Infrastructure: Apache Spark on AWS EMR
As Securonix collects, analyzes and responds to massive volumes of data in real time, most of their operating expenses are attributed to cloud spend. At the point of first contact with Intel Tiber App-Level Optimization, 50% of their overall infrastructure cost was allocated to Spark Streaming on AWS EMR.
Securonix had already started their cloud optimization journey by prioritizing observability with dashboarding and metrics, which confirmed that data processing costs represent the largest portion of cloud spend. As a result, Securonix turned their focus towards optimizing the underlying EMR infrastructure used for Spark Streaming.
Securonix had several requirements for optimization solutions given the dynamic nature of their cloud environment. Any added solution had to be able to optimize the entire EMR fleet, require minimal engineering efforts for implementation, and assure over 15% in cost savings. They also needed the solution to work in parallel with other optimization and cost saving initiatives already in place.
Additionally, the security and resiliency of any new solution remained highly critical given the company's responsibility to maintain sensitive customer data.
Initially, Securonix chose five EMR clusters on which to deploy Intel Tiber App-Level Optimization for optimization. During the benchmark period, Securonix processed higher amounts of traffic while using fewer cores and met the 15% infrastructure reduction goal with up to 33.57% fewer cores in one environment.
Securonix was able to achieve these results by leveraging App-Level Optimization’s ability to automatically and dynamically allocate Spark executor resources based on job patterns and predictive idle heuristics. The autonomous solution also applied continuous YARN optimizations by more efficiently allocating resources based on CPU and memory utilization. Following these results, the cost saving potential was clearly identified and Securonix quickly decided to expand Intel Tiber App-Level Optimization fleet-wide.
In a matter of weeks, Securonix successfully optimized all 45 EMR clusters with Intel Tiber App-Level Optimization, enabling them to achieve their cost savings goal through a capacity reduction in excess of 15%.
*Intel does not control or audit third-party data. You should consult other sources to evaluate accuracy
Cybersecurity
HQ: Texas, USA
Employees: 900+
Securonix delivers a next generation security analytics and operations management platform for the modern era of big data and advanced cyber threats. The platform unlocks security analytics at scale, achieving unmatched detection in a unified, modern SIEM that lets customers deploy everywhere.
