Comments on: Asynchronous HTTP Cache Validations http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/ A goal is a dream with a deadline. Mon, 15 Mar 2010 19:40:07 +0000 http://wordpress.org/?v=2.7.1 hourly 1 By: victori http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-226142 victori Tue, 15 Dec 2009 23:28:08 +0000 http://www.igvita.com/?p=194#comment-226142 Good article, just a heads up Squid does not support stale-while-revalidate and stale-if-error in the headers. It can only be defined in refresh_pattern rulesets. Second, to really get the most out of squid you need to normalize request headers and not let clients by pass the cache via cache-control request header. I have a written a patch to optimize this. http://letsgetdugg.com/2009/12/06/squid-headers-normalization-patch/ Good article, just a heads up Squid does not support stale-while-revalidate and stale-if-error in the headers. It can only be defined in refresh_pattern rulesets.

Second, to really get the most out of squid you need to normalize request headers and not let clients by pass the cache via cache-control request header. I have a written a patch to optimize this.

http://letsgetdugg.com/2009/12/06/squid-headers-normalization-patch/

]]> By: Ilya Grigorik http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-208907 Ilya Grigorik Tue, 11 Aug 2009 03:29:28 +0000 http://www.igvita.com/?p=194#comment-208907 Alex, thanks a lot for the link, very interesting IE feature! Never heard of it before, but definitely a nice one. One thing I can't figure out is whether there is an official spec for pre and post-check, or whether it's an IE only feature. Alex, thanks a lot for the link, very interesting IE feature! Never heard of it before, but definitely a nice one. One thing I can’t figure out is whether there is an official spec for pre and post-check, or whether it’s an IE only feature.

]]> By: Alex http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-208856 Alex Mon, 10 Aug 2009 13:23:21 +0000 http://www.igvita.com/?p=194#comment-208856 The IE caching extensions pre-check and post-check, don't these provide a browser based version of stale-while-revalidate? http://blogs.msdn.com/ieinternals/archive/2009/07/20/Using-post_2D00_check-and-pre_2D00_check-cache-directives.aspx The IE caching extensions pre-check and post-check, don’t these provide a browser based version of stale-while-revalidate? http://blogs.msdn.com/ieinternals/archive/2009/07/20/Using-post_2D00_check-and-pre_2D00_check-cache-directives.aspx

]]> By: Masking Latency & Failures with Squid - igvita.com http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-208507 Masking Latency & Failures with Squid - igvita.com Wed, 05 Aug 2009 16:27:30 +0000 http://www.igvita.com/?p=194#comment-208507 [...] again reminded about the great amount of effort they have put in to make it all work. In fact, I've written about Mark Nottingham's proposals for stale-while-revalidate and stale-if-error in the past (with a [...] [...] again reminded about the great amount of effort they have put in to make it all work. In fact, I’ve written about Mark Nottingham’s proposals for stale-while-revalidate and stale-if-error in the past (with a [...]

]]> By: Ilya Grigorik http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-172794 Ilya Grigorik Thu, 26 Feb 2009 18:49:58 +0000 http://www.igvita.com/?p=194#comment-172794 Yep, I've played around with that and it's a great feature. The only difference is that a grace period in Varnish guards against the 'dog-pile' effect, but does not actually serve up stale content on the first cache-miss. Minor difference, but one to keep in mind. Yep, I’ve played around with that and it’s a great feature. The only difference is that a grace period in Varnish guards against the ‘dog-pile’ effect, but does not actually serve up stale content on the first cache-miss. Minor difference, but one to keep in mind.

]]> By: Mike Perham http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-172781 Mike Perham Thu, 26 Feb 2009 17:57:24 +0000 http://www.igvita.com/?p=194#comment-172781 Varnish supports a "grace period" for stale objects: http://varnish.projects.linpro.no/wiki/Performance Varnish supports a “grace period” for stale objects:

http://varnish.projects.linpro.no/wiki/Performance

]]> By: Ilya Grigorik http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-140992 Ilya Grigorik Thu, 16 Oct 2008 03:36:50 +0000 http://www.igvita.com/?p=194#comment-140992 TJ, you're right, this pattern is not for everybody, but as Nate described, for certain types of content this can be a life savior. (ex: dynamic widgets, news blocks, etc... anything that's not directly affected by user actions) Ben, I've done a quick pass over merb-cache but did not similar pattern you're referring to. Is there a specific file I should be looking at? (I still believe this code belongs higher up in the stack). TJ, you’re right, this pattern is not for everybody, but as Nate described, for certain types of content this can be a life savior. (ex: dynamic widgets, news blocks, etc… anything that’s not directly affected by user actions)

Ben, I’ve done a quick pass over merb-cache but did not similar pattern you’re referring to. Is there a specific file I should be looking at? (I still believe this code belongs higher up in the stack).

]]> By: Ben Schwarz http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-140479 Ben Schwarz Thu, 09 Oct 2008 08:44:43 +0000 http://www.igvita.com/?p=194#comment-140479 This sounds very similar to the cache pattern that we've implemented for Merb's recent "merb-cache". Read more about <a href="http://merbunity.com/tutorials/15" rel="nofollow">how to get running with merb-cache</a> or <a href="http://github.com/wycats/merb-more/tree/master/merb-cache" rel="nofollow">watch the progress on github</a> Cache strategies can be mixed, matched and even joined together, this; with 'eager cache' and you've got yourself a very agile content distribution system. This sounds very similar to the cache pattern that we’ve implemented for Merb’s recent “merb-cache”.

Read more about how to get running with merb-cache or watch the progress on github

Cache strategies can be mixed, matched and even joined together, this; with ‘eager cache’ and you’ve got yourself a very agile content distribution system.

]]> By: links for 2008-10-08 « Bloggitation http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-140282 links for 2008-10-08 « Bloggitation Wed, 08 Oct 2008 08:35:25 +0000 http://www.igvita.com/?p=194#comment-140282 [...] Asynchronous HTTP Cache Validations (tags: ruby rails cluster memcached 247up programming) [...] [...] Asynchronous HTTP Cache Validations (tags: ruby rails cluster memcached 247up programming) [...]

]]> By: Nate http://www.igvita.com/2008/10/07/asynchronous-http-cache-validations/comment-page-1/#comment-140080 Nate Tue, 07 Oct 2008 15:03:03 +0000 http://www.igvita.com/?p=194#comment-140080 There are many situations where an app doesn't require real time info. For example, the plethora of apps out there pulling news from RSS feeds for say each individual user, or individual accounts. And too many RSS feeds take seconds to pull down from their original server. It's such a waste to just run a cron job to refresh each users RSS feed. Asynchronously invalidating a cache of that RSS content is a wonderful pattern for this situation. There are many situations where an app doesn’t require real time info. For example, the plethora of apps out there pulling news from RSS feeds for say each individual user, or individual accounts. And too many RSS feeds take seconds to pull down from their original server. It’s such a waste to just run a cron job to refresh each users RSS feed.

Asynchronously invalidating a cache of that RSS content is a wonderful pattern for this situation.

]]>