The ideal computational system keeps all data in the fastest accessible place – typically system memory. However, memory volatility and cost have traditionally limited its size. Data that needs to be persisted shuffles back and forth to and from disk. Shuffling is a drag on performance, particularly when an application consumes large volumes of data. A disk-induced latency is of particular concern for AI and machine learning applications where real-time information delivery is the ultimate objective.
Thankfully, the same kinds of technology advancements that are driving the flash storage revolution are also quietly driving significant changes in the ways that computer system memory is implemented and used as persistent storage. It is now economically feasible to run an entire database for example in server memory—a significant technological advancement that promises to alter the way computing systems are designed and implemented that is now known as in-memory computing.
Non-volatile memory (NVRAM) is riding down the same cost for capacity curve as SSD. And technologies are now coming to market that allow discreet memory modules inside clustered servers to be networked together to form a scalable memory fabric. Intel (Optane) and Micron (QuantX) are advancing affordable 3D XPoint non-volatile memory modules for example, that are capable of supporting real time information applications.
One of SAP ’s hottest offerings right now is HANA, an in-memory data platform and its S/4HANA business suite built on HANA technology. IBM and Oracle also have products in this category and SAP currently claims over 21,000 HANA and 7,900 S/4HANA licensed customers. However, unlike the other two, SAP is all-in on in-memory. In a February 28, 2018 corporate fact sheet, SAP stated that “…our strategy is to become the most innovative cloud company powered by HANA.”
I began following in-memory computing technology as a bleeding-edge trend three years ago when the search was on for ways to run Big Data analytics applications in real time. Now, the significance of SAP’s strategy statement to in-memory computing—vendors and users alike—lies in the fact that a software company with over 378,000 customers in over 180 countries is staking its future on in-memory advancement. SAP aims to take in-memory mainstream.