--- title: HTTP/3 is faster with some work and help from Cloudflare description: Delivery Hero cut its apps' end-to-end home screen response time by 17 percent after moving most users to HTTP/3. Chidera said handshake latency dropped 51... canonical: https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: HTTP/3 is faster with some work and help from Cloudflare | daily.dev og:description: Delivery Hero cut its apps' end-to-end home screen response time by 17 percent after moving most users to HTTP/3. Chidera said handshake latency dropped 51... og:url: https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd og:image: https://api.daily.dev/og/posts/YVmoTvWWd.png og:image:alt: HTTP/3 is faster with some work and help from Cloudflare og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # HTTP/3 is faster with some work and help from Cloudflare **[DEVCLASS](https://daily.dev/sources/devclass)** · 5 min read · 0 upvotes · 0 comments ## Summary Delivery Hero cut its apps' end-to-end home screen response time by 17 percent after moving most users to HTTP/3. Chidera said handshake latency dropped 51 percent, and Pakistan's response time was cut in half. Moving to HTTP/3 via Cloudflare's CDN was simple, but Android's Cronet library took about 800ms to initialize at P90, so the team forked it into their own bridge using OKCronet and overlapped initialization with other background tasks, cutting P95 latency from 1.2 seconds to 120 milliseconds. Secondary screens in packet-loss-prone markets like Pakistan still suffered, which Delivery Hero's engineers attributed to Cloudflare's choice of the Cubic congestion control algorithm over BBR, and Delivery Hero is working with Cloudflare to resolve this. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much did switching to HTTP/3 reduce handshake latency for a global mobile app? Delivery Hero cut handshake latency between its apps and servers by 51 percent after moving to HTTP/3 via Cloudflare's CDN. In Pakistan, where networks previously took nearly a second at p90 to complete a secure handshake, response time was cut in half. Overall P95 latency dropped from 1.2 seconds to 120 milliseconds after additional engineering work. _Weighing an HTTP/3 migration? daily.dev surfaces real production benchmarks like this one._ ### Why does Google's Cronet library slow down HTTP/3 on Android instead of speeding it up? Cronet, Google's networking library used to add HTTP/3 support on Android, took approximately 800 milliseconds to initialize at P90, which ate into the 0-RTT speed benefits HTTP/3 is supposed to provide. Delivery Hero's engineers worked around this by forking the integration into a custom bridge called OKCronet and overlapping Cronet's initialization with other background tasks rather than waiting on it. _Debugging slow Android network startup? daily.dev tracks real fixes for Cronet-style initialization lag._ ### Why does Cloudflare's Cubic congestion control algorithm perform worse than BBR on lossy networks? Cubic, the congestion control algorithm Cloudflare uses for HTTP/3, historically handles packet loss poorly, causing latency issues on secondary app screens in markets prone to packet loss like Pakistan. Testing with Google HTTP/3-aware servers using BBR (Bottleneck Bandwidth and Round-trip propagation time) instead avoided getting hung up on the same packet loss, and Delivery Hero is working with Cloudflare to resolve the issue. _Choosing between Cubic and BBR for lossy networks? daily.dev tracks how teams resolve congestion-control tradeoffs._ ## Similar posts on daily.dev - [HTTP1 vs HTTP2 vs HTTP3](https://daily.dev/posts/http1-vs-http2-vs-http3-wbpqmgwlo) · System Design Newsletter · 170 upvotes · 5 comments - [HTTP/2 and HTTP/3: what changed and why it matters](https://daily.dev/posts/http-2-and-http-3-what-changed-and-why-it-matters-kxd7rny7x) · Flavio Copes · 1 upvotes · 0 comments - [EP200: HTTP/2 over TCP vs HTTP/3 over QUIC](https://daily.dev/posts/ep200-http-2-over-tcp-vs-http-3-over-quic-o83jfjirb) · ByteByteGo · 10 upvotes · 0 comments - [Agents Week: network performance update](https://daily.dev/posts/agents-week-network-performance-update-q23bdp8rb) · Cloudflare · 1 upvotes · 0 comments - [How We Moved Discord Voice to the Edge](https://daily.dev/posts/how-we-moved-discord-voice-to-the-edge-xeeigwjrp) · Lobsters · 1 upvotes · 0 comments --- Tags: [#cloudflare](https://daily.dev/tags/cloudflare) [View this post on daily.dev](https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"HTTP/3 is faster with some work and help from Cloudflare","url":"https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd"},"datePublished":"2026-10-31T20:05:28.130Z","dateModified":"2026-10-31T20:06:00.313Z","description":"Delivery Hero cut its apps' end-to-end home screen response time by 17 percent after moving most users to HTTP/3. Chidera said handshake latency dropped 51...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3f37760eb2fe96f77d33c751e84f2058?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3f37760eb2fe96f77d33c751e84f2058?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"DEVCLASS","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"DEVCLASS","logo":"https://media.daily.dev/image/upload/s--quq5_zKP--/f_auto/v1732552064/logos/devclass","url":"https://daily.dev/sources/devclass"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"cloudflare","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"DEVCLASS","item":"https://daily.dev/sources/devclass"},{"@type":"ListItem","position":3,"name":"HTTP/3 is faster with some work and help from Cloudflare"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/http-3-is-faster-with-some-work-and-help-from-cloudflare-yvmotvwwd#faq","mainEntity":[{"@type":"Question","name":"How much did switching to HTTP/3 reduce handshake latency for a global mobile app?","acceptedAnswer":{"@type":"Answer","text":"Delivery Hero cut handshake latency between its apps and servers by 51 percent after moving to HTTP/3 via Cloudflare's CDN. In Pakistan, where networks previously took nearly a second at p90 to complete a secure handshake, response time was cut in half. Overall P95 latency dropped from 1.2 seconds to 120 milliseconds after additional engineering work. Weighing an HTTP/3 migration? daily.dev surfaces real production benchmarks like this one."}},{"@type":"Question","name":"Why does Google's Cronet library slow down HTTP/3 on Android instead of speeding it up?","acceptedAnswer":{"@type":"Answer","text":"Cronet, Google's networking library used to add HTTP/3 support on Android, took approximately 800 milliseconds to initialize at P90, which ate into the 0-RTT speed benefits HTTP/3 is supposed to provide. Delivery Hero's engineers worked around this by forking the integration into a custom bridge called OKCronet and overlapping Cronet's initialization with other background tasks rather than waiting on it. Debugging slow Android network startup? daily.dev tracks real fixes for Cronet-style initialization lag."}},{"@type":"Question","name":"Why does Cloudflare's Cubic congestion control algorithm perform worse than BBR on lossy networks?","acceptedAnswer":{"@type":"Answer","text":"Cubic, the congestion control algorithm Cloudflare uses for HTTP/3, historically handles packet loss poorly, causing latency issues on secondary app screens in markets prone to packet loss like Pakistan. Testing with Google HTTP/3-aware servers using BBR (Bottleneck Bandwidth and Round-trip propagation time) instead avoided getting hung up on the same packet loss, and Delivery Hero is working with Cloudflare to resolve the issue. Choosing between Cubic and BBR for lossy networks? daily.dev tracks how teams resolve congestion-control tradeoffs."}}]} ```