Software AG has announced the results of joint testing of its Terracotta BigMemory platform on Intel’s new Xeon E7 v2 processor. The Intel Xeon E7 v2 processor in 8 socket configurations can address up to 12 terabytes (TB) of memory. When combined with BigMemory, enterprises can accelerate critical business applications with real-time in-memory access to ever-increasing volumes of fast Big Data. For large installations, BigMemory can cluster RAM across many Intel processor based servers but, thanks to the increase in RAM in a single Xeon E7 v2 server, Software AG and Intel can now significantly reduced the total cost of ownership of large in-memory data clusters. For example, tests that showcased 6TB of RAM on a single 4 socket server, demonstrated that as the volume of data increased, latency remained at sub-milliseconds levels. Given BigMemory’s ability to scale-out as well as scale up, Intel and Software AG can reduce the number of servers required for a cluster in the 100 Terabyte by more than 75 percent.
Software AG and Intel will be providing a demonstration of the Terracotta BigMemory platform with Intel’s new Xeon E7 v2 processor on both the Software AG and Intel stands at CeBIT 2014.
“As the deployment of wearable devices and the Internet of Things proliferates, every enterprise must be concerned with collecting and mining today’s hugely escalated data streams to identify opportunities in time and before their competitors. The scale is huge, which might lead, for example, to the demand for real-time personalized services for every person on the planet,” said John Bates, CTO for Intelligent Business Operations and Big Data at Software AG. “Today, with Intel, we are bringing that ability to all enterprises and significantly reducing the costs of large scale Big Data analytics.”
With wearable devices forecast to generate five percent of sales for the Global 1000 by 2020, Software AG is also demonstrating how in-memory and event processing technologies are supporting digitally empowered customers. The live demo shows how Google Glass can be used by customers to interact with the business processes of the retailer and its partners.
“Organizations that can take action based on timely, relevant, and accurate information as events occur can gain a serious competitive advantage,” said Patrick Buddenbaum, director of the Enterprise IT Segment at Intel Corporation. “By taking advantage of the boost in performance, memory capacity and reliability of the Intel Xeon Processor E7 v2 product family, Terracotta BigMemory solutions let applications store, retrieve and analyze extremely large datasets to enable use cases like real time fraud detection that were difficult or impossible five years ago.”
Terracotta’s BigMemory product was used with Intel’s Xeon E7 v2 processor to scale data volumes to 6TB in-memory.
- The results demonstrated that as the volume of data increased at each memory configuration point, latency and throughput remained stable and predictable.
- The latency remained constant at 4.2 to 4.3 milliseconds, while the transactions per second (TPS) remained stable as per the requirements of the workload.
- BigMemory can scale out to work in a cluster of hundreds of servers and an application can easily scale to support hundreds of Terabytes of data in-memory.
- Support for 99.999% availability by building on Terracotta’s and Intel’s experience with mission-critical deployments
- Lower TCO with reduced operation costs / fewer servers, lower power consumption and reduced cooling costs in the data center