Learning Functional Programming in Go

Buffered implementation

Let's try using input/output buffers.

In the following diagram, each stage of the pipeline reads from its input buffer and writes to its output buffer. For example, the Decrypt filter reads from its instream buffer, coming from the data source and writes its output buffer:  

Since there are two orders, the buffer size is two. Since concurrent queues' buffer shared inputs and outputs, if we had four orders, then all filters in the pipeline could execute at the same time. If we had four CPU cores available, then all filters could run concurrently.

As long as there is room in its output buffer, a stage of the pipeline ...