Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
HTTP/3 runs over QUIC and UDP instead of TCP, with encryption built into the transport. It improves privacy and performance while making network inspection much harder.

What Is HTTP/3 and How Does It Affect Security?
HTTP/3, published as RFC 9114 in June 2022, keeps the semantics of HTTP — methods, headers, status codes — and replaces everything underneath. Instead of TCP with TLS layered on top, it runs over QUIC, a transport that uses UDP and builds encryption into the protocol itself. There is no unencrypted HTTP/3. By 2026 roughly a third of all websites support it, and the major browsers and CDNs negotiate it by default.
Three differences matter.
Head-of-line blocking is gone at the transport layer. HTTP/2 multiplexed many streams over one TCP connection, so a single lost packet stalled every stream waiting for retransmission. QUIC tracks streams independently, so loss affects only the stream it belongs to. On a mobile connection with real packet loss this is a substantial improvement.
Connection setup is faster. QUIC combines the transport and cryptographic handshakes, giving a one round-trip connection, or zero round-trips when resuming a previous session.
Connections survive network changes. A QUIC connection is identified by a connection ID rather than by the four-tuple of addresses and ports, so a phone moving from Wi-Fi to mobile data keeps the same session instead of rebuilding it.
Header compression moved from HPACK to QPACK, which avoids the ordering dependency that would otherwise reintroduce head-of-line blocking.
Encryption is not optional and covers far more than it did before. In TLS over TCP, the transport headers — sequence numbers, acknowledgements, flags — are visible and modifiable by anything on the path. QUIC encrypts almost all of that, leaving only a short header with the connection ID. The practical consequences are good: no RST injection to tear down a connection, no transport-level manipulation by middleboxes, and much less metadata leakage. Protocol ossification, the phenomenon where middleboxes prevent any change to a protocol because they parse it, is largely designed out.
Amplification is handled deliberately. Before validating a client address, a server may send no more than three times the bytes it received, which prevents QUIC becoming the next reflection vector. Address validation tokens handle the rest.
The visibility cost is real and needs planning for.
HTTP/2 Rapid Reset, tracked as CVE-2023-44487 and disclosed in October 2023, abused stream multiplexing by opening and immediately cancelling streams to exhaust server resources. HTTP/3 shares the multiplexing model and is affected by the same class of attack, so stream concurrency limits and request rate controls must be configured for it explicitly rather than inherited.
QUIC flood attacks have also appeared, using the handshake as a resource exhaustion vector. And zero round-trip data carries an inherent replay risk, so 0-RTT should be restricted to idempotent requests — never to anything that changes state.
Decide deliberately rather than by default. If your inspection stack cannot handle QUIC, blocking UDP 443 outbound is a legitimate interim control, and clients will fall back to HTTP/2 cleanly; just be aware you are trading user performance for visibility and that the trade should be revisited. If you enable it, confirm your next-generation firewall and proxy actually parse QUIC rather than pass it, set stream limits on your own servers, restrict 0-RTT, and make sure application-layer logging is shipped to your SIEM.
On your own public services, enabling HTTP/3 is generally a performance win with no security downside, provided rate limiting is applied at the edge. Our Web Application Security Services test HTTP/3 endpoints alongside the rest of the application stack.
In transport terms, yes. Encryption is mandatory and covers most transport metadata, which prevents connection injection, reset attacks, and middlebox tampering that are possible with TCP. Application-layer risks are unchanged: injection, broken access control, and logic flaws behave exactly as before. The main trade-off is reduced network visibility for defenders rather than reduced security for users.
It is a defensible choice if your inspection tooling cannot parse QUIC, because clients fall back to HTTP/2 over TCP without errors. You lose performance for users on lossy connections and you push a growing share of traffic onto an older path. Treat it as a temporary measure and reassess when your firewall or proxy gains proper QUIC support.
Only if the WAF sits inline at the network layer. A WAF deployed at a CDN or reverse proxy that terminates HTTP/3 sees fully decoded requests and works normally. A network appliance that parses TCP streams does not see QUIC traffic at all. Check whether your WAF terminates the connection or inspects it in passing, because the answer determines your coverage.
Data sent in the first flight of a resumed connection can be captured and replayed by an attacker, because there is no round trip to prove freshness. For a GET that only reads data this is harmless. For anything that changes state, such as a payment or a configuration change, it is not. Restrict 0-RTT to idempotent requests at the server.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.