My First All Things Open

The list of advantages of #opensource runs long. As a software developer, professionally and for most of my life, I breathe a sigh of relief that the truth about the reliability of most shipped software is out in the open. This is why open source, observability and continuous improvement are such exciting fields to me. Open source software is more reliable, full stop. It is more reviewed, has a more diverse community of contributors, and in some instances, is better maintained. Along with reliability is the benefit of “going further, faster”. Communities of invested contributors tend to turn out content and fix defects at a faster pace. Google Open Source provides a concise response to the question for further context: “Why Open Source?” . It should come as no surprise then that software organizations are increasing their use of open source software by leaps and bounds.

All Things Open , a “polyglot technology conference focusing on the tools, processes and people that make open source possible” has been held in Downtown Raleigh for the past 10 years. I've lived in the Raleigh area for over 20 years and have been a software developer all my life (it feels like) so imagine my surprise when I discovered #ATO22 right here in my own backyard! Our Instaclustr teammates are well versed in the importance of open source and as such, they were already in the know about the conference. I became aware through conversation with Sharan Foga, Director of DevRel and Community Relations for Instaclustr, and my teammate, Will Stowe, Principal Cloud Architect. Instaclustr was a silver sponsor of the conference and our own Paul Brebner, Open Source Technology Evangelist for Instaclustr, had a workshop accepted for facilitation at the conference (a pretty big deal).

Web development was front and center at the conference this year, as I suspect it commonly is, and key innovation was the subject of several keynote discussions. For example, Angie Jones, Global VP of Block’s decentralized web division (TBD ), gave a keynote on decentralized identity and how it can empower users by transferring identity ownership back into their hands. While there was plenty of technical content, everything was infused with humanity. The user experience was a main ingredient of every talk. Keynotes also touched on ethics in tech, the deleterious effects of gatekeeping knowledge, and how important it is to build the next generation of technologists to be better.

My current focus is #DevOps and the cloud, so I was pleasantly surprised to find that both topics had curated content organized into dedicated tracks. I found the scheduler to be convenient and could easily find topics that matched my interest. The sessions were brief, 20-30 minutes long, and never marketing or product promotion. They were either updates on existing projects or introductions to new concepts, and always ended with a call to action and reference for where to go to learn more. Many of the sessions I attended along the DevOps track were run by speakers whose title was either Developer Relations or Developer Advocate. The topics ranged from cultural to theoretical to practical. I listened to the importance of the developer communities and also witnessed a live demo of the construction of a DevOps pipeline for a simple Kubernetes deployment using GitHub actions in under 30 minutes.

Notable Sessions

There were so many great sessions but I want to call out a few that were of particular interest to me.

Why WebAssembly is the Next Wave of Cloud Computing

The only session I attended where I encountered “standing room only” was the session on Cloud Native WebAssembly . The speaker was Matt Butcher, CEO, and founder of Fermyon, a frictionless WebAssembly platform complete with IDE and built-in CI/CD pipeline framework (Spin). Matt has a rich history of open source and application development innovation. He is a co-founder of Helm; the Kubernetes application package manager and he served as maintainer and chair of CNCF Helm.

In this session, Matt covers the evolution from VMs to WebAssembly (#wasm ) and demonstrates, conceptually, where Fermyon would like to go with cloud as the key enabler. Part of the emergence of WebAssembly has to do with containers not being the right answers for all problems, especially in the cloud. Scale to 0 was provided as an example of such a problem but there are others. To address these issues, WebAssembly is emerging as the next wave of cloud computing. Matt asserts that wasm is actually what serverless computing only ever claimed to be but never achieved. Wasm is cross platform, cross language, and cross architecture. Wasm capabilities were on display with the Finicky Whiskers game, a simple application that demonstrates how quickly http payload is sent to a server and the microservices spin up, analyze and spin down. This single application makes use of at least 5 different microservices, written in at least 3 different languages (Go, Rust and Ruby) and makes use of http listeners as well as a Redis queue. The application, however simple, demonstrated nicely the near native performance considering the coordination required between all those moving parts.

