I've been fortunate to witness the internet revolution firsthand. I remember when deploying a web service meant buying and installing expensive hardware, installing and configuring indecipherable server software, and endless worry about not knowing what I didn't know. Today, I can (we all can) deploy a website and HTTP API with a few clicks, pay only for what we use, and scale up without a second thought.
This accessibility sparked an explosion of innovation that changed everything about how we build software, and the software that we build.
But after the past several years working with a Wi-Fi provider, I noticed something troubling. The world of network services -- "low level" protocols like DNS, DHCP, MQTT, RTSP/RTP, RADIUS and more, on which authentication, media streaming, games, IoT are built -- hasn't enjoyed the same revolution. While the web moved forward, many of these critical network services are still running on aging software with little beneficial innovation.
The problem isn't talent or ideas -- it's that the barrier to experimenting with non-HTTP services is simply too high. The time, money and/or risk involved in trying something new are prohibitive for most developers and organizations.
I've witnessed teams sit on great ideas because setting up a test environment, building server infrastructure, and understanding the complexities of protocols like UDP required too much investment for something that might not work.
My vision for Proxylity is to bring the magic of quick, low-risk development to network software and help foster innovation in areas that desperately need it.
I'm starting with what people keep telling me is a "niche" -- UDP. But it's a niche that powers critical parts of our digital experience and desperately needs better tooling. Media streaming, observability, real-time games, IoT -- they all depend on UDP. In fact, you're probably reading this article using UDP since HTTP/3 is based on QUIC, which is based UDP. If only we all could take the risk Google and other giants did in developing QUIC.
UDP is a perfect base to build from for Proxylity -- it has interesting problems and high constraints that will help me learn as I build toward supporting all message-based network protocols. The lessons from building "carrier grade" UDP solutions will inform everything else we do.
My goal is simple: I want to see what happens when we lower the barriers to network innovation. Really, really lower them.
What if you could develop a custom DHCP service as easily as you build a REST API today? What if testing a new DNS delivery method was a afternoon project rather than an expensive infrastructure commitment? What if your game networking layer could be serverless, and free when not in use?
I don't know exactly what people will build when we make these tools more accessible -- and that's what excites me most. I'm looking forward to seeing what you create!
If you're working on something interesting in this space or just want to chat about the future of network services, I'd love to hear from you.
Thanks for reading!
-- Lee