In late June I participated in EMC’s second annual University Day, which deepens the collaboration between EMC and its academic partners. (See the list of universities we partner with below.)
EMC established University Day to foster an open dialogue between EMC, academic researchers, and students. It provides a perfect venue for us to strengthen existing relationships, discuss potential research engagements, and recruit top students.
Held in Santa Clara, California, the agenda for this event included discussions on challenging high-tech issues in next-generation data centers, including new developments in solid state storage.
My part in the agenda was to present how EMC uses data analytics to manage our global university research portfolio. University research is a component of our larger innovation analytics strategy. In addition to academic research, this strategy encompasses idea contests, innovation conference engagement, paper publication, and intellectual property.
The highlight of the day was the student competition, in which nine students presented their research via posters and competed for first prize. What I found remarkable about their proposals—which explains why they were chosen!—is how well they align with current industry trends.
The top three winners and the relevance of their ideas follow:
- First place: Vasily Tarasov, Stony Brook University – Effective Block Layer Deduplication via File System and Application Hints. Customers want IT features (e.g., deduplication, backup) to be seamlessly integrated into the applications they are using. This prevents them from having to use separate tools for separate features. Having the application layer provide “hints” to a deduplication tool is a great first step in this direction.
- Second place: Stan Park, University of Rochester – Flash-Oriented I/O Scheduling in Software Defined Network. The phenomenon of Application Nearness means that applications are getting further and further away for the final resting place (e.g., a storage device) of their data. Complex, multi-hop networks have inserted themselves in between applications and their data, and software-defined networks are becoming an important arbitration point to guarantee performance. Adding a flash layer to the network makes a lot of sense.
- Third place: Wenfei Wu, University of Wisconsin-Madison – Detecting Bottleneck Middleboxes in Virtual Networks. Along similar lines, bottleneck detection within virtual networks has been an area of recent innovation at EMC (e.g., RSA Security Analytics). Wenfei’s approach provides another mechanism for getting more intelligence and optimization between application and storage.
The students, faculty, and EMC employees greatly enjoyed the dialogue that all of the student projects inspired. For further information on some of this work, feel free to browse the academic research portals of the universities EMC works with:
- Carnegie Mellon University
- Case Western Reserve University
- Florida International University
- Harvard University
- Northeastern University
- Princeton University
- Stony Brook University
- UC Irvine
- UC San Diego
- UC Santa Cruz
- University of Minnesota
- University of Rochester
- University of Utah
- University of Wisconsin
