Academic research and the year ahead

[juliaaustin]

It's 2011 and it seems the creative juices are flowing world-wide as we received over fifty submissions to our request for research funding proposals.  VMware's Sponsored Academic Research program was announced last year to foster in-depth exploration in enterprise technologies.  Recipients of this bi-annual award will receive up to $250,000 in funding.  An important part of this program is VMware's commitment to collaboration.  The funded research projects not only offer financial support, but also the investment of time from our senior-most engineers.  It is our belief that collaboration between industry and academia not only enables a real-world approach to research, but it also establishes long-term relationships that extend the research beyond the initial ideas that were funded.

We started our Sponsored Academic Research program early last year and researchers at the Parallel Data Lab (PDL) at Carnegie Mellon University were the first benefactors of this program.  The PDL team is establishing an academic cloud at CMU and are researching the automation, measurement, and evolution of cloud technologies.  One of VMware's partners, HP, has generously stepped up to support this initiative by providing critical infrastructure for this research.  Using HP CloudStart, the university will replace multiple dedicated clusters with a single cloud environment for performing simulations and data analyses, as well as supporting data storage and data-intensive applications.

Our most recent request for proposals (RFP) focused on Performance Management Challenges in Virtualized Environments.  Virtualized environments are proliferating in the IT industry. They are becoming a foundation of many computing and communication environments from large enterprises and multi-tenant clouds to virtualized desktops, as well as mobile endpoints. The management of performance in such virtualized environments provides many interesting challenges.  These challenges include:

• Development of large-scale statistics gathering and analysis in scalable virtualized environments such as health models for multi-tier applications as well as correlations application, VM and host performance.
• Improvements in coordinated resource management across applications, VMs and hosts. For example, transparent solutions for the double-swapping problem and elimination of redundant disk I/O across VM and host, as well as better management of runtime systems such as JVMs and databases in overcommitted situations in VM and host.
• Performance improvements of an emerging class of distributed, latency-sensitive programming and middleware frameworks, such as Hadoop, Memcached, GemFire Data Fabric and traditional High Performance Computing (HPC).
• Performant and scalable handling of all virtualized environment management data, including data consistency, data distribution and levels of coupling between management and managed elements. For example, design of a scalable management and monitoring infrastructure for millions of VMs.

I am quite impressed with the response to this RFP and the breadth and depth of the proposals from top universities all over the globe.  Our Program Committee is hunkering down to read and review all proposals and in the coming week we will be narrowing down the set to a list of semi-finalists who will move on to the second round of the selection process.  The finalist(s) will be announced in April and I will blog about it, of course.

Have ideas for our next RFP that will go out this spring?  Comment on this post or let's start a discussion!