November 2014

Sponsored by NuoDB Inc.

Download pdf

We explain how and why the current layered DW architecture developed and examine why today’s business imperatives of speedy decision making and data driven action taking demand a new operational/informational approach. This environment, also called HTAP (hybrid transaction/analytical processing), needs a combination of in-memory operation and novel database techniques to enable simultaneous read/write and long-read activities on the same data. 

Abstract

We are experiencing a seismic shift in the way business works. As business and technology merge in the biz-tech ecosystem, real-time data has moved to the center of upfront analytics and immediate action. The old approach to delivering decision support data through a layered architecture is no longer sufficient. A new combined operational/informational world is emerging.

This paper explains how and why the current layered architecture developed and examines why today’s business imperatives of speedy decision making and data driven action taking demand a new approach. This operational/informational environment, also called HTAP (hybrid transaction/analytical processing), needs a combination of in-memory operation and novel database techniques to enable simultaneous read/write and long-read activities on the same data.

We describe NuoDB, a modern, distributed, memory-centric database, composed of processes scaling out over multiple hosts in one or more datacenters, which addresses these operational/informational needs. With reduced contention between reads and writes, and a peer-to-peer networked architecture, this database strongly supports the emergent operational/informational needs of today’s most advanced business models.