Database II and modern CPU architechture

I take Database II (cmpt454) this semester, and feel the way to speed up database is very similar to modern CPU.

One of the most important issues for implementing database is that the read/write speed of disk is too slow comparing to memory read/write. Therefore, algorithms, such as B+ tree,hash index and clustered record, are solutions for disk’s tardiness.

Now we have 3 layers of cpu cache, and each layer is faster than layer, and, also, CPU caches are more speedy than memory . When we use a listed list, the efficiency is not good because it does not take the advantage of CPU cache. On the other hand, if we use an array to hold data, it should be better than listed list because array is stored in sequential block, and CPU could pre-fetch elements of array to cache.

In my point of view, algorithms for implementing database will be more popular in general software development.