Sverre Jarp: The Intersection of Big Data, Enterprise, and HPC

Sverre Jarp, the Chief Technology Officer in the CERN openlab, has been working in computing at CERN for over 35 years and has held both managerial and technical positions promoting advanced but cost-effective computing solutions for the Laboratory. As the General Chair of the ISC Big Data’13 Conference, he is preparing what promises to be an interesting line-up. We caught up with him for a few questions about the space, and the conference.

Data-intensive computing has been around for years in areas like data mining, data warehousing and business intelligence. What is the criterion that elevates applications to the big data realm? Or has big data just subsumed all those legacy applications and added new ones?

The new form of Big Data analytics is typically based on novel types of data, often collected from Internet sources. It has an emphasis on rapid extraction of information that can assist enterprises in improving their sales and marketing activities. It does not replace traditional Data Warehousing, but complements it.

You have a mix of vendors and users in the program. What is the main audience you are targeting?

We firmly believe that most conference delegates will want to hear from both vendors and “colleagues” who have already embraced the Big Data paradigm, hence the mix.

As indicated on our Web site, The ISC Big Data’13 Conference is targeting representatives, managers and decision makers from industry and research and their staff, responsible for Big Data R&D and deployment within their organizations. The program is in particular designed for people covering data analytics, data storage and data centre management, system architectures, systems- and software-engineering, big data software tools, and specialists who are implementing and running enterprise, scientific and engineering applications. The conference is also meant for companies facing technological challenges in big data hardware, software and algorithms. 

The theme of the conference is “ISC Big Data – where Enterprise and HPC Meet.” While vendors may have a convergent product strategy around big data, their enterprise and HPC customers rarely meet. What do you think is the value of bringing together end users from such divergent application areas?

As already indicated we believe that there is a lot of synergy between the two communities, so we are offering – maybe as the first ones – a forum where experience and wisdom can be shared. We will have several talks that lend credibility to our approach; In addition to the keynotes we have, for instance, Michael Feindt who will talk about NeuroBayes which is an analytics tool that originated in science but is now used for business analytics. We will also hear from Paypal how they used an HPC-inspired Big Data system for fraud detection.

Felix Wortmann’s keynote for the conference is “Big Data – Hype or Disruptive Innovation?” But given that you and the ISC organizers went to the trouble of putting together this event, the assumption is that big data is indeed disruptive. Can you make a case for that?

We do not know exactly what Prof. Wortmann will tell the audience, but the guess is that he will stress the positive aspects of the new “Big Data” era. Companies now have much easier access to very large hardware configurations – easily reaching Petabytes in size – and such amounts of data, typically acquired from new sources, can be exploited relatively easily by people who want “new” knowledge that empowers them in their work, far beyond what the traditional approaches allowed.

Big data conferences are not exactly unique to the IT event circuit these days. What’s different about ISC Big Data’13?

Our new conference is building on the strength on the International Supercomputer Conference with over 30 years of experience. We believe that the experience gathered over many years in the HPC/HTC (high performance/high throughput) community can be very valuable to people embarking on a Big Data strategy based on data often acquired via Internet or via new types of sensors.

Related items:

Manufacturing Real-Time Analytics on the Shop Floor 

The Three T’s of Hadoop: An Enterprise Big Data Pattern 

Bare Metal or the Cloud, That is the Question…