Googlebot on a tcp/ip level

Apart from the auto-backslapping this a an interesting confirmation that performance of your website impacts search engine results.<p>As for google sending multiple requests, from the way the article is written it sounds as though google sends the requests all at once and then waits for the answers to come back one by one, you can cure this by switching keep-alive off on the server side.<p>Typically in your http configuration you would add a line like this (example for apache):<p><pre><code> KeepAlive Off </code></pre> You could even do this just for the googlebot:<p><pre><code> BrowserMatch "GoogleBot" nokeepalive </code></pre> That way you can 'fix' the google bot issues without affecting the normal users of the site.

While the article was interesting the first line "We did it. We solved one of the unsolved big SEO (Search Engine Optimization) mysteries of the modern time." had me reaching for the close button. Is it just me or is there a LOT of spin from SEO types?

Someone needs to point out that this isn't the TCP/IP level, it's just HTTP.<p>I know it is pedantic technical discussions need precision in their agreed terminology. HTTP and TCP/IP are at different stack levels, and everything discussed here is HTTP.

It would be useful if Google hooked some info like this into the error message somehow. Can anyone pass it on to the right people at Google?

"If you don’t understand a single word i just wrote, please remember, we are geeks."<p>Hum. An SEO discovers HTTP pipelining and gets all excited about it. <i>yawns</i>

Thank you so much. The timing of this article is fantastic I have actually been battling this exact bug/quirk all week.

I wish we had a downvote arrow as well. This is SEO spam and it reached nr 1 on HN.