--- title: We're making Tailscale faster | daily.dev description: Tailscale is rolling out a set of performance improvements: reduced memory overhead for small packets by avoiding unnecessary buffer copies (roughly 5%... canonical: https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: We're making Tailscale faster | daily.dev og:description: Tailscale is rolling out a set of performance improvements: reduced memory overhead for small packets by avoiding unnecessary buffer copies (roughly 5%... og:url: https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w og:image: https://api.daily.dev/og/posts/Xf9fFgy1w.png og:image:alt: We're making Tailscale faster 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. # We're making Tailscale faster **[Tailscale](https://daily.dev/sources/tailscale)** · 9 min read · 1 upvotes · 0 comments ## Summary Tailscale is rolling out a set of performance improvements: reduced memory overhead for small packets by avoiding unnecessary buffer copies (roughly 5% speed-up), a new multi-queue architecture for subnet routers, app connectors, and exit nodes to spread work across CPU cores, throughput gains from using Linux's writev to reduce packet copying, and netmap caching that lets devices reconnect to peers using a locally cached network map when the control plane is slow or unreachable, cutting startup latency by one to two orders of magnitude in poor network conditions. Memory reduction and netmap caching (via feature flag) are expected in the v1.104 client, with multi-queue and additional throughput gains planned for later releases. The team is also seeking feedback on a planned Tailscale-aware performance monitoring and testing toolkit. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What version of Tailscale will include the memory reduction and netmap caching improvements? The v1.104 client release is expected to include memory reduction via buffer changes on Linux and Android, plus netmap caching enabled by default following further testing. Multi-queue support for subnet routers and app connectors, along with additional throughput gains, is planned for a release after v1.104. Mobile clients get netmap caching in a later release too. _Track version-specific rollout details like this on daily.dev before upgrading Tailscale in production._ ### How does Tailscale netmap caching speed up device startup on bad networks? Netmap caching lets each device store a copy of its network map on disk, so it can establish direct connections with other devices using the cached data even before reaching Tailscale's control plane. This helps most on poor connections like airplane Wi-Fi or filtered hotel networks, where warm-cache starts have been observed sending data one to two orders of magnitude faster than cold starts. It requires the device to have connected at least once before and have persistent disk space. _Developers debugging flaky VPN startup latency can follow feature rollouts like this via daily.dev._ ### Why does Tailscale copy small network packets into 64 KiB buffers? Tailscale's wireguard-go implementation historically only offered one 64 KiB buffer size to unpack incoming data into, so even a 1 KiB packet was copied into its own 64 KiB buffer every time, wasting memory and CPU cycles. On Linux and Android, Tailscale now leaves packets where they land and tracks their boundaries within the larger read instead, producing roughly a 5% speed-up in many configurations. _Engineers optimizing network throughput can follow low-level fixes like this on daily.dev._ ## Similar posts on daily.dev - [Hypergrowth isn't always easy](https://daily.dev/posts/hypergrowth-isn-t-always-easy-f6jvryfcy) · Tailscale · 2 upvotes · 1 comments - [Tailscale Services GA: App-aware connectivity with more control](https://daily.dev/posts/tailscale-services-ga-app-aware-connectivity-with-more-control-j2xxhzw1i) · Tailscale · 1 upvotes · 0 comments - [Tailscale Monthly Update: December 2025](https://daily.dev/posts/tailscale-monthly-update-december-2025-glcyo6tdw) · Tailscale · 0 upvotes · 0 comments - [Everyone loves Tailscale, but it's masking a critical gap in how most people think about networks](https://daily.dev/posts/everyone-loves-tailscale-but-it-s-masking-a-critical-gap-in-how-most-people-think-about-networks-v4sdxgjmz) · XDA Developers · 4 upvotes · 0 comments --- Tags: [#linux](https://daily.dev/tags/linux), [#vpn](https://daily.dev/tags/vpn), [#tailscale](https://daily.dev/tags/tailscale) [View this post on daily.dev](https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w) ```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":"We're making Tailscale faster","url":"https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w"},"datePublished":"2026-09-22T14:08:21.096Z","dateModified":"2026-09-22T14:32:28.208Z","description":"Tailscale is rolling out a set of performance improvements: reduced memory overhead for small packets by avoiding unnecessary buffer copies (roughly 5%...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3848a8ed2c67d6fe650a28f2f505e3fa?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3848a8ed2c67d6fe650a28f2f505e3fa?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Tailscale","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":"Tailscale","logo":"https://media.daily.dev/image/upload/s--QeDb_zoC--/f_auto/v1711188591/logos/tailscale","url":"https://daily.dev/sources/tailscale"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"linux,vpn,tailscale","timeRequired":"PT9M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Tailscale","item":"https://daily.dev/sources/tailscale"},{"@type":"ListItem","position":3,"name":"We're making Tailscale faster"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/we-re-making-tailscale-faster-xf9ffgy1w#faq","mainEntity":[{"@type":"Question","name":"What version of Tailscale will include the memory reduction and netmap caching improvements?","acceptedAnswer":{"@type":"Answer","text":"The v1.104 client release is expected to include memory reduction via buffer changes on Linux and Android, plus netmap caching enabled by default following further testing. Multi-queue support for subnet routers and app connectors, along with additional throughput gains, is planned for a release after v1.104. Mobile clients get netmap caching in a later release too. Track version-specific rollout details like this on daily.dev before upgrading Tailscale in production."}},{"@type":"Question","name":"How does Tailscale netmap caching speed up device startup on bad networks?","acceptedAnswer":{"@type":"Answer","text":"Netmap caching lets each device store a copy of its network map on disk, so it can establish direct connections with other devices using the cached data even before reaching Tailscale's control plane. This helps most on poor connections like airplane Wi-Fi or filtered hotel networks, where warm-cache starts have been observed sending data one to two orders of magnitude faster than cold starts. It requires the device to have connected at least once before and have persistent disk space. Developers debugging flaky VPN startup latency can follow feature rollouts like this via daily.dev."}},{"@type":"Question","name":"Why does Tailscale copy small network packets into 64 KiB buffers?","acceptedAnswer":{"@type":"Answer","text":"Tailscale's wireguard-go implementation historically only offered one 64 KiB buffer size to unpack incoming data into, so even a 1 KiB packet was copied into its own 64 KiB buffer every time, wasting memory and CPU cycles. On Linux and Android, Tailscale now leaves packets where they land and tracks their boundaries within the larger read instead, producing roughly a 5% speed-up in many configurations. Engineers optimizing network throughput can follow low-level fixes like this on daily.dev."}}]} ```