Having been around at ServiceNow for a while now the highlight of my week is still getting the opportunity to meet with customers and hear how they are using the product to solve business challenges day after day, week after week.
In the first couple of years the message from customers was consistently around Service Management; "how should I configure ServiceNow for Change Management" or perhaps questions about scaling the Service Catalog to a size that exceeded the developers best estimates whilst coding it.
But during the past 6 months your message has changed. I know that people are still working on Change Management and Service Catalog (pretty sure Change isn't a solved problem for most of us yet!) but the emphasis has most certainly shifted.
Recently I've spoken to grocery mega-chains, news networks, pest control exterminators and financial services companies about the same subject. How can we use ServiceNow to start a transition to a DevOps style of working.
The opportunity to talk to customers about DevOps and ServiceNow is too good to pass on. As a company ServiceNow has scaled to its position as "The Enterprise Cloud Company" though high levels of automation, orchestration and a strong alignment between its Development, Site Reliability Engineering and Customer Support organizations. I'm lucky to have witnessed it happen over the past years.
And all of this is done using our own product as the process workflow platform, the orchestration tool and the single system of record. Customers are really interested to know our views on DevOps and to see a little of how it's done.
Why are customers getting excited about DevOps now?
DevOps isn't a new phenomenon, but like its buzzword ancestor "cloud" it took a while after the initial hype to get to a point where customers feel they should be doing something about it.
From the "Technology Trigger" moment back in 2009 where the guys from Flickr showed us their "10+ deploys a day" presentation it's been 6 years until tool vendors like ServiceNow are being questioned heavily about our thoughts on Process, Product and People
We've all seen the Gartner Hype Cycle (pictured above) and in my opinion customers are now scrambling to the top of the "Slope of Enlightenment".
Whats happened in the meantime? I think the availability of engineers that now have experience in DevOps tooling like Puppet and Chef are reaching more of a critical mass (at least in London, where I hang out). Additionally those tools that were 6 years ago an effort in community support, are now coming of age, and it's possible for enterprises to get support and consultancy on the tools they need to achieve success.
And lastly process workflow and orchestration platforms like ServiceNow that can handle both Development and Operations team requirements are here to meet the challenge!
...and what are customers hoping to achieve by investing in DevOps?
Each company I spoke to in recent months about DevOps was at a different level of capability but the benefits seem to be well understood. Investing in DevOps process, products and people could result in the following 4 characteristics:
- Continuous Delivery of features/components/configuration into production
- Continuous Integration of components
- Heavy use of version control for both code and infrastructure components
- Automated testing to increase speed of deployment
We'll come back to this list in detail later, maybe in another post. It's a scary moment for those customers that I spoke to as they mentally say "Nope, nope, nope and nope!" to the list. Doesn't feel good to look at a list of characteristics of high-performing DevOps organizations and feel like all of these are unattainable.
But as the title of this blog post suggests: "Unicorns are cool, but DevOps horses pull a lot of weight"
As I talk to customers and ask them what they want to achieve from a move to DevOps I get a clear and consistent message. It isn't necessarily about speed or number of deployments. It's about quality and the ability to roll back.
This is consistent with the 2014 State of DevOps report that tells us "High performing IT organizations were more agile and reliable, deploying code 30 times more frequently with 50 percent fewer failures". Almost as one, customers told me that it's the "fewer failures" they are after, although the "more frequently" bit is interesting.
What were you saying about Unicorns and Horses?
The list of 4 characteristics of highly performing IT teams is formidable. All of the customers that I met knew about automated testing, but few were able to say that was a strong point for their teams. The concept of treating infrastructure as code is well known, as DevOps climbs up the "slope of enlightenment", but few of our customers are doing it today.
Here is where I can offer a word of encouragement. In the world of DevOps you can roughly classify companies as Unicorns or Horses
Unicorns are the magical companies, in the enchanted position of having unit tests for all of their applications; Puppet configuration history for each server and a release deployment time measured in minutes rather than weeks. The Unicorns have certainly led the way and powered the DevOps conversation to its current position in the hype cycle.
Horses are the rest of us. All of the customers that I spoke to have to look after legacy applications, hardware that was installed and running mission critical apps long before they joined the company. Server cabinets full of mismatched hardware and different operating systems through mergers and acquisitions.
If your job is running a team that supports a critical business system, but you have no continuous delivery, no continuous integration, no version control and no automated tests... well, I'm sure a lot of people have closed the book on DevOps at this point and gone back to business as usual.
I've written this post to tell you that there is a way forward for the horses - I've heard a lot from customers in this position, and the good news is that a lot of the benefits of DevOps are available to you, no matter where you are starting from.
More DevOps and Service Management blogs
I'm going to be posting here on Mondays, Wednesdays and Fridays on DevOps and Service Management related topics. Please come back and catch the next installment of "Advice for Horses" on Friday.
In the meantime - I have some homework for you:
- Tweet out the link to this article and tag @simo_morris - tell me if you are a horse or a unicorn and whether you identify with what you've read so far.
- If you haven't watched the original DevOps presentation that started this all off - it's only 45 minutes, find the time before Friday
- If you are a ServiceNow customer in London and want to meet to chat about DevOps, or meet online internationally - get in touch
See you on Friday
2 Comments
