You're tasked with writing a spec for a generic local cache with the following property: If the cache is asked for a key that it doesn't contain, it should fetch the data using an externally provided function that reads the data from another source (database or similar). What features do you think such a cache should offer? How, in general lines, would you implement? You are tasked with improving the efficiency of a cache heavy system, which has the following properties/architecture: The system has 2 components, a single instance backend and several frontend instances. 1. The backend generates data and writes it to a relational database that is replicated to multiple remote data centers. 2. The frontends handle client requests (common web traffic based) by reading data from the database and serving it. Data is stored in a local cache for an hour before it expires and has to be retrieved again. The cache’s eviction policy is LRU based. There are two issues with the implementation above: 1. It turns out that many of the database accesses are redundant because the underlying data didn't actually change. 2. On the other hand, a change isn't reflected until the cache TTL elapses, causing staleness issues. Offer a solution that fixes both of these problems. How would the solution change if the data was stored in Cassandra and not a classic database?