this is all just from the top of my head, so don't quote me on this, just use it as a general guideline and read more on wikipedia or something...i'm not responsible if you fail your class 🙂
cache effectiveness (not only the hardware, but software as well) can be measures by the Hit/Miss ratio. a cache Hit is when the data requested is already residing in the cache, and a Miss is when the data has be fetched from the physical storage or in case of a calculation, computed on the fly.
if you have a high Hit count and low Miss count it means your cache is holding the highly requested data and is helping the component perform better. a high miss count means your cache is not effective.
there are several ways to increase cache effectiveness:
1) tweak the algorithm - the algorithm decides which data to keep in the cache and which data to evict. even small changes in eviction policy can have great impact on the hit/miss ratio.
2) allocate more resources - if possible, enlarge the cache. this always helps BUT can be expensive or have diminishing returns. on the other hand, a too small cache which can hold only a single "page" of data is not going to be effective as you would always need to fetch from physical memory unless all your requests are the same. wouldn't it be great if we could hold ALL of our data in cache and always have a 100% hit ratio? 🙂
3) pre-load data onto the cache - this is more of an optimization technique, trying to predict which data is going to be highly requested
