This will help determine how much we can cache the entire announce response (Which includes seeder count, leech count), since that would not have changed
Specifically, this can help determine whether we can:
Use inaccurate caching - just set an expiry on the cache (e.g. 60 secs), and accept that within 60 secs the real value might have changed
Accurate caching - always update the cache when seedcount / leechcount changes.
If the NOCHANGE ratio is high, then implementing (2) will have a very low cost, since writes won't be as necessary.
This will help determine how much we can cache the entire announce response (Which includes seeder count, leech count), since that would not have changed
Specifically, this can help determine whether we can:
If the NOCHANGE ratio is high, then implementing (2) will have a very low cost, since writes won't be as necessary.