Learn Python by Building Data Science Applications

Two types of problems

Now, let's get back to the task at hand. There are two general types of problems concurrency can solve—CPU-bound and I/O-bound tasks. As you can guess from their names, CPU-bound tasks require more computing power than one CPU can provide. For obvious reasons, this kind of problem can't be solved by threading or asynchronous execution, and so multiprocessing and cluster computing are our only options.

The second type, I/O-bound, is limited by the input/output (for example, it has to wait for the database or network). Network resources are usually way slower than the CPU, so in this case, our computer just waits for data to come. This is where threading and asynchronous execution shines.