Terracotta BigMemory

Terracotta BigMemory is the solution which implements information storage in inexpensive RAM (RAM) allows to accelerate access to terabytes of data in real time to 1000 times in comparison with any other technology that in turn helps to disclose advantages of Big Data fully.

In the fall of 2012 Terracotta announced that its flagship product of BigMemory became the winner in the Technologies of Big Data nomination of the competition Top Innovator held within the DataWeek 2012 conference. The experts of DataWeek estimating achievements in different areas of the world of IT gave a victory to the Terracotta BigMemory platform in acknowledgement of the technologies of data management put the innovative in it. As the winner of Terracotta provided the achievements at the DataWeek 2012 conference, the action, largest in the USA, devoted to data management questions.

2014: Testing of the Terracotta BigMemory platform based on the new Intel Xeon processor E7 v2

On March 12, 2014 the Software AG company announced project deliverables of testing of the Terracotta BigMemory platform based on the new Intel Xeon processor E7 v2. The research showed: the processor in eight-socket configurations is capable to service to 12 TB OZU! In combination with opportunities of BigMemory technologies it allows to accelerate work of critical business applications and to process really huge arrays of Big Fast Data of in-memory in real time. In case of need (for example, at especially large deployments) the BigMemory platform allows to integrate OZU of several servers based on Intel processors in one cluster. However now, when one server based on Xeon E7 v2 can service considerably the bigger volume of OZU, clients had an opportunity to considerably reduce costs on processing of Big Data. So, tests from 6 TB OZU on one four-socket server showed that despite growth of volumes of the processed data the delay makes less than one millisecond. Considering horizontal and vertical scalability of the BigMemory platform, the combination of technologies of Intel and Software AG will help more than to quarter the number of the servers necessary for a cluster on 100 TB.

"In propagation conditions of mobile devices and Internet of Things effective processing of Big Data becomes not luxury, but a question of survival of business. Rates in this game are high, and requirements — too. That day when we are faced the need to provide individual information services to each inhabitant of our planet is near" — Dr. Dr. John Bates, the director of technologies of Software AG in the field of intellectual business processes and Big Data says. — Our cooperation with Intel will help our users to cut down radically expenses on the analysis of Big Data".

By 2020, according to forecasts, about five percent of profit of the companies entering into Global 1000 will fall on a segment of portable and wearable devices. The solution of Software AG it is better than any words prove how in-memory technologies and event handlings are capable to support millions of mobile users worldwide. In our demonstration we will show you as, for example, Google Glass glasses can be used by clients for connection with business processes of the seller and his partners.

Data on testing

During testing the Terracotta BigMemory platform based on the Intel Xeon processors E7 v2 proved the capability to be scaled up to 6 TB data of in-memory.

• The received results showed that at increase in volume of the used OZU the level of a delay and capacity remains stable and predictable.
• Level of a delay is 4.2-4.3 milliseconds; the indicator of the transactions processed in a second remains stable.
• The BigMemory platform is capable to be scaled to the cluster including hundreds of servers; the used business application can be scaled for support of hundreds terabyte of data of in-memory.
• Availability at the level of 99.999% provides compliance to requirements of business and critical environments.
• Economy thanks to reduction of number of servers and, as a result, decrease in power consumption by a power supply and cooling of a data processing center.