algorithm - Choosing a proxy caching approach to work with Google Calendar API -

i need suitable caching approach use enterprise portal showing data google calendar api. algorithms or design patterns best applicable?

the google calendar api limited number of requests per day (defaults 10,000 requests/day - have requested more) , rate of access (5 requests/second/user).

there 2 core api methods expect use, 1 list of user calendars (1 api hit) , 1 download events of individual calendar (1 api hit per calendar).

both calendar list , individual calendars contain etag values can used avoid unnecessary api requests. if have list of individual calendar etag values can see if of these have changed querying calendar list. (unfortunately http 304 not modified response still counted api hit).

also don’t want download , cache entire calendar contents (so maybe few days or weeks @ time).

i need find approach tries minimize number of api calls doesn't try store everything. needs able cope fetching data unchanged calendars because "time sliding window" on calendar data has moved on. system backed data storage multiple portal instances share same data.