[openstreetmap-website] Increase cachablility for individual OSM elements on the API (#129)

Matt Amos notifications at github.com
Fri Oct 12 01:16:36 GMT 2012

I can see some benefits to the ETag headers. Primarily because for individual versions of elements it makes sense to add this information and, if it might help downstream caches it seems worthwhile. 

The last time I ran analysis on API call performance (admittedly a while ago http://matt.dev.openstreetmap.org/analysis/20110403.html), individual element calls didn't rank high enough to appear on the leaderboard. It's possible this has changed, though, or would change if downstream caches became more widespread.

I'm less convinced by the idea of using redirects from the current version to a specific version. Since the redirect itself isn't cacheable, it doesn't seem to help downstream caches very much. It would be potentially slightly faster on the server side (only need to fetch the version for the element) and potentially lower bandwidth, but only where a downstream cache is available. Even then, it seems likely that any potential saving would be small compared to the latency of having to round-trip to re-check the redirect.

Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/rails-dev/attachments/20121011/a796eb91/attachment.html>

More information about the rails-dev mailing list