Scaling with Azure Synapse and Azure Data Lake

Many data analytics platforms are based on a symmetric multi-processing (SMP) design. This involves a single computer system with one instance of an operating system that has multiple processors that work with shared memory, input, and output devices. This is just like any desktop computer or laptop we use today and extends to many server technologies too. An alternative paradigm is massively parallel processing (MPP). This involves a grid or cluster of computers, each with a processor, operating system, and memory. Each machine is referred to as a node.

In practical terms, consider computing a sum across 100 billion rows of data. With SMP, a single computer would need to do all the work. With MPP, you could logically allocate 10 groups of 10 billion rows each to a dedicated computer, have each machine calculate the sum of its group in parallel, and then add up the sums. If we wanted the results faster, we could spread the load further...