Transcript Metro-Area High Performance Networks and Emergency

Metro-Area High Performance Networks
and Emergency Preparedness
Invited Talk to
SBC Technology Resources, Inc.
Austin, TX
December 4, 2001
Larry Smarr, Department of Computer Science and Engineering,
Jacobs School of Engineering, UCSD,
Director, Cal-(IT)2
Driving a Future Internet Mass Market for
New Applications and Services
•
•
•
•
Broadband to the Home and Small Businesses
Parallel Lambdas in Long Haul and Metro
Wireless Access--Anywhere, Anytime
Vast Increase in Internet End Points
– Embedded Processors
– Sensors and Actuators
– Information Appliances
• Emergence of a Distributed Planetary Computer
– Storage of Data Everywhere
– Distributed Scalable Computing Power
Governor Davis Has Initiated Four New
Institutes for Science and Innovation
California Institute for Bioengineering,
Biotechnology,
and Quantitative Biomedical Research
UCD
UCSF
Center for
Information Technology Research
in the Interest of Society
UCM
UCB
California
NanoSystems Institute
UCSC
UCSB
UCLA
UCI
California Institute for
Telecommunications and
Information Technology
UCSD
www.ucop.edu/california-institutes
Cal-(IT)2
An Integrated Approach to the Future Internet
220 UCSD & UCI Faculty
Working in Multidisciplinary Teams
With Students, Industry, and the Community
The State’s $100 M
Creates Unique Buildings, Equipment, and Laboratories
www.calit2.net
Over Fifty Industrial Sponsors
From a Broad Range of Industries
Akamai Technologies Inc.
AMCC
Ampersand Ventures
Arch Ventures
The Boeing Company
Broadcom Corporation
CAIMIS, Inc.
Conexant Systems, Inc.
Connexion by Boeing
Cox Communications
Diamondhead Ventures
Dupont
Emulex Corporation
Network Systems
Enosys Markets
Enterprise Partners
Venture Capital
Entropia, Inc.
Ericsson Wireless
Communications, Inc.
ESRI
Extreme Networks
Global Photon Systems
Graviton
IBM
Computers
Communications
Software
Sensors
Biomedical
Startups
Venture Capital
IdeaEdge Ventures
The Irvine Company
Intersil Corporation
Irvine Sensors Corporation
JMI, Inc.
Leap Wireless International
Link, William J. (Versant
Ventures)
Litton Industries, Inc.
MedExpert International
Merck
Microsoft Corporation
Mission Ventures
NCR
Newport Corporation
Oracle
Orincon Industries
Panoram Technologies
Printronix
QUALCOMM Incorporated
Quantum
The R.W. Johnson
Pharmaceutical
Research Institute
SAIC
Samueli, Henry (Broadcom)
SciFrame, Inc.
Seagate Storage Products
SGI
Silicon Wave
Sony
STMicroelectronics, Inc.
Sun Microsystems
TeraBurst Networks
Texas Instruments
Time Domain
UCSD Healthcare
The Unwired Fund
WebEx
$140 M Match From Industry
Cal-(IT)2 “Living-in-the-Future” Laboratories
• Undergraduate Education
– UCSD Sixth College
• Graduate Education and Scientific Research
– UCSD Computer Science Department
– UCSD Scripps Institution of Oceanography
• Metro Broadband
– San Diego Telecommunications Laboratory
– Chicago Metro Net
• Medical Research and Clinical
– UCSD SOM / Veterans Admin / Scripps Clinic / UCSD Healthcare
• Transportation
– UCI Campus Parking
– UCI ZEVNET
• Environmental Monitoring
– SDSU Santa Margarita Ecological Reserve
Cal-(IT)2 is Partnering with Chicago Researchers
Creating Metro LambdaGrid Laboratories
Apps Middleware
•
NSF-funded
GMPLS
Extensions
•
Developing
GigE and
10GigE
Applications
and Services
•
Testing Optical
Switches
•
Linking Chicago
to San Diego
Metro Labs with
Long Haul Fiber
Clusters
Dynamically
Allocated
Lightpaths
Switch Fabrics
Physical
Monitoring
C
O
N
T
R
O
L
P
L
A
N
E
LambdaGrid On-Ramps Phase I (2002-2003)
Gigabit Ethernet NICs to 10 Gigabit Ethernet MAN
StarLight/Northwestern
UIC
LAC
TND1
TNV1
EVL
TNV1
TNC1
Router
Router
10GigE
Router
10GigE
Router
10x10
GigE
2x 10GigE
DWDM
•
TND1
DWDM
TND1 = Datamining Clusters at NU and UIC Lab. for Advanced Computing
