Urika-GX Use Cases
The Urika®-GX platform’s agility and speed make it uniquely effective in supporting the entire pipeline required for successful analysis.
Financial Services: Surveillance
The last thing a chief compliance officer at a financial services provider wants is a letter from the SEC, or any other financial regulator.
Governance, risk and compliance (GRC) investigations on illegal actions committed in financial firms usually appear as a series of anomalous small events associated with a perpetrator. If an auditor can identify these anomalous small events, there is high probability that the illegal action can be prevented. Unfortunately, a GRC investigation is a highly compute-intensive problem that requires a highly versatile platform, one that can continuously digest and process structured and unstructured data for all stakeholders. The system should have the ability to represent this data in a network with visibility to connections and relationships. And finally, the analytics platform should generate alerts in time to prevent losses.
The Urika-GX platform fits the bill. The two dedicated I/O nodes allow seamless, high-speed ingress of data without interruption. That data can then be processed in the Cray Graph Engine (CGE) for developing a rules-based compliance system that learns over time. The inclusion of CGE also enables the fast discovery of a series of anomalous events.
Additional financial services use cases:
- Alpha discovery data scientist toolkit
- Risk management
Life Sciences: Estimating Patient Outcomes
Advances in the field of genomics have enabled medical researchers to gather highly granular patient data. It is not unusual to see a complex collection of data elements: gene expressions data, cancer types, and current status for thousands of patients, for example.
Before the Principal Investigator (PI) runs any statistical analysis she must organize the data to filter out relevant information, “munge” different data files into one and “wrangle” the data into a proper format. Next, the PI requires a set of tools to run complex statistical analysis on that “prepared” data and some way to visualize it.
The Urika-GX platform enables the PI to run end-to-end complex and mixed workloads, including scientific simulation-type workloads, without the need to move the data to different machines for different steps. Pre-integrated Python notebooks allow the PI to wrangle the data quickly and share findings with other researchers. The Spark-MLlib machine learning library enables the use of state-of-the-art machine learning algorithms — for example, to classify genes based on cancer types. The high-performance Cray Graph Engine allows the PI to create knowledge graphs and discover new dependencies and connections in data — connections that would take her days, if not weeks, to find without the Urika-GX system.
Additional life sciences use cases:
- Drug repurposing
- De novo assembly
Cybersecurity: Network Reconnaissance
Maintaining a secure network to ensure uninterrupted business operations is an ever-growing need in today’s hyper-connected world.
Network reconnaissance is an old problem that chief security officers and IT teams have continuously grappled with. But the huge amount of machine-generated data as log files and net flow data has always tested technologies to their limits. The problem is simple in theory: detect anomalies in network data, in time to take action. But with the huge amount of machine-generated log file data, net flow data is pushing technologies to their limits.
The Urika-GX system provides the optimum platform for solving problems associated with high-velocity, high-veracity and high-volume data. For a complex network reconnaissance workflow, users can configure dedicated I/O nodes, allowing data ingestion from a variety of sources. Captured data can be processed in memory, leveraging Apache Spark™, and then fed into the Cray Graph Engine to run graph and machine algorithms — thus uncovering network issues and attacks in near-real time.
Additional cybersecurity use cases:
- Insider threat detection
- Advanced threat detection
- Real-time system log analysis
More Examples of Urika-GX Platform Use Cases
- Root cause analysis in manufacturing
- Operations optimization in manufacturing
- Digital twin in manufacturing
- Machine log mining in datacenter operations
- Crowd-sourced data integration for earth sciences
Data scientists today need to do more and discover more.
IT departments need to keep up with the pace and explosion of analytics infrastructure requirements.
Sheer horsepower using Intel® Xeon® E5-2600 v4 processors up to 48 nodes and 1,728 cores
