Balancing Service and Research: Q&A with Vijay Samalam
Vijay Samalam, SDSC's Executive Director, manages the Center's day-to-day activities.
You are SDSC's Executive Director. Tell us about what you do.
Samalam: My main job is to make sure that the day-to-day running of the Center is smooth and works towards realizing the overall goals and vision of SDSC - both the short term tactical goals and the long term strategic vision. This is particularly challenging for a center like SDSC because our charter combines being both a service organization for domain scientists nationally, providing cyberinfrastructure for their scientific research, and at the same time being a leader in research and implementation of the tools that make up cyberinfrastructure.
Balancing these two sometimes contradictory roles is challenging in terms of how resources and budgets are allocated, personnel are hired and assigned to various projects, and goals are set for various groups in the Center. My daily job consists of balancing these sometimes disparate goals among all the different groups in the Center. And to the extent I am successful in realizing this I owe a lot to the senior management here with whom I try to work very closely.
What are your goals for SDSC?
Samalam: SDSC is a leading national center in data cyberinfrastructure. We are known for providing data intensive high performance computing services. It is especially significant that we have firmly established ourselves as a leader in the emerging area of data management and long term preservation. The memorandum of understanding signed recently between SDSC and the National Archives and Records Administration is indicative of how seriously we take all aspects of data management as well us how seriously the rest of the data community views our ideas and accomplishments.
I have three goals for SDSC in the next couple of years.
- I would like us to be firmly established as providing the best data cyberinfrastructure service nationally. This entails organizing ourselves so that we can provide technical services easily, defining service metrics and monitoring them so that we are confident of providing the right mix and level of service, and finally and most importantly, establishing a culture where service fulfillment is an honorable goal.
- I would also like us to be the leading center for developing new and innovative cyber tools and integrating these tools seamlessly into a unified data cyber service. We are well known for having produced some wonderful tools in the areas of high performance computing, networking, data management and system management software, and visualization. The Synthesis Center we have established jointly with Calit2 is an instance of seamlessly integrating some of these cyber tools. I would like us to continue to be innovative with new cyber tools but also be known for taking these tools to the next step and hardening them so that they can be deployed in a production environment. This step is not easy and we will have to work closely with our federal funding partners to make this possible.
- Finally I would like SDSC to be known as a collegial and fun place to work, where management leads by example rather than by fiat, and where researchers at the center feel that decision making is, as much as possible, collective and transparent.
What do you think are the challenges facing SDSC in the next few years?
Samalam: The biggest challenge is of course long term sustainability in the face of federal budget crunches. The senior management team has been working on various long term sustainability plans and we have also been trying to engage the federal funding agencies in a dialog on this issue.
But to me the more immediate challenge is trying to find the right balance between being a service organization and a leading technical innovator. In the past it was clear that centers like SDSC were primarily technical innovators. But in the past few years our users have made it amply clear that they expect us to be above all a seamless provider of cyber services. At the same time there is the expectation both from our users and from the funding agencies that we will continue to be technical innovators. Finding the right balance in the next few years will be key to our long term sustainability.
The other challenge is organizational and cultural. We are too big to be a department, and too small to be a large technical R&D organization. We are both an academic institution in some respects but also an industrial technical service organization in others. Finding the right management structure, and establishing the right technical culture to advance these goals is challenging. It is also critically important for our long term success.
You've had an interesting background -- tell us a little bit about your previous positions.
Samalam: After my graduate studies I spent a few years in academia but I have spent most of my career in industry. I worked for some time in an industrial R&D laboratory where I worked on applied problems in the areas of IC fabrication, neural networks, and networking protocols. I also worked for a leading manufacturer of networking equipment as a product director, chief technology officer, and as an architect for their next generation products. While the culture in industry is very different from SDSC, and while the long term strategic visions are different, I can honestly say that the day-to-day management issues are the same and almost universal in any medium-to-large technical organization.
We know you have an interest in quantum computing. Can you explain what this is for our readers and talk about a problem you think is interesting.
Samalam: My interest in quantum computing is more of a hobby. When I first came to SDSC I asked people if we should be worried about quantum computing replacing traditional high performance computing any time soon. No one seemed to be quite sure about it and given my background in condensed matter physics I thought I might spend some time understanding the technical issues and the promise of quantum computing.
Quantum computing has the potential for achieving massive parallelism on a scale that would be very difficult for classical computing systems to emulate. But we are still a ways from realizing this potential as researchers try to solve a host of practical problems. While I do not expect the National Science Foundation to withdraw their petascale computing initiative in favor of a quantum computing initiative in the next five years, I do think that the high performance computing community will be keeping a close watch on this up-and-coming technology.