– 32 Pentium Processors With Raid Disks and GigE Networking Each
•
TNV1 = Visualization Clusters (Gigapixel/sec) at NU and UIC EVL
– 27 Pentium processors (5x5 screen array plus control/tracking) and GigE
•
TNC1 = TeraGrid Computing Clusters at EVL
– 16 Itanium processors for code construction and testing for the TeraGrid
•
Match Cluster I/O to MAN Speeds
– at 1990’s Workstation Prices
Source: Tom DeFanti, EVL, UIC
LambdaGrid On-Ramps : Phase II (2004-2005)
10x10 Gigabit Ethernet and 2x80 Gbps MAN
UIC
LAC
TND2
EVL
TNV2
TNV2
10x10GigE
TNC2
router
router
O-O-O
switch
TND2
router
O-O-O
switch
10x 10GigE
2x40GigE
DWDM
•
StarLight/Northwestern
2x40GigE
DWDM
…
TND2 = Datamining Clusters at NU and UIC Lab. for Advanced Computing
– 32 Deerfield processors with 10GigE networking each, NetRam storage
•
TNV2 = Visualization Clusters at NU and UIC EVL
– 27 Deerfield processors with 10GigE networking each, 25 screens
•
TNC2 = TeraGrid Computing Clusters at EVL
– 32 Deerfield processors with 10GigE networking each
Source: Tom DeFanti, EVL, UIC
Research Topics in Metro Optical Networks
• Provide Integrated Services in the Tbit/s Range
– Lambda-Centric Communication & Computing Resource Allocation
– Middleware Services for Real-Time Distributed Programs
– Extend Internet QoS Provisioning Over a WDM-Based Network
• Develop a Common Control-Plane Optical Transport
Architecture With GMPLS:
– Transport Traffic Over Multiple User Planes With Variable Switching
Modes
– Lambda Switching
– Burst Switching
– Inverse Multiplexing (One Application Uses Multiple Lambdas)
– Extend GMPLS:
– Routing
– Resource Reservation
– Restoration
UCSD, UCI, USC, UIC, & NW
Research Topics in Metro Optical Networks
• Enhance Security Mechanisms:
– End-to-End Integrity Check of Data Streams
– Access Multiple Locations With Trusted Authentication Mechanisms
– Use Grid Middleware for Authentication, Authorization, Validation,
Encryption and Forensic Analysis of Multiple Systems and
Administrative Domains
• Distribute Storage While Optimizing Storewidth:
–
–
–
–
Distribute Massive Pools of Physical RAM (Network Memory)
Develop Visual TeraMining Techniques to Mine Petabytes of Data
Enable Ultrafast Image Rendering
Create for Optical Storage Area Networks (OSANs)
– Analysis and Modeling Tools
– OSAN Control and Data Management Protocols
– Buffering Strategies and Memory Hierarchies for WDM Optical
Networks
UCSD, UCI, USC, UIC, & NW
The NSF TeraGrid
Partnerships for Advanced Computational Infrastructure
This will Become the National Backbone to Support Multiple
Large Scale Science and Engineering Projects
Applications
Caltech
0.5 TF
0.4 TB Memory
86 TB disk
Visualization
Note: Weakly Optically Coupled
Compared to Metro Clusters
TeraGrid Backbone (40 Gbps)
Data
SDSC
4.1 TF
2 TB Memory
250 TB disk
Intel, IBM, Qwest
Argonne
1 TF
0.25 TB Memory
25 TB disk
Compute
NCSA
8 TF
4 TB Memory
240 TB disk
Creating Metro, Regional, State, National, and
Planetary Optical Networking Laboratories
AsiaPacific
SURFnet,
CERN
CANARIE
Seattle
Portland
NYC
TeraGrid
AsiaPacific
Caltech
SDSC
Potential
Extension
*ANL, UIC, NU,
UC, IIT, MREN
Source: Tom DeFanti, Maxine Brown
AMPATH
AMPATH
A Layered Software Architecture is Needed
for Defense and Civilian Applications
SPAWAR
Systems
Center
San Diego
www.ndia-sd.org/docs/NDIA_20June00.pdf
Grid Requirements for
Internet Middleware Systems
•
•
•
•
•
•
•
•
•
•
•
Identity & authentication
Authorization & policy
Resource discovery
Resource characterization
Resource allocation
(Co-)reservation, workflow
Distributed algorithms
Remote data access
High-speed data transfer
Performance guarantees
Monitoring
•
•
•
•
•
•
•
•
•
Adaptation
Intrusion detection
Resource management
Accounting & payment
Fault management
System evolution
Etc.
Etc.
…
IBM Adopts