.
There are some amazing ideas coming out of the DevOps community, and the energy and general friendliness at the meetups is always impressive. Equally we have a strong community in ITSM, but it does feel sometimes that ITSM and DevOps are at different ends of the same spectrum. In IT service management, we assume that most organizations now have the basics in place (incident, problem, change) and are reaching for new, more complex ideas (e.g. value systems, XLAs). In the DevOps world, organizations don’t always have the basics in place and are starting to see the value of a service desk, incident management, a level of change control around their services. How do we link up these two ends and make sure no one reinvents the wheel? Will DevOps invent ITSM?
Things have changed a lot since I started in ITSM and we need to be T-shaped professionals. Only understanding one perspective, whether that’s ITIL, DevOps, Agile or whatever, isn’t enough. This was one of the key ideas behind the VeriSM project, to help organizations blend different bits of management thinking and ways of working to create their own model.
With the launch this year of ITIL 4, COBIT 2019 and the ISO20000 update as well as the ongoing evolution I’m intrigued to see where we go next. I’d love to hear what you think. For some more articulate writing on this topic, I’d also recommend you read up on Jon Hall’s articles.
Wht’s diff btwn DevOps Vs ITIL🤔, can DevOps delivers wht ITIL has ?
What is in ITIL were there long before ITIL, so no, DevOps would not have had to develop it. As mainframe "systems engineer" we used all of the processes in ITILv2 at least in the early 80's.
Thanks for the post Claire. I can only really echo that which is already said especially that said by Joel P. who sums it up so eloquently.
Crickey, never been described as "indefatigable", I shall look it up later. Here's my view from the devops perspective... 1) The devops movement spawned from a conference about how to make agile development teams and the more "cautious" ops teams work together and deliver value. It was about breaking down the barriers between the teams, having empathy and and adopting working practices that made life better. It was a deliberate move by the community to leave the term devops definition-less. 2) Off the back of this many principles, ideas and sometimes models have spawned, not only based around empathy/collaboration but delivering value. e.g. CAMS/CALMS/3 ways/VSM. - borrowing a whole lot from Lean and manufacturing,, not IT 3) Then we spent 2014 and 2015 telling the world devops wasn't about tools, it's not a job title, there isn't such a thing as a devops team and there most certainly aren't certificates and qualifications for devops - clearly not everyone listened 4) ITIL imho complements a lot of the ideas devops ideas, so long as a practical amount is applied and not everything is implemented. I see it as a framework to make sure changes/incidents are safely managed, provided there is a sensible amount of review (by people that are empowered to do it - shift left kids!) - what I don't see it as a way to "manage work", that's for agile practices. 5) I see the SRE ideas as a way implementing a bit of both, hopefully the best of both worlds, but I'm still to see evidence of someone other than google implement such and pure/opinionated incarnation with success. Please note I've deliberately missed the tooling - IAC, containers, monitoring, metrics etc as that detracts away from the point of it being about value and collaboration.