Peter Alvaro, assistant professor of computer science and Jishen Zhao, assistant professor of computer engineering were awarded Faculty Early Career Development (CAREER) awards from the National Science Foundation (NSF). Each award provides $475,000 over a five-year period.
NSF presents the CAREER award to faculty "who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research." The awards provide funding to support the recipients' research, teaching and outreach activities.
Professor Alvaro studies distributed systems. He seeks to provide new tools, languages and models that will help make these systems more reliable, more robust and easier to understand.
Alvaro describes distributed systems as being inherently complex due to uncertainties in their executions: uncertainty about the ordering and timing of events, and uncertainty about failures such as machine crashes. Although we’ve known about these difficulties for decades, Alvaro says it’s increasingly important that we find solutions to them sooner rather than later. “With each passing day, we become more dependent on large-scale, cloud-based distributed systems for everything from commerce to transportation to communication to storage,” says Alvaro.
His NSF CAREER project, “Lineage-Driven Fault Injection” (LDFI) aims to improve the overall availability and resiliency of these increasingly critical systems. Equally important, his work targets the new generation of programmers, who will be responsible for building, maintaining and debugging the large-scale distributed systems of the future. Although he is reluctant to claim that a single approach (such as LDFI) will be sufficient to solve all of the problems inherent to distributed systems, Alvaro believes that understanding how programs are sensitive to failures on large-scale distributed systems “takes a nice chunk of the complexity off the table.”
"When asked why he thinks NSF funded his research, Alvaro said that “the idea of using explanations of what went right in one execution to understand what could go wrong in another is a novel idea that (one hopes) will give rise to a new research area.”
In the classroom, Alvaro uses a combination of in-class exercises, group work and theoretical material to help students connect theory with practice, which he says is particularly challenging when it comes to distributed systems. But he enjoys the challenge and likes to help students overcome barriers in their understanding. A love of teaching drew Alvaro to an academic career, and a love of northern California and the coast brought him to UC Santa Cruz.
Professor Zhao’s focus is on persistent memory, and on how to introduce non-volatile memory into existing large, data center storage systems. This is important because memory is temporary, but it’s fast. On the other hand disk storage is persistent, but it’s slow. So researchers like Zhao are looking for ways to transform memory technology into something both fast and permanent. The fleeting, temporary nature of memory means that when memory failures occur at the large scale, on big servers such as those utilized by financial institutions, hospitals, research centers and other myriad large enterprises, critical data can be lost forever. The first hurdle Zhao plans to tackle is the issue of scalability: current memory devices don’t scale to larger sizes, so they can’t accommodate large amounts of memory.
Zhao is excited about the direction of her research. She’s drawn not only to the memory device itself, but also to the theory and the implementation process by which new memory technology can be introduced into existing systems. She says that the idea of memory that acts like storage will change how we store and move data around – and while it is in fact revolutionary, Zhao believes we have to introduce the technology in a way that preserves our existing investment in large data centers.
Zhao also loves teaching and finds big classes with lots of students energizing, and also challenging in the sense that students have such varied backgrounds. She works hard to balance the pace of her classes and to design of course materials so that students learn what they need to learn. Special projects such as poster sessions and paper reviews provide opportunities for motivated students to go beyond the basic requirements of the class.
NSF proposal reviewers said that Zhao’s project is likely to answer some big questions about how to integrate permanent memory into data centers. Reviewers also identified Zhao’s compelling vision, clear five-year plan, and strong industry connections as strengths of the proposal. And the educational components, which include workshops that will help students find internships in Silicon Valley, were also considered to be very strong.
When asked if she has any advice for junior faculty, Zhao recommended that new faculty seek out input and feedback from others, particularly from more senior faculty members. She has found that the experienced faculty here at the Baskin School of Engineering are usually willing to help young faculty. For those considering submitting a proposal to the NSF CAREER programs, Zhao recommends talking to the program directors, who can provide guidance regarding which directorate and organization is the best fit for the project, and can provide suggestions about the research scope and how to deal with the integration of teaching and research.
Congratulations to Professors Zhao and Alvaro! We are looking forward to hearing how your research progresses!