PowerPoint Presentation research computing at rit

1
Research Computing at RIT Mission, Vision
Plan Gurcharan Khanna 05/11/07
2
MISSION
3
What is Research Computing?

• Research into computing technologies (RD)
• Computing resources services for researchers

4
Why do research computing?

• Save money through economies of scale via
• Centralized FTEs,
• Collocated systems,
• Efficient utility consumption
• Coordinated security
• Shared, effective use of facility spaces
• Increase researchers productivity
• More time on research, less time on technical
  support
• Higher levels of expert support and enhanced
  research collaborations
• Access to greater computational resources
• Advance science ask bigger questions, do better
  science, increase scope of research

5
Indirect Benefits

• Enhance RITs prestige
• Provide a competitive advantage with peer
  institutions
• Increase public recognition
• Improve recruitment of faculty, staff and
  students
• Growth opportunities in staff and resources

6
What architecture should we choose?
7
Current environment at RIT

• Ad hoc, home grown environments not shared
• Some faculty need more resources (CAST) while
  others need support for what they already have
  (COS)
• Existing equipment not always stable -- in need
  of upgrades and repairs (e.g. IBM Cluster in ITS,
  Condor System in CIS)
• Faculty devoting time to systems administration
  when they should be focused on research and grant
  writing

8
Examples of unmet research needs

• Astronomy. Gravity models. David Merritt.
• Remote Sensing. Algorithm development, image
  processing. Schott, Rhody.
• CAST. New image compression algorithm
  development. Chance Glenn.
• Physics. Molecular modeling for cataract
  research. George Thurston.
• EE. Autonomous robotics, algorithm development.
  Ferat Sahin.
• CS. Combinatorial Math, Ramsey Numbers. Staczek
  Radziszowski.
• CE. Parallel Computing teaching and research. M.
  Shaban.

9
VISION
10
(No Transcript)
11

• 3 Tier Compute Systems Plan

12

• Envisioned Computational Architecture

• HTC High throughput computing
• HPC High performance computing
• SMP Large memory multiprocessor systems
• GRID computing

13

• Integrated Research Environment

• Computers
• Software
• Networks
• Data
• Security
• Support
• Accessibility
• Collaboration

14

• PLAN

15

• Research Computing Systems

• 47 node dual cpu IBM linux cluster in ITS
• 5 node dual cpu IBM linux cluster for NYSGrid
• 100 client Condor flock in CIS
• 4 core 32GB RAM 1.2 TB Disk SMP in ITS (soon)

16

• Research Software

• Matlab site licensing proposal submitted
• Mathematica site licensing
• Other

17

• Research Networks

• Building 76 upgrades to 1 Gigabit
• Building 74 upgrades to 1 Gigabit
• Building 74 IPv6 testing
• 10 Gigabit proposals for video and compute
  infrastructure on campus and off

18

• Research Data

• Storage capacity
• Performance
• Management
• Access
• Security

19

• Research Security

• Goal at or above campus standards

20

• Research Support

• Paul Mezzanini Senior Sys Admin/Engineer (RC)
• Rick Bohn Sys Admin (part time RC)
• Jim Bodie Sys Admin (CIS)
• Brett Matzke Sys Admin (CIS)
• Ed Brannin Lab Manager (CASCI ICE Lab)
• Andrew Ford CS Co-op (CASCI ICE Lab)

21

• Research Accessibility

• Interface
• Middleware

22

• Research Collaboration

• Interactive Collaboration Environments Lab
• (CASCI)
• Teaching Learning Venue
• R D Projects
• Facilitate Collaboration
• Evaluative Studies

23

• Research Activities

24

• Research Projects Under Way

• Smarter Classrooms Project immersive environments
  (Dean Gatley, CIS Rosen, IT Schull)
• Condor1000 increasing the size of the Condor pool
  at RIT to over 1000 clients (RC, various depts.)
• CyberPortal Project connecting one department
  occupying three different buildings with
  persistent live high quality video and audio.
  (CIS Baum)
• Access Grid Project establishing three enhanced
  Access Grid nodes in CASCI and Golisano College
  (CASCI Johnson)
• Virtual Theater on the Grid (CIAS Schweppe, CS
  Geigel)
• 3D Modeling of Pittsford Area using LIDAR and
  Pictometry Data (CIS Lach, RC)
• Simulation of PET images using the IBM cluster
  (CIS K. Baum)
• Campus Grid connecting individual, departmental,
  and central compute and data resources in a
  shared campus facility (IBM, RC, ITS)

25

• Research Projects Under Discussion

• TeraGrid Storage for Archival Images of Ancient
  Manuscripts (EE Mukund, CIS Easton)
• Planet Lab Site (CS Kwon, RC)
• High Definition Video Classes (Masaryk
  University, Czech Republic)
• Remote Thesis Defense (Bio, Skuse)
• Medical Imaging Collaboration (CIS Dr.
  Pearlman, Director, AIC, Dartmouth College)
• Advanced Visualization Capabilities investigate
  immersive, 3D, virtual environments for research
  use (Bischoff, Batcheldor, et al.)
• Shared Faculty Research Cluster joint ownership
  and central maintenance of a high performance
  computing cluster (CoS, CoE, CCIS, CAST)
• Live Stereo 3D Internet Broadcast live stereo 3D
  transmission of surgery over Internet2 to other
  schools (Doolittle)
• Molecular Modeling research into developing and
  implementing fully detailed modelling of water
  molecules (Thurston)
• Remote Translation Service for Deaf/HOH Students
  potential joint proposal with NTID and University
  of Arizona (Albertini)
• Parallel Java implementing parallel java in a
  standard linux cluster environment (Kaminsky)
• Teaching MPI on the IBM Cluster experiment
  managing teaching and research jobs on a single
  system (CE)
• Experimental MPI across subnets (RC, ITS, CE)
• Advanced Research Infrastructure Network Test
  deployment of 10GigE network for research for
  reduced latency, increased bandwidth, and
  protocol testing. (RC, ITS, vendor)

26
How will we measure success?

• Increased number of grant proposals
• Greater external funding
• More publications
• More citations
• More use

27
http//www.rit.edu/rc