HTTP/3 and IPv6: Which Hosting Buzzwords Actually Matter When You're Paying Nothing
Photo: Shixart1985, CC BY 2.0, via Wikimedia Commons
Open the features page of any hosting provider and you'll be hit with a list that reads like a networking textbook index: HTTP/3, IPv6, TLS 1.3, Brotli, LiteSpeed, QUIC, Edge Caching, Early Hints. It's designed to make the $30 plan look like a worse deal than the $3 plan, and the free plan look like it's missing everything.
Here's the strategic reality: for a five-page small business site with a few hundred visitors a day, most of these decide nothing. A small handful matter a lot. Sorting them is the whole game.
HTTP/3: real, but you can get it for free anyway
HTTP/3 runs over QUIC instead of TCP. Translation: it handles packet loss better. On a stable office connection you won't notice. On a phone with two bars in a parking garage, it can mean the difference between a page that loads and one that spins.
Here's the part that annoys budget hosts: you often don't need to upgrade your plan to get it. Put a free Cloudflare account in front of your site and HTTP/3 is available on the free tier. If your host is behind a CDN that already supports it, you have it and don't know it.
Verdict: nice to have. Rarely a dealbreaker. Free to acquire through a CDN either way.
IPv6: invisible until it isn't
The internet ran out of IPv4 addresses years ago. IPv6 exists so that doesn't matter. Most US mobile carriers are well down the IPv6 path, and a growing share of traffic is IPv6-first.
What happens if your host is IPv4-only? In most cases, nothing — carriers run translation layers that make it work. In a few edge cases, mostly on certain mobile networks and corporate connections, reaching an IPv4-only site is slower or occasionally fails.
Verdict: nice to have, occasionally matters. If a host in 2026 can't do dual-stack, note it as a yellow flag about how much they've invested in infrastructure.
HTTP/2: the one that should be non-negotiable
HTTP/2 multiplexes requests over a single connection. Without it, a page with twenty images and three scripts queues up like a single-lane drive-thru. This is not exotic technology — it's been standard for a decade.
If a host is still serving HTTP/1.1 only, that tells you something about the age of their stack. That's not a performance footnote. That's a signal.
Verdict: dealbreaker if missing.
TLS 1.3 vs. TLS 1.2: smaller difference than the marketing claims
TLS 1.3 cuts a round trip off the handshake, which saves a few milliseconds. Good. But the far more important question is whether the host gives you HTTPS at all, and whether it renews the certificate automatically without you thinking about it.
A site with TLS 1.3 and a certificate that lapses in March is worse off than a site with TLS 1.2 and automated renewal.
Verdict: automated HTTPS is a dealbreaker if missing. The TLS version is a nice-to-have.
Brotli compression: small win, free win
Brotli compresses text files about 15–20% better than gzip. On a page of HTML, CSS, and JavaScript, that's a modest but real reduction in bytes shipped. Most modern stacks enable it by default. You can verify it in five seconds (see the testing section below).
Verdict: nice to have. Almost always already on.
The server software debate: mostly noise at your scale
LiteSpeed vs. Apache vs. Nginx is a legitimate argument for a site doing serious traffic with dynamic PHP. For a static or near-static small business site, the difference is dwarfed by two other factors: where the server physically sits and whether there's a caching layer in front of it.
A well-cached site on an old Apache box in the right region will beat an uncached site on the newest LiteSpeed hardware every time.
Verdict: marketing noise for your use case. Ask about caching and server location instead.
What actually moves the needle (and the rankings)
Google's Core Web Vitals — LCP, INP, and CLS — are what show up in search tooling. None of them are protocol features. They're outcomes. The inputs that matter most for a small site, ranked:
- Time to first byte. Driven by server location, caching, and whether the host is oversold. This is where cheap plans actually fail.
- Image weight. A 4MB hero JPEG will destroy your LCP regardless of HTTP/3.
- Third-party scripts. Chat widgets, ad pixels, and font loaders cost more than any protocol upgrade saves.
- Render-blocking CSS and JS. One badly placed script tag outweighs every item on the buzzword list.
Protocols are the last 5%. Most small sites are losing the race in the first 50%.
How to test what your host actually supports
Five minutes, no paid tools:
- Open Chrome DevTools → Network tab → right-click the column headers → enable Protocol. Reload. You'll see
h2,h3, orhttp/1.1per request. - Run
curl -I https://yoursite.comin a terminal. The response headers showalt-svc: h3if HTTP/3 is advertised. - Check the response headers for
content-encoding: br— that's Brotli working. - Visit ipv6-test.com to see whether your domain resolves over IPv6.
- Run PageSpeed Insights on a real mobile connection profile, not desktop.
If you're on a free plan and the results look grim, the fix is usually not a new host. It's a free CDN in front of the existing one, compressed images, and one fewer script.
The short version
Dealbreakers: no automated HTTPS, HTTP/1.1 only, no caching layer, server on the wrong continent for your customers.
Nice to have: HTTP/3, IPv6, Brotli, TLS 1.3, Early Hints, whatever the new acronym is next quarter.
Buy for the dealbreakers. Get the nice-to-haves free from a CDN. And spend the money you save on something that actually changes your customer's experience — like a domain that's easy to say out loud.