Andy 'Bob' Brockhurst :donor:<p><span class="h-card" translate="no"><a href="https://hachyderm.io/@coldclimate" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>coldclimate</span></a></span> </p><p><a href="https://infosec.exchange/tags/dns" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>dns</span></a> <a href="https://infosec.exchange/tags/techarchitecture" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>techarchitecture</span></a> <a href="https://infosec.exchange/tags/routing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>routing</span></a> <br>I worked a long time ago on BBC News (15+ years) and compiled redirects.<br>IIRC was a massive apache rewrite rules config (and associated acceptance tests to confirm correct source/dest and no looping).</p><p>Since then I worked on sites that had compressed 10 sites to 1, and regularly re-routed "for SEO purposes" (sigh).<br>Started out as nginx with a lua block that connected to redis. couple of layers of checks for exact match full url as key, then partial match on url prefix, the final domain only match.</p><p>final iteration was similar but using dynamodb instead of redis.</p><p>We had about 11 million redirects in redis for scale.</p>