Infrastructure for the Future Cloud - Mary Fernández, AT&T Labs Research

Mary Fernández
AT&T Labs Research
http://www2.research.att.com/~mff/
Date: Thurs., Feb. 9, 2012

Title: Infrastructure for the Future Cloud

ABSTRACT: Despite its many interpretations, the "Cloud" is fundamentally a digital utility that will undoubtedly change, and hopefully improve, how we interact with and manage our digital lives. The Cloud utility promises both individual and enterprise users near instant access to seemingly unlimited, low-cost, high-performance, and secure compute and storage resources, from any device, anywhere, any time. At AT&T Labs Research, we are designing and prototyping the Cloud infrastructure necessary to deliver on these promises. I will begin with a brief description of the virtual machines (VMs) and virtual network elements that are the building blocks of Cloud infrastructure and then describe two research projects. The Distributed Management Framework (DMF) simplifies the specification, provisioning, and management of complex cloud services by providing high-level programming abstractions and transactional management semantics. A video demo illustrates how DMF migrates a game server across the continent reliably and seamlessly. Kaleidoscope is a tool that helps manage VMs at a fine granularity, making it possible to spin-up, migrate, and consolidate VMs efficiently, improving infrastructure utilization and potentially reducing energy usage. I'll conclude with some open problems of interest to our research team. Our team includes Andres Lagar-Cavilla, Simon Xu Chen, Matti Hiltunen, Kaustubh Joshi, Yun Mao, Kobus van der Merwe and student interns from University of Pennsylvania, Carnegie Mellon and Rutgers Universities.

BIO: Mary Fernández is Executive Director of Dependable Distributed Computing at AT&T Labs Research, a department of principal investigators whose expertise includes security, formal modeling, and reliability of networks and distributed systems. She has been at AT&T Labs Research, having a lot of fun, since receiving the PhD in computer science from Princeton in 1995. Mary is a long-time mentor of students in STEM fields and proudly serves on the boards of MentorNet and the Computing Research Association.