Network Reliability Engineering - Getting The Word Out
May 7, 2019
It's been over a year since I started at Juniper, and nearly 1.5 years since I started having discussions with industry peers (and now customers) about this emerging Network Reliability Engineering term. We've had some internal iterations on it as well, identifying key behaviors that we feel can be adopted by the broad enterprise market, not just the hyper-scale companies from which a lot of these ideas tend to originate.
In an industry challenged by ambiguity and lack of clear direction when it comes to automation, I still feel very strongly that Network Reliability Engineering represents the true nature of what those that have automated in production have always been about. The idea that existing skill-sets matter in this new world, coupled with a new "true north" of reliability as the goal of automation we've been striving for all along, means we can start having very real, very tangible discussions about skill-sets, and moving beyond the hyperbole of automation as a "job killer" or simply a vehicle to do things more quickly.
After all, there are really good reasons why our customers first came to us with this term. For our part, we're maniacally obsessed with highlighting the bright spots within our customers and helping the industry as a whole learn from them. That's one of many reasons we're investing heavily into building the vendor-agnostic NRE Labs community as a vehicle for those who want to learn a better way of running network operations to do so with as few barriers as possible, and for those with bright spots to share, to be able to do so in an unbranded, community-first manner.
Getting The Word Out
I was very pleased to have the opportunity to discuss Network Reliability Engineering at length on a recent Short Take by the Network Collective podcast, where we dove into some of the specifics of NRE, and of course, what we're putting out there into the community to make this real for customers and industry advocates alike:
This was a great conversation, and I want to thank Jordan for having me on. It's always fun to chat with those folks in general, but especially regarding this topic, which has been near and dear to my heart for over a year.
I also feel like moving the industry forward requires tackling the problem from multiple ends. We are and will continue to invest in augmenting the skill sets of network engineers in our industry with automation and reliability engineering process and culture. However, while my background certainly contains your garden variety, meat-and-potatoes network engineering, I have also spent a lot of time on the "other side of the fence" in the world of software development. I know first-hand that there are a lot of misperceptions about infrastructure in this space; especially networking - which is often still thought of as a form of dark magic. As the meme goes:
In addition to breaking down these barriers, I also think it's important for application owners and developers to understand a little bit more about the challenges in networking, and just how different the network engineering domain has become from just about any other area of IT. Many network engineers could benefit from sitting with developers and just listening to them talk about their processes and workflows, but without a common vocabulary, it's not a natural conversation to have. Many of the network engineers I talk to these days assume they'd have as much in common with software teams as they would with accounting, or legal.
So while we're going to continue pushing for augmentation in the networking discipline, I am striving to get the word out about networking within the Cloud and SRE communities. In the age of distributed applications and microservices, the argument can easily be made that the network is more important now than it ever has been, and there are changes a-brewing in the networking industry that developers and cloud architects should be aware of, so that we can work more closely together in an ever-more-rapidly changing industry. One way I've been fortunate recently to have a platform for this is when Derick Winkworth and I sat with Brian at the CloudCast to talk about some of these challenges and what we're planning to do about it:
As I mentioned, while talking about NRE is a big part of what we want to do to help spread the success of the bright spots we're seeing, so too is NRE Labs a community we're continually investing in, not only as a way for us to prove that we mean what we say, but of course, to vigorously and with extreme prejudice, knock down the barriers that have made automation a challenge for so many to adopt for so long. I discussed this at length at the recent Network Field Day event, and if you haven't seen NRE Labs in action yet, you really owe it to yourself to see the video below.
To me, there are a few common themes here that I think should be highlighted:
NRE is more than just "the new automation". It's a new operational model that centers around the idea that reliability is the crux of why we should be doing everything, including automation.
There's a time and a place for technology and products, but NRE is the natural outcome of the need to talk about process and cultural improvements in today's network operations.
These new principles are for you. There's no room here to assume that this is just something for the hyperscalers, or people with "dev" in their title - this revolution is ours to own.
This is not just talk for us. This isn't real for us until we put skin in the game, and that's exactly what we're doing with the vendor-independent NRE Labs community, and we are committed to continuing to invest in this as a symbol of our commitment to the betterment of the industry.
We're Just Getting Started
A year is a long time. We've already learned a lot, and expect to learn more, as we get more folks on the NRE bandwagon, and we continue to make progress on these practices in production with our customers and across the industry. You can expect us to continue to champion and refine the principles we stand for, with the express intent to move the industry forward.
To join us on this journey, I encourage you to check out the community we're building around NRE Labs and the open source project that powers it - Antidote. Your feedback and involvement are not only encouraged but to us, it's the whole point. Our desire is to be the amplifier that continues to get the word out about the bright spots that are taking place all over the industry. Join us, and together, we'll make networking the envy of the entire IT discipline.