Advanced Networks Day

1
High Performance Computing and Networking _at_
UNB Virendra C. BhavsarDirector, Advanced
Computational Research Lab.Dean, Faculty of
Computer Sciencehttp//acrl.cs.unb.caMay 5,
2005
2
C3.ca Association
3
Atlantic Computational Excellence Network
(ACEnet)
People, Research, Excellence
Hosting sites
Member sites
4
ACRL Infrastructure
5
Major Users
UNB BiologyGary Saunders UNB ChemistryScott
BrownridgeLarry CalhounGhislain
DeslongchampsFriedrich Grein UNB Computer
ScienceEric AubanelVirendra BhavsarBrad
NickersonRuth Shaw UNB Text Processing
CentreAlan BurkDavid Gants UNB GeodesyPetr
VanícekRichard Langley UNB MathematicsKeith
DeBellAbraham Punnen
UNB Mechanical EngineeringMohammad Bagher
AyaniDavid BonhamAndrew GerberMarwan
HassanEsam Hussein UNB PhysicsEugene K
HoZong-Chao YanLi-Hong Xu UNB ForestryEvelyn
Richards UNB BiomedicalKevin Englehart DAL
PhysicsAndrew Rutenberg MTA ChemistryStacey
Wetmore MUN Computer ScienceDwight Kuo
Sick Kids Hospital, TorontoRegis
PomesChing-Hsing YuLen Zaifman StFX Computer
ScienceLaurence Yang UofCalgary Computer
SciencePeter TielemanJustin MacCallum UdeM
Environmental StudiesYves Gagnon UdeM Computer
ScienceJalal Almhana UPEI PhysicsSheldon
OppsJames Polson UofT Computer ScienceHue Sun
ChanMaria Sabaye Moghaddam
Supporting over 170 professors, grad, undergrad
students from UNB and Atlantic Canada
6
ACEnet Plans

• Create and operate world-class HPC facilities
• Behave as a single, regionally distributed
  computational power grid
• Create and operate sophisticated collaboration
  facilities to bind together our geographically
  dispersed research communities

7
ACEnet Networking

• ACEnet projects will require enormous networking
  capabilities
• Atlantic Canada is a distributed environment
• Waterways make networking solutions difficult
  (e.g. Cabot Strait)
• St.FX is not (yet) connected to the CANARIE
  backbone

8
Computational Grid

• ACEnet resources should be as homogeneous and
  easy to access as the electrical power grid
• The location of compute resources should be
  irrelevant due to the network
• Users will regularly access distributed remote
  computing resources

9
Data Grid

• User data should be ubiquitous
• Hundreds of terabytes of data
• With job and data migration, ACEnet will be
  sending petabytes (1015 bytes) of data across
  CAnet 4

10
Collaboration Grid

• ACEnet will put in place collaboration gear to
  communicate with Atlantic Canada and the world
• Lecture rooms equipped so ACEnet sites can share
  seminars and participate remotely
• ACEnet cafés at each site sharing continuous
  video feeds
• Desktop level collaboration equipment for
  personal communication
• Access Grid will continuously stream tens to
  hundreds of MBytes across the CANARIE network

11
Networking and Computer Science

• Many aspects of computer science depend on
  advanced network infrastructure
• Real-time processing
• Distributed and Grid computing
• Data Mining
• Network Protocols and Architecture
• Complex System Modeling and Analysis
• Sensor Grids

12
PaGrid - Mesh Partitioning

• Large scale experiments require access to a
  computational grid
• Work is partitioned across multiple HPC
  facilities and networks
• Network speed is crucial

Partitioned Data
Cluster 1
Cluster 2
13
Conclusion

• Advanced networks are conducive to sharing
• Our researchers can access remote resources
• Remote researchers can access our resources
• Advanced networks are needed for collaboration
  tools
• Virtual cafe or desktop style meetings
• Videoconferencing
• Advanced networks must grow to keep pace with
  compute infrastructure
• A strong CANARIE backbone is required for the
  success of ACEnet