What are the benefits of OLAP in memory for traditional systems with significant memory?

Question

Do operating systems with OLAP memory have advantages over traditional OLAP models supported by enough RAM to contain the entire cube?

For example, if I use the MOLAP mechanism (SSAS) and GB / TB RAM, where the entire cube (or even the star schema) is a resident of RAM, what is the difference compared to something like TM1 / SAP HANA?

Answer 1

It basically boils down to the following:

A system that is truly optimized for in-memory operations takes into account several aspects, such as random access, memory page size, different cache levels (CPU, ...), etc.

This allows you to maximize the opportunities offered by RAM, and the hard drive does not offer, which in turn provides excellent performance.

A traditional engine optimized for access to the file system usually takes into account several aspects related to file processing and file system OS, etc.

Even when such an engine loads everything into its cache (memory), it still works with AS IF data if it were on disk, which makes sense, since the code should work in situations where not everything fits into memory. Using the same implementation for both situations allows better testing / stabilization / correction of errors / maintainability, etc. But this leads to the fact that "does not take advantage" of everything that makes access to RAM different from access to a file / disk. Such an engine can usually be made faster if it implements RAM optimization, so that it offers the best in every world (RAM compared to a disk) ... I don’t know which engine does this ...