Perhaps one of the most important and overlooked concept in computer science is Caching. Caching without a doubt is one of those computer science concepts that has helped me in organizing my life for the better.
Caching, merely is just another scheme or a technique to organize digital information for faster access when needed. But how can this be applied to our daily lives, you might wonder. The amount of relate caching has organizing not just digital information but physical information as well in human environment is quite impeccable. Who knows, maybe you have been following the underlying principle of cache all these days and you didn’t even realize that it’s computer science again that made your life easier.
Let’s say you have a book that definitely need to checkup everyday. It is quite obvious for you to keep this book just on top of your desk on in one of your desk closets as this inherently facilitates to quick retrieval. You obviously don’t want to keep this book back in your book shelf that you don’t touch if not just for this book. This is exactly what Caching aims to offer, quick and efficient retrieval.
But the problem is, what happens when you don’t have a compartment to store books in your desk table or what if your desk storage is full already and it can’t hold this book you definitely need to checkout everyday. This problem applies to Caching in computer science as well and of course we need to evict an item from our desk and replace that item with our book. But what item do I remove?
There are different types of eviction techniques when it comes to evicting items from cache store when it is full. The first and the most simplest is Random Eviction where you just toss off an item at random from the cache store. The second strategy is FIFO (First In First Out) where you evict the first time that you had put in your desk. This also means tossing out the oldest item that entered your desk, paving way for the new one. The third idea is called LRU (Last Recently Used) where you evict an item that’s gone long untouched.
Of course, the Random Eviction is probably the easiest and most compelling things to do as you don’t need to analyze anything and just pick one item at random from your desk and evict that. FIFO looks to be logical at first as it is quite understood that you also want to remove the oldest item from your desk. But LRU happens to be the most effective principle. The reason why it is so is because of “Temporal Locality”. Temporal Locality states that if you have recently accessed a piece of information, it is more likely that you will access the same item in your near future. Correlate this with LRU and you shall understand why it’s the most optimal and effective principle to follow. A straight forward example to understand Temporal Locality is possible when you intend to switch tabs in your machine. Isn’t it quite obvious that you might definitely return to the tab from which you switched to another ?