Hardware and Software Support for Managed Languages in Data Centers

Tuesday, April 11, 2017
11:00 AM to 12:00 PM
POB 2.302
Free and open to the public

Data center applications and frameworks are often written in managed languages such as Java, Python or Scala. While managed-language issues like garbage collection and JIT compilation have seen significant research on single-node systems, modern data center workloads run across a large number of independent language VMs and face new challenges that were not previously addressed. At the same time, the data center environment also creates new opportunities, as it enables rethinking layers of the infrastructure stack that have traditionally been fixed.

In this talk, I will discuss some of these opportunities and challenges, spanning both the hardware and the software layer. On the software side, I will show how to better support distributed managed-language applications using a "Holistic" Language Runtime System, which treats the runtimes underlying a distributed application as a distributed system itself. On the hardware side, I will describe how custom data center SoCs could provide an opportunity to revisit the old idea of hardware support for garbage collection. Finally, I will discuss how this type of hardware-software research can benefit from the emergence of open-source hardware and how infrastructure based on the free and open RISC-V ISA may provide an opportunity to achieve a higher degree of realism and industry adoption for academic computer architecture research.

x x


Martin Maas

Martin Maas

UC Berkeley

Martin Maas is a Ph.D. candidate in the Computer Science department at UC Berkeley, working with Krste Asanović and John Kubiatowicz. His research spans managed language runtime systems, computer architecture, operating systems and security, with a specific focus on warehouse-scale computers. He is currently working on hardware and software support for managed languages, and he has previously worked on co-scheduling of parallel runtime systems and architectural support for memory-trace obliviousness. Before coming to Berkeley, Martin received his BA in Computer Science from the University of Cambridge.