Systems research: Is it relevant?
Three-day Haifa SYSTOR conference explores a myriad of experimental systems and storage
Virtualization is revolutionizing the systems industry, as one speaker explained, but participants at the SYSTOR 2009 conference enjoyed being at the conference, opening themselves up to new trends, people, and ideas.
Haifa Research Lab and one of the conference organizers. “By choosing a program that integrated both worlds, we can encourage joint projects and collaboration.”“The goal behind SYSTOR 2009 was to organize a first-class international conference that combined the presentation of qualified academic papers and the exciting work going on in the industry,” explained Miriam Allalouf, staff member at the IBM
The program covered a wide range of topics, from experimental research in systems and storage to existing advances in virtualization, power management, multicore systems, and optimization. A careful selection of the 24 top submissions held the audience’s attention for three days with a definitive collection of practical expertise.
Most conferences that include presentations from industry tend to be marketing oriented, but the speakers here went into detail on implementation and challenges,” reported one enthusiastic participant. “We even found out what’s being done in Intel in China and research work being done in Romania.”
Friendly, fertile and thought-provoking
Prof. Marc Snir from the University of Illinois, the world-renowned expert on multicore systems, who gave a fascinating talk on Supercomputing: from Petascale to Exascale. “We’re going to be seeing systems with billions of parallel threads,” noted Snir. “It’s interesting to think about how we will manage such huge systems.”One of the conference's highlights included keynote speaker
Another big drawing card was the keynote ‘Towards Invisible Storage’ given by Alain Azagury, director of the XIV business unit. Other popular sessions included Next Generation Data Center Architecture by Michael Kagan, CTO of Mellanox, and talks on deduplication by experts from Diligent (IBM), Marvel, and several universities.
“I enjoyed being brought up-to-date on such hot topics,” noted David, a participant from the Technion. “Hearing about new ideas sparked fertile conversations and did wonders for my networking.”
Does it really matter if the system works or not?
The third day of the conference held an especially thought provoking panel on the essence and relevance of systems research, with experts from IBM Research, Hebrew University, Google, the University of California at Santa Cruz, and the University of Illinois. The moderator Prof. Dror Feitelson from the Hebrew University, panelists and the audience examined several serious questions: For research to be valuable, does it really matter if the system works or not? What’s the effect of research on systems that already exist? What do we consider a success and what is failure? Can we really predict at the outset what technologies will be successful?
SYSTOR 2009 focused on experimental systems issues along with more established systems and storage technologies. Perhaps that's what research is all about: exploring something new that may evolve into a future trend or even killer application.