Projects

A few of our current and pending consulting and research projects.

A Campus Network for the Emerging World of Data-Intensive Science

PI: Ed Lazowska (eScience Institute); Co-PIs: Virginia Armbrust (Oceanography), Tom Quinn (Astronomy), Hamish Robertson (Physics), Robert Waterston (Genome Sciences); Senior Personnel: Tom Ackerman (Atmospheric Sciences), Tom Anderson (Computer Science & Engineering, Andrew Connolly (Astronomy), Valarie Daggett (Bioengineering), John Delaney (Oceanography, OOI), Terry Gray (UW Technology), Arvind Krishnamurthy (Computer Science & Engineering), Juan Mascarello (UW Technology).

Project Summary: A proposal to the National Science Foundation (submitted Aug 2009), to revitalize University of Washington campus network to enable the emerging world of data-intensive science – eScience.

We will re-architect our approach to meeting the needs of leading-edge scientific investigators, which will catalyze advances in specific high-impact science projects that require "extreme networking". We will do this in a way that is generalizable, over time, to the campus as a whole, laying the foundation for broad transformation.

No facility is more important to our research enterprise, impacts a broader cross-section of our key NSF investigators, or is experiencing more dramatic increases in demand, than our campus network. Advances in data acquisition technology are driving a new regime in science, where the analysis of exponentially increasing data volumes is the key to discovery. Multi-disciplinary, multi-institutional, and multi-national collaborations amplify the demand for networking. The result is an "analysis ecosystem" in which the real-time transfer of vast amounts of data is the norm rather than the exception.

This ongoing rapid transformation has revealed severe limitations in our current facility. Although we have dedicated "R&D" routers that are separate from our "general-use" network core, they have not met the "extreme networking" needs of leading-edge investigators, resulting in ad hoc, non-sustainable, and non-generalizable approaches. We propose to address this gap by upgrading the research component of the campus network facility, providing dedicated L1, L2, and L3 services for researchers. Links to our regional network provider, the Pacific NW Gigapop, also will be upgraded for research traffic, as will the network's local "access layer".

Environmental-Genomic Data Integration

(Armbrust Lab; Bill Howe)

Biochemical Visualization

(Lidstrom Lab; Bill Howe, David Beck)

Longitudinal Analysis of Oceanographic Simulations

(CMOP; Bill Howe)

SciFlex: A Cross-Scale Shared Data Management Platform for Scientists

(Suciu, Balazinska, Howe)

Managing Uncertainty in PetaScale Databases

(Suciu, Connolly; Bill Howe)

High-throughput Simulation & Annotation of Protein Domains

(Daggett Lab; David Beck)

Parallelized BLAST for sequence data in an RDBMS

(Armbrust Lab, Lidstorm Lab; David Beck, Bill Howe)

Project Jar

• Co-author a proposal to collaboratively develop new algorithms for analysing large-scale datasets from next-generation short-read sequencers.
• Audit an existing relational database design for managing data acquired from a sensor network or other field experiments.
• Establish a set of software engineering "best practices": source control, issue tracking, meeting planning, code reviews, project management
• Revise the technical computing sections of a non-collaborative proposal to increase competitiveness.
• Develop a proof of concept for using offloading your application to the cloud using Amazon's virtualization service EC2, their data storage and access service S3, Google's App Engine, or others.