RAMCloud Papers

The general gist of RAMCloud makes sense, but the value is hidden by the academic perspective.

The notion of fitting tons of data into memory is happening. As one of the founders of MemSQL, a YC company that ships a distributed in-memory database, it's been my focus the last 5 years.

In addition the growth of IMDBs, Apache Spark is another cue that main-memory processing is the next phase of Big Data.

In-memory processing is enabling the creation of real-time data pipelines next to existing batch systems like Hadoop and other data warehouses, and is critical for real-time recommendations, personalization, and decision-making for businesses.

Call it a side effect of the sharing or 'on-demand' economy, but companies are now moving faster to meet higher expectations from their customers/userbases.

SAP HANA is already based in a diskless RAM database only