The evolution of FAST

We interviewed the co-chairs of FAST ‘15 about the upcoming conference, what they are excited about in storage research, and the role of academics in today’s marketplace. Jiri Schindler, Principle Engineer at SimpliVity and Erez Zadok, Associate Professor at Stony Brook University. There is still time to register to attend the conference Feb. 16-19, 2015 in Santa Clara, CA.

Q: How would you describe FAST to someone in the technology industry that has not attended it?

Erez Zadok: If you want to have impact, I cannot think of a better place to have as many people who are really interested in storage-related technologies be exposed to this work.

FAST is also very unusual in a way—and I think that it is a source of strength—in that it has a very healthy mix of both academics and industry folks, and so you’ll find that you are very easily bridging communities that don’t tend interact very much. Academics live in their own world, and the industrial folks live in a different world. Here, people actually publish together, collaborate together, meet together, attend each other's talks, and discuss topics. A lot of future collaborations happen in the hallways.


Q: What’s the relationship between the research presented at FAST and industry?

Jiri Schindler: The time from having a paper that would appear at a conference like FAST to something that would be a company or product has shrunk. Back in the first or second FAST, it would take a few years. Now that has shrunk thanks to the healthy environment, especially in the Bay Area. But I think there's an important role to take a step back and ask what are the common principles, so that all these budding startups don't reinvent the wheel.

Q: How has the scope of FAST changed over the years?

JS: There is evolution of FAST in my view where if you look at the attendee list from 10 years ago, you wouldn't have members of the "fringe" areas. One of the reasons that FAST was created was, in fact, that we would bring the piecewise research and publications in the database community or general OS community or architecture or community together, and then over time that sort of focus on storage technologies and file systems—hence the name—grew into drawing in application writers, the hyperscalers, people who work for those companies. Of course, each individual will take a different thing and will translate it into their needs and systems and how they are going to run their business or company, but I think to me, it’s important that FAST has means of synthesizing everything into one conference and of now reaching out and drawing on attendance from infrastructure-type people. There's obviously huge interest from young companies in the Bay Area that look at new technologies. And I think what's evolving, and what's particularly of interest to me is that there is broader participation from the consumers of storage and the application layers and the IT administrators who leverage all those technologies.

EZ: I've seen how in the past 15 years computer science as a profession, as a field, has grown considerably. In academia these days we are debating, What should a degree in computer science look like? What should we keep, what should we take out? Because it simply has expanded; there's just too much stuff that we have to cover. And so we start to see a lot of specializations in various fields, and I see the same thing happening in the world of conferences. It used to be just a handful of conferences that covered just about every topic and as they grew, you had to create separate conferences for more specialized topics. So FAST in the first few years was very specialized on storage technologies. But now you see it growing and expanding to cover lots of other topics, from networking and security to operating systems to social media. And that's sort of a natural progression and growth of the field.

In terms of attendees, I've often seen at FAST attendees from large companies and small companies. There is, I think would say, a healthy recruiting effort going on: lots of people looking for new employees to recruit, and even some startups can get formed from discussions that have happened because you just happened to see a bunch of people and discussed it with them. I've noted myself, having placed a lot of students over the years in various companies in the Bay Area, that the market has been pretty good. Even in the bad 2007–2008, there was still a lot of demand for qualified masters and PhDs with storage skills and programming skills and Linux skills and it has only gotten better and better in the last few years in terms of offers and packages that people get, so I'd say this is a good time to get into this field.

The kind of research that we do is not on storage systems; it is on the entire storage stack. And there are lots and lots of layers that affect what happens with storage all the way at the bottom. It matters very much if you're at the file system level, the application, middleware, networking. Lots of different things happen at different layers, and they all affect storage in one way or another. That's the attraction also. And that's why for me I can say that I work on just about anything that could be related to storage—it could be compilers, security, netowkring, anything that could affect it is related.

Q: Anything else you’re excited about at FAST this year?

EZ: One thing that excites me in particular in addition to seeing all the latest and greatest work that is going to be published, is to hear Kirk McKusick’s retrospective talk. Our field is growing and getting old enough that you can actually go back and say, How did we start? How were things 20 or 30 years ago? I'm looking forward to hearing from him and having the younger crowd recognize where did we start and where did we go because a lot of the new generation has been born with computers at hand, and they don't know anything better or different.

JS: Erez and I were fortunate enough to be the custodians of this year’s program through the program committee and the great work that the members of the committee have put into it. I am happy to see the continued branching and inclusion of other fields. I am looking forward to the growing branching or encompassing other fields and having attendees from those fields that would not be traditional storage as defined 10 years ago.

Q: What does FAST and USENIX mean to your work?

EZ: Of course we all travel a lot, more than we would like to, and so we have to figure out what's worth travelling for. For me, FAST is one of the first things I want to go to, because I know in this one week I'm going to get a huge amount of information that will take me months to digest and then will help me think about my research agenda and what I'm going to do next and guide my students in the future. I really love the so-called hallway track. It's just meeting colleagues, meeting new people, interesting people, talking to them. I often find that starting to talk to people who are not quite exactly in my subarea makes for the most interesting future collaborations. When you start to work with people who are a little off in other subareas, that's what makes the most interesting discussions. You find common ground and ways to do something innovative that hasn't been thought of before. Even though the conference is only two and a half days, I usually travel for the whole week because if I'm there already I might as well go and visit various companies. It's usually a whole week's worth of events for me that are extremely helpful and benefical to me and my studetns.

JS: I think the community, especially FAST, is thriving. I think USENIX is very good at redefining itself. It had very different mission purpose 15 or 20 years ago. And I think it's a tremendous organization. This modern world of fast-paced Twitter and everything is a fantastic way of facilitating communication, but I think nothing can replace the "Oh, I happened to hear something in the hallway" at an event that is relevant. The human factor. We are computer scientists here talking about technology. But USENIX has providing a platform for human interation. Nothing can replace that, not even a million retweets in an hour or whatever else is a measure of success.

John Troyer is the founder of TechReckoning, a communitiy for IT professionals. He is looking forward to attending FAST '15.