NGINX Conf 2018 kicked off in Atlanta yesterday, with hundreds of developers, operators and architects in attendance to explore how the web, cloud, and microservices are intersecting. Thousands of people also tuned in to the livestream from around the world, making it the biggest NGINX conference yet (videos of the keynotes from both days are now available on demand).
As NGINX CMO and event emcee Rob Whiteley outlined, this year’s theme is the Journey to Microservices. We looked at how everyone from traditional enterprises to digital pioneers to cloud‑native companies are undergoing modernization initiatives. The goal? Mastering digital disruption.
This was my fifth Conf and I think it was the best lineup of speakers and content we’ve had yet. Looking back at Day 1, I took away these five things.
Takeaway #1: It’s a Digital World Now, And You Have to Master Digital Delivery
I led off the morning with a call to arms: it’s a digital world we live in, which means that all companies are becoming tech companies, whether we like it or not. We have to put technology at the core of what we do. In the last year, four billion people were online globally, representing more than half of the entire planet. With those people spending roughly six hours a day connect to the Internet, this will represent one billion collective years online in 2018.
It’s driving tremendous amounts of new behavior and paradigm shifts in our economy. One‑sixth of consumer spending – $1.5 trillion! – now takes place online. And we’re each consuming 3 GB of bandwidth each month on our phones. This has increased 50% year‑over‑year, which is putting tremendous new pressure on telecommunications and IT infrastructure.
The demands to get the online experience right for consumers are huge – they want instant satisfaction. But when the experience is right, the benefits are huge: consumers are happy to spend 15% more for a seamless customer experience.
To survive and thrive in this environment, companies have to rethink digital delivery. We’re seeing physical and digital delivery mechanisms converging. You need a platform that helps you optimize the elements of delivery you can control – performance, reliability, resiliency, and security – as well as the flexibility and agility to react quickly to what you can’t control.
Takeaway #2: Customers Mastering Digital Delivery Have a Competitive Advantage
At NGINX, we’re lucky to have some pretty amazing customers. We’re even more lucky that they’ll join us on stage at Conf and tell their stories. This year, Jamie Panagos from Charter Communications and Josh Ryder from AppNexus joined me on stage during my keynote.
Jamie talked about how the recent merger between Time Warner Cable and Charter Communications resulted in a multi‑year effort to change how video is delivered to customers. To provide the best experience, Jamie was tasked with building a CDN so that Charter owns the end‑to‑end digital experience. He talked about how NGINX Plus is part of the pluggable architecture that gives Charter significantly more control over digital delivery, while also reducing costs and improving flexibility.
Josh picked up on the same theme and described how AppNexus – recently acquired by AT&T – also relies on NGINX Plus to reduce costs, improve customer experience, and deliver high performance. The conventional wisdom is that you can only pick two of these benefits, as Josh quipped. But he described how with NGINX Plus AppNexus gets all three. While saving millions of dollars by replacing hardware load balancers, the company also have gained new capabilities to support their ad auctions and serving, which all have to happen in 100ms or less!
Takeaway #3: You Must Reduce Complexity By Collapsing Infrastructure Tiers
Sidney Rabsatt, NGINX’s VP of Product Management, used NGINX’s own research to show how the microservices journey breaks down.
About 60% of companies are still using monolithic, legacy applications. Roughly 30% of the market is companies currently looking to modernize by building external microservice add‑ons into their existing systems; as an example, Sidney described a retailer that uses the Google Maps API to provide a store locator on its website. The last 10% of companies have only microservices, mostly from cloud‑native providers and companies that just started developing apps in the last 5 to 10 years.
The specific tech needs vary for companies at each stage of the process, making today’s app infrastructure increasingly fragmented. Sidney announced how the NGINX Application Platform can be deployed as a dynamic application gateway, which greatly reduces complexity by supporting both monolithic and microservices application delivery. The dynamic application gateway consolidates several infrastructure components – API gateways, reverse proxies,and load balancers – and manages all of them through the same control plane. It’s intelligent and distributed and absorbs spikes in demand without system failure.
Takeaway #4: Building Distributed Services Changes How Your Teams Need To Operate
Isa Vilacides, Director of Quality and Productivity Engineering for CloudBees, outlined how the shift from monolithic applications to microservices changes engineering teams from large groups all working on the same codebase to smaller teams each with its own narrowed responsibility.
These new possibilities around agility bring new considerations around preventing bugs. As Isa said, while we ultimately want to go fast without breaking things, things will break. So when things go wrong in production, the priority becomes being the first one to know why. And when teams are smaller and distributed and often spread out across geographies, it can become much harder to find out.
Isa talked at length about consumer‑driven contract testing as an answer to this challenge. It’s an additive step that complements traditional unit, integration, and end‑to‑end testing. Consumer‑driven contract testing helps with canary releases and dogfooding, methodologies that give developers more accountability for their microservices in production. Isa also talked about how your efforts have to shift to monitoring so you can catch the inevitable breakages that occur. She concluded by saying adopting microservices is ultimately a communication and collaboration issue, and to be cautious about undertaking it if you’re not ready to address the human side of change.
Takeaway #5: Enterprises Can Modernize with a Three‑Step Approach
Lakshmi Sharma, Head of Product Management for Google Cloud, broke out the increasingly complex software environment that established enterprises have to navigate: a mix of legacy apps hosted on premises, private hosted clouds, and public cloud products. It has created systems that are hard to migrate and can be a deterrent for companies that would ideally like to be working toward more greenfield solutions.
Companies do want to modernize, but as Lakshmi outlined, their requirements are significant – they want minimal down time while they re‑architect monolithic architectures, reduce technical debt, and create a single interface to seamlessly manage a broad amount of inputs. Unfortunately, too often they set out to do this by building complicated new structures on top of their legacy infrastructure. No matter what you add in, if the backend is accessing data and compute storage in the same way as it always was, your problems aren’t going away.
Luckily, Lakshmi spelled out a clear path to modernization. The first step is containerization, packaging applications in a platform‑independent manner. Next is orchestration, learning how to deploy the application and using some sort of CI/CD tool to manage apps. The final step is a service mesh, connecting and securing apps and new services into the infrastructure. Lakshmi urged companies to adopt only the “layers” of this stack that make sense for their business. For example, not all companies need the sophistication of a service mesh but most will benefit greatly from the portability and agility of Docker and Kubernetes.
Mapping NGINX Products to the Microservices Journey
We’re proud of the pace of innovation here at NGINX, and we use our annual conference to announce new open source advancements and commercial products.
You can see the recap of the announcements we made on Day 1 in yesterday’s announcement blog. However, the big takeaway for me is that the creative and varied ways customers adopt our technology is truly astounding.
I’m excited to see how our new API management solution and upcoming service mesh solution in NGINX Controller, dynamic‑clustering features in NGINX Plus R16, and TLS and Node.js support in NGINX Unit will help customers undertake the journey to modernize and adapt to the new digital world.
[Editor – The API management solution in NGINX Controller is now API Connectivity Manager, part of F5 NGINX Management Suite. NGINX Service Mesh is a free, stand‑alone offering.]