With Machine Learning, Can HPC Be Self Healing?
September 26, 2017 Timothy Prickett Morgan
High performance computing, long the domain of research centers and academia, is increasingly becoming a part of mainstream IT infrastructure and being opened up to a broader range of enterprise workloads, and in recent years, that includes big data analytics and machine learning. At the forefront of this expanded use is MasterCard, a financial services giant that is looking to drive the real-time business benefits of HPC.
As MasterCard has learned, however, alongside business value come additional needs around data protection. As HPC systems are more likely to hold customer-facing and other compliance related data, such infrastructure has the potential to become a more likely cyberattack target as well as a compliance risk for example through data aggregation.
HPC systems not only need to be treated in terms of governance, compliance and risk, but potentially hold the keys to the solution. For example, HPC-based machine learning can be used for detection of anomalies or false positives, to support human decisions around fraud or to provide early warning in the case of zero-day attacks.
Not only this but time to action can be increased from a matter of days to minutes, potentially impacting processes. . . if the organization is ready to deal with real-time information, that is. In this webinar, we learn from Nick Curcuru, vice president of the big data practice at MasterCard, about what needs to be in place both technically and in terms of management models and processes so that the benefits can be fully achieved.
Joining Curcuru is Onur Celebioglu, HPC engineering director at Dell EMC, Bryan Betts of Freeform Dynamics, and Jon Collins of The Register. Tune in to find out how HPC and security can work together as a package, to deliver on even the most leading edge needs of the enterprise. You can sign up to participate in the discussion, which happens on October 18 and 10 am Eastern, at this link.