Distributed Systems Group

Research project to explore distributed computing problems relevant to Big Data Analytics, Cloud Infrastructures and modern networking and storage technologies.

Project Details

Distributed Systems Group

Distributed Systems Group

Research project to explore distributed computing problems relevant to Big Data Analytics, Cloud Infrastructures and modern networking and storage technologies.

Project Overview

Cloud Computing has emerged as the greatest paradigm shift in the computing industry in recent memory. Distributed systems and infrastructures are at the heart of these modern Cloud Computing infrastructures; be it foundational layers such as file systems and object stores, or Big Data analytics and Machine Learning platforms. Recent technological trends toward high performance RDMA network fabrics and storage technologies such as Intel/Micron's 3D XPoint are creating interesting new opportunities and challenges in modern Cloud infrastructures, as well as traditional enterprise systems such as databases. At the same time, processor vendors are packing more cores per CPU than ever before, enabling greater degree of scaling on a single node.

Oracle Labs' Distributed Systems Group was formed with the charter to explore the implications of all these technological trends in the new Cloud era. This includes applying new techniques to solve distributed computing problems, studying the role of new technologies in data center scale distributed infrastructures, and also understanding the utility of single node performance and scalability in the larger distributed context. We study scalability and performance problems relevant to Cloud Computing infrastructures, Machine Learning platforms, databases, and other foundational systems such as the Linux kernel, that are of core interest to Oracle.

Principal Investigator

Virendra Marathe

Consulting Member Technical Staff

I head the Distributed Systems Group at Oracle Labs, where we study distributed computing problems relevant to Oracle's Cloud infrastructure and Machine Learning platforms. I am interested in all aspects of distributed systems including consistency, fault tolerance, load balancing, scheduling, programming models, use of advanced networking and storage hardware, and modern distributed Machine Learning platforms.  These days I am studying privacy preserving Machine Learning.

More broadly, I am interested in system software research. In the past I led the Penumbra project that researched implications of emerging byte addressable persistent memory technologies on enterprise systems. Before that, I was a member of the Scalable Synchronization Research Group, where we studied various aspects of concurrent programming, including concurrent algorithms, programming models, synchronization primitives, run time systems, languages, compiler and architectural support.

Education:

* Ph.D. Computer Science 2008, University of Rochester