Classic OLAP and OLTP systems are considered to be opposites – OLTP systems process simple, ongoing data such as sales, job orders, stock fluctuations, etc., and need to deal with it in real-time; OLTP is about processing masses of data in real time, but not about doing complex processing, that is where OLAP comes in.
OLAP systems are designed to do complex processing on aggregated data and multi-dimensional information collected from different sources (including OLTP systems). However, OLAP processes are heavy and slow, using many resources, therefore they are usually run in batches, at times when transactional flow is low.
These two types of action groups contradict, since systems that need to quickly process transactional data cannot do the complex analytical processing, while systems that do complex processing cannot provide quick analysis. That is why traditionally, each of these jobs was done by a separate system.
Then again, why not have one system that does both? Many enterprises today need to have the ability to run complex analytic queries on real time transactional data. Integrated OLAP and OLTP systems can support the day-to-day operational needs as well as decision-making processes in an enterprise, which require continuous or on-demand analysis of current or historical information, and the flexibility to handle dynamic enterprise data. Perhaps it is possible to solve the technical contradictions and achieve a real time OLAP system that processes the ongoing data flow while also doing the complex processing in real time.
As it turns out, this idea is not fiction and such integrated systems do already exist in production today. This exact issue was discussed at VLDB 2011 conference held in Seattle last August. Members of the Hasso Platner Institute in Germany presented a paper on HYRISE (A Main Memory Hybrid Storage Engine), where they discussed hybrid OLAP and OLTP systems, and the integration of mixed workload systems. However, the system presented is not a production system.
There are production-ready hybrid systems currently available on the market that do provide integrated OLAP and OLTP functions. One example of an integrated system is our ActivePivot solution.
The ActivePivot solution has been used in the market for the past five years by major enterprises, including ING, HSBC, and JP Morgan Chase. It was originally built as a fast OLAP engine, using in-memory and multi-threaded technology, and a business logic architecture that continuously processes events, so when new data arrives in the system, it is handled in the same way an OLTP system handles events. ActivePivot is a classic OLAP engine that functions like an OLTP system, providing an integrated real-time transactional environment with complex analytic capabilities.