Supersize my systems and storage
Researchers push the boundaries for clouds, distributed systems, virtual environments, and energy issues
Research company IDC last year forecast that 2010 would see the volume of digital data stored by people on the planet reach 1.2 billion terabytes. By 2020, that volume will have grown by a factor of 44 to 35 trillion gigabytes. It's no wonder researchers in IBM are looking ahead for new ways to store, access, manage, and secure all that data.
As part of these efforts, IBM Research – Haifa recently hosted SYSTOR 2011, a three-day international conference for systems and storage experts from industry and academia. The conference brought scientists from Israel together with colleagues from around the world to share their visions and leading-edge research in these areas.
"This year we're seeing more work focusing on distributed systems and cloud storage," explained Paula Ta-Shma, researcher at IBM Haifa and the general chair of the conference. "With the immense volume of data coming at us, new paradigms for storing the information and connecting these data systems have become crucial."
The systems and storage community is truly multi-faceted. Students and professors from academia networked with industry players from IBM, EMC, SAP, and others—who were all joined by employees from the many storage startups for which Israel is now famous. "Bringing everyone together opens new dialogs and presents opportunities for joint collaboration that may not otherwise happen," continued Ta-Shma.
The program included an exciting lineup of world-renowned experts, including Hank Levy from the University of Washington, who spoke about Pushing the Boundaries of Distributed Storage Systems; Daniel Keren from Haifa University, who discussed Efficient Monitoring of Large Distributed Systems; and Kai Li from Princeton University, whose talk was titled Challenges in Building a Commercial Deduplication Storage System. Other talks included discussions on nested virtualization, solid state disks, performance tradeoffs, memory issues, energy consumption, and more.
Another highlight was a talk by Haifa researchers David Carmel and Dafna Sheinwald, who helped build the Watson deep question-answering system that recently beat leading human contestants in the Jeopardy! quiz show. The conference also featured an excursion to the famous archeological remains at Caesarea, the former home of the ancient Romans and King Herod.
Participants expressed their appreciation for the well thought-out program and the high-level speakers at SYSTOR 2011. Only 30% of the research papers submitted to SYSTOR were accepted by the program committee this year, pointing to a new level of maturity for the yearly event.
Program co-chair Liuba Shrira is a visiting professor at the Technion, a professor at Brandeis and a research associate in MIT. She explained that research into systems is usually done by small groups. "This kind of conference gives researchers an opportunity to become part of a community and face the challenges together, instead of working in isolated islands, each with its own ivory tower," she noted.
"We all see the volumes of information multiplying on YouTube, Gmail, or Facebook, which have given us new roles as consumers of data," continued Shrira. "Everyone wants to know how this revolution will play out and enjoy the results. But, most of all, all this data is opening new realms in storage and systems. We've just begun to scratch the surface of what can be done with all this information—and there is just so much potential all around."