Google's HTTP/2 codec slows Envoy

(apoxy.dev)

40 points | by dilyevsky 5 hours ago ago

9 comments

  • basilikum 3 hours ago ago

    this looks like it could be an interesting read that i would be bothered to read if a human bothered to write it.

    • dilyevsky 2 hours ago ago

      I did write it, the machine did all the react tho (with my help)

    • adolph 2 hours ago ago

      I found it worthwhile and interesting. I'm not in deep with proxies every day and have never had to profile performance like this. The writing was very comprehendable to me and the summing conclusions rounded things out nicely. The first conclusion was probably the most practical takeaway: Codec defaults are performance changes. Its obvious in retrospect but also easy to glaze over when reading release notes.

  • rightbondsadmin 3 hours ago ago

    this reminds that protocol compliance and performance aren't always aligned

    • dilyevsky 2 hours ago ago

      Im not entirely sure what the motivation was for replacing nghttp2 with oghttp2 - nghttp2 is probably the most used codec for http/2 - so probably not compliance. I think it was mostly about code reuse since oghttp2 is part of QUICHE which is also used in Envoy for QUIC.

  • nosefrog 3 hours ago ago

    So tired of these AI generated articles.

    • dilyevsky an hour ago ago

      This is a company blog. I run some paragraphs through ai to fix my dogshit sentence structure since english is not my first language (come to think of it, my russian has always been pretty lame too). Get over it

    • alt227 an hour ago ago

      So tired of comments that do nothing other than point out things are ai.

  • yawniek 3 hours ago ago

    [flagged]