We have switched to in-memory database technology.

Real time analysis of big data causes major issues on classic database engines, even more so when they reach their limits!

As our data volumes steadily grew, we found ourselves in a unique situation of searching for a database system that is capable of storing billions of data records whilst still allowing complex analysis in real time. This is where a state-of-the-art in-memory database system saves the day.

In-memory databases put the working data set into the system memory, thereby reducing the latency issues found between the CPU and the data in traditional database engines.

The benefits are outrageous! Highly available, scalable and blazingly fast! The decision to upgrade our systems to this super power was a no-brainer!

The new in-memory database now provides enough capacity for fast, subsequent inserts (due to global latency issues or system maintenance) while normal operations and analytics continue unaffected.

The beauty of our new database  engine is that it does not  slow down as more data is added – in fact it may even become faster as more cluster nodes are provided.

Improved processing capabilities providing more detailed stats in real time….what more could a business with fast growing data requirements wish for?

About msales

msales is the leading programmatic traffic monetization platform for affiliate networks, ad networks and large publishers. By running 3,000 direct offers against 10’s of thousands of indirect targets, the platform is constantly auto-optimizing the traffic flow. The smart algorithms take into account all available metrics, simultaneously generating the highest yield for publishers and delivering the best quality traffic to advertisers.

Combining global reach with access to almost all offers in the market, msales’ German-made intelligent adserver technology guarantees to always find the sweetspot for your traffic – evaluating more than 16mio cross combinations in less than 50ms for every single click.

msales. Convert Everything

One Response

  1. Nicholas Wiersma says:


    Nice article. Was wondering what in-memory database you went with. How hard was it to cluster and provide fault tolerance. I was recently looking at Hadoop for this kind of thing, it never even occurred to me to look at an in-memory database.


Leave a Reply