There are two sides to every coin, so the old adage goes. It goes without saying that both sides of the coin are the same color even if they bear different images.
So we discovered when we dove into the middle of DevOps and NetOps. We found that both groups are far more generous in their opinion of the other with respect to prioritization of efforts than traditional archetypes purport them to be, and they have a lot in common. Even though they may disagree on details.
We surveyed 884 NetOps and DevOps professionals during July 2017. We asked primarily about automation efforts, but also about those measures often cited by DevOps such as frequency and success rate of deployments.
But those are just cold numbers. We wanted to know whether the perceptions held by each group were in line with the animosity portrayed by caricatures or had the two moved on and become if not fast friends, then at least frenemies. After all, both groups have the same goals – get an app to market that will propel the business to new heights.
And from the results, it certainly seems so. Both desire greater collaboration and interaction when it comes to moving apps to market, offering up specific examples as to just how they’d like to see that fall out. And both groups share common perceptions of the security, reliability, and performance of the applications they’re delivering and then deploying. Perhaps surprisingly, neither is immune to the impact of the pace of change in the industry. DevOps and NetOps alike are not entirely confident their roles will be relevant in five years, and both identified at least some gap between what they need to know to do their jobs and the skills/training they currently possess.
But we did uncover differences as to just how much technology should bridge the divide between development and production pipelines. That, in turn, validates a widely held belief that “the network” is a primary factor in the rise of multi-cloud as it drives developers and DevOps to seek solutions outside IT, primarily in the cloud.
That’s a direct quote, and a common theme amongst those DevOps in the minority who believed their counterparts in NetOps were not prioritizing the “right things.” That could be because they were pitted against the small number (8%) of NetOps who indicated “None” of their production pipeline was currently automated. Or perhaps they operate as part of a small group of DevOps (4%) who believe developers and DevOps and development should have NO access to the production pipeline via automation/self-service capabilities, and simply disagree.
The good news is they are in the minority on both counts. The majority of NetOps not only operate in environments where 50% or more of the production pipeline is automated, but a plurality (61%) also believe DevOps and developers should have greater access (50% or more) to that pipeline.
That means NetOps are squarely on board with DevOps' desire for more automation and self-service capabilities.
Furthermore, both groups overwhelmingly believe the other prioritizes “the right things,” with 82% of DevOps and 76% of NetOps giving a head nod of approval to their counterparts' prioritization. Not much call for duels at dawn in the parking lot after all.
Greater Automation, Greater Chance of Success
A common belief within DevOps circles is that automation not only enables greater frequency of delivery and deployment, but improves its overall success rate. Whenever manual intervention is required, the chances of errors creeping in increases. Human error is a significant factor in many an outage, after all. It’s not only logical, it’s borne out by our survey. Comparing deployment success rates against the percentage of deployment pipeline current automated yields a positive correlation – at least most of the time.
Based on the data it seems that there exists a “sweet spot for success”: deploying once a week and automating at least 75% of the pipeline.
Our survey found that in spite of growing agreement on automation, there were distinct differences of opinion. One of these differences appears with respect to core concepts required to bring about the real changes needed to succeed in a fast-paced, multi-cloud world driven by an application economy.
NetOps were, on average, twice as likely to describe current deployment frequencies as “too frequent” (8%) than their DevOps counterparts (4%). Conversely, far more DevOps were dissatisfied with the frequency of delivery, with 26% describing theirs as “not frequent enough.”
Still, clear majorities in both groups felt their current speed was “good enough for them.”
Of the 74% of NetOps who declared the frequency of deployments “good enough,” 37% deploy once a week. Interestingly, of the 9% who declared their deployment frequency “too frequent,” 44% are also deploying once a week. The majority of the 18% of NetOps who want to deploy more frequently are deploying either once a week (27%) or once a month (28%).
There is little disagreement between DevOps and NetOps on the importance and impact of automation on the application lifecycle – from development to delivery to deployment. There was a positive correlation between how NetOps rates the reliability, performance, and security of applications and automation of the pipeline. There is more confidence in the reliability and performance ratings of applications when the pipeline is at least 50–75% automated.
Confidence in application security ratings is stronger when 75% and more of the pipeline automated. Unsurprisingly, DevOps seem similarly impacted by pipeline automation. They were most confident in their ratings of application reliability and performance when more than 75% of the pipeline is automated.
Based on the results of our survey (you can get the full report here) the general consensus around automation is definitely two thumbs up – one from NetOps and one from DevOps.
Automation is important, no matter which side of the coin you’re looking at.