Re: Using HTTP cache control for invalidating a cache of parsed objects

Brian Smith wrote:
> I don't understand your example very well. Why wouldn't the normal
> HTTP cache work as is? What is happening in the higher-level cache?

I see now based on the word "parsed" in the subject line. If it were me, I would try this strategy first:

Ask the HTTP cache for the unparsed resource If the HTTP cache didn't return something new, then keep your parsed objects as is, otherwise rebuild them. Rely on the server to properly specify cache-control directives to avoid revalidations, so that the cache only has to fetch the resources when they expire and/or are evicted.

This assumes that the cache implements an effective (LRU) eviction scheme.

By the way, disk space has become a scarce resource again because of mobile. Make sure you have a good strategy for managing whatever disk space you take up for your higher-level cache.

Cheers,
Brian