
HTTP3, the latest version of the internet’s core data transfer protocol, represents a fundamental shift in how devices communicate. While previous versions relied on the transmission control protocol—which ensures reliable data delivery through careful packet tracking and error checking—HTTP3 embraces the user datagram protocol, a lighter-weight approach that prioritizes speed over perfect delivery. This architectural change fundamentally reshapes how engineers approach system design.
Coupled with the rise of web real-time communication (WebRTC), these protocols promise real-time and direct device-to-device communication for embedded systems. In practice, this translates to smart home security systems with near-instantaneous video or industrial sensors that transmit critical data without lag.
Such great potential, however, demands big decisions from businesses. Executives must therefore ask themselves: How will the rise of UDP shift impact our roadmap? What infrastructure investments are needed? And how can we balance the promise of improved performance against the risks of early adoption?
These three-letter acronyms might not mean much to the average person, but they dictate how data traverses the internet. Therefore, replacing TCP with UDP represents a fundamental change in how the internet speaks and, as a result, how we speak.
I have previously compared TCP to a telephone connection. There is an initial “handshake” to set up the connection and then successfully place a “call” for two-way communication. It is reliable but takes time to get started. In comparison, HTTP3 uses quick UDP internet connections (QUIC), a multiplexed transport protocol built on UDP that is like a high-speed conveyor belt continuously sending multiple packages to their destinations without waiting for confirmation. This method allows for faster and more efficient delivery, even if the occasional “package” goes missing.
This shift reshapes internet traffic in profound ways. For one, UDP eliminates the need for TCP’s time-consuming “handshake,” lowering latency in a win for real-time applications. Moreover, UDP’s ability to continue data transmission even when some packets are lost proves invaluable in areas with unreliable connectivity. Additionally, HTTP3’s multiplexing capabilities allow for multiple data streams over a single connection, a feature that reduces overhead and boosts efficiency.
Despite launching in 2022, connected device developers are only scratching the surface of integrating HTTP3 in the real world. And this is in addition to other new protocols like WebRTC gaining ground in IoT.
In tandem with the rise of video surveillance over the past few years, it is hard to ignore the growth of WebRTC in IoT. Originally designed for browser-based real-time communication, WebRTC also moves the needle toward UDP with improved latency, enhanced reliability and efficient peer-to-peer communication. This is something we have seen at Nabto with a few video device customers who want to bypass middlemen servers and stream directly over browsers.
Like HTTP3, WebRTC leverages UDP as its foundation, specifically using stream control transmission protocol (SCTP) built on UDP. SCTP provides message-oriented communication with multiple streams, flow control and congestion avoidance mechanisms.
One protocol favoring UDP over TCP might be chance. But two? That is the start of a trend. Developers are clearly responding to market conditions and seeking protocols that combine TCP-like reliability and congestion control with the speed and flexibility of UDP. Whatever way you look at it, UDP is on the up and bringing faster, more secure and more efficient device communication to the sector.
While the one-two punch of HTTP3 and WebRTC promises reduced latency and improved reliability for real-time applications, it also introduces new challenges. Successfully moving with this protocol shift therefore requires understanding the trade-offs and implications.
On the infrastructure front, adoption faces some challenges. While server software can be updated quickly, consumer-grade hardware like routers and firewalls evolve more slowly. Many homes still run on network equipment designed a decade ago, with TCP-optimized software that may not handle UDP efficiently. Moreover, load balancers—crucial for high-traffic websites—currently lack robust UDP support. This creates a complex landscape where modern protocols must coexist with legacy infrastructure.
My advice? Prepare and proceed with caution. HTTP3’s already been around for a few years, but the winds of protocol change are blowing harder. Likewise, QUIC and WebRTC are both from Google, and one can expect the tech giant to build out their respective ecosystems to address the above issues and make protocol transitioning easier or incentivized. Another plus? Both protocols also come with built-in encryption.
If you do change your device protocol, a measured approach is key. Start by assessing your infrastructure and identifying which applications would benefit most from UDP’s characteristics. Invest in building expertise within your development team and implement any changes gradually. Begin with non-critical applications to gain experience, mitigate risks and go from there.
As UDP reshapes internet communication, IoT companies must balance preparation with progress. Assess, identify, invest and implement. If your industry demands low latency and secure transmission, start building your expertise today while the infrastructure catches up tomorrow.
Carsten Rhod Gregersen is the founder and CEO of Nabto.
You must Register or Login to post a comment.
This site uses Akismet to reduce spam. Learn how your comment data is processed.
Advertisement

More Stories
Anatomy of a Scam
Climate and Environmental Sustainability Within the IETF and IRTF
From Commitments to Practice: Internet Society’s Priorities for WSIS+20 Implementation