One of the interesting concepts for me was the idea of maintaining persistent storage for a collection of wasm components that comprise an application. Add to this the complexity that wasm based applications can also run components side by side with containers as well as VMs and this makes storage even more interesting. Spin provides a mechanism to maintain an application-wide datastore to do just this.

As our colleague Tryggvi Larusson, Technical Director here at NetApp, has told us, wasm is something to keep an eye on. It has big implications for areas we care about such as application development (DevOps), storage management and even cost management to name a few.

Keeping an Eye on your Cloud Native Applications with OpenTelemetry

Another very well attended session was the one on OpenTelemetry . The speaker for this session was fellow Georgia Tech alum, Brad Topol. Brad is an IBM Distinguished Engineer but also a contributor to both Kubernetes and OpenTelemetry and chairs the Kubernetes SIG Docs localization subgroup.

#OpenTelemetry is a “collection of tools, APIs, and SDKs. Use it to instrument, generate, collect, and export telemetry data (metrics, logs, and traces) to help you analyze your software’s performance and behavior.” It is a way to enable effective #observability which is exceedingly important in the time of microservices and distributed architectures. The OpenTelemetry open-source project is the result of the 2 CNCF projects OpenTracing and OpenCensus. The goal of the project is to standardize the gathering and exporting of data that allows the observer to determine the state of an application at any point in time in the face of any variety of unknowns.

As an open-source project, OpenTelemetry has robust language support and is natively supported and contributed to by an impressive list of organizations . The standardization prevents vendor lock in and offers flexibility in consumption with key backend integrations.

Brad focused mainly on the signal receiver, exporter and collector components of OTEL and took the audience through a handful of walkthroughs. These components have agentless as well as serverless options. Brad emphasized the importance of standardization of data formats and API for telemetry data with the rise of the cloud as the de facto platform for containerized applications. While there do still exist some proprietary protocols for data transport, OpenTelemetry Protocol (OTLP) is gaining traction and is vendor agnostic.

This project is of great interest to me in that it significantly advances debugging and triage capabilities. Solutions like this help address alert fatigue and democratizes software problem solving. Collecting high cardinality, high dimensionality data removes the “reproduction” requirement and enables engineers at varying levels of competency to ask the right questions to get to the heart of even the most complex issues. That’s cool.

Conclusion

In addition to the sessions and the stickers (did I mention the stickers? You will never see a better collection of stickers) the vendor hall was a “who’s who” in DevOps and automation. Just about every company on Digital.ai’s periodic table of DevOps tools was there. I talked to the folks from Perforce, whose products I've used for years, and learned that they have an open source management service offering (OpenLogic). They’ve added integrations to their portfolio that helps their customers manage their open source components. Perforce open sourced it’s cli some years back. I bring attention to this because it is an example of how companies are finding new and interesting ways to bring value to software development organizations based on the increase in use of open source.

And as can be expected there was serious recruiting going on. Groups looking for devs that are well versed in open source would do well to recruit at this conference. Open source talent is still in very high demand and there was no shortage of eager would-be recruits in attendance. I spent a little time chatting with Major League Hacking, an official student hackathon league that organizes open source based internships for qualifying students. They’ve been very successful in establishing partnerships with major companies and finding rich work experience for the students.

As we continue to explore the impact of community, Developer Relations and opportunities in open source and DevOps, conferences like All Things Open are a great resource for everyone to learn and get involved. As purchasing decisions continue to shift to the professionals that are closer to the actual application development itself, exploring ways to connect with this particular audience becomes increasingly important. Events like ATO most certainly help in that effort as is demonstrated by the "by the numbers" graphic. As innovation continues to accelerate in the area of application development and software development lifecycle, creative strategies to support the data needed by these cutting-edge solutions need to keep pace. I want to make sure NetApp is well positioned to innovate in lock step.

I had a tremendous experience, made many impactful connections, and learned a bunch. Can’t wait for the next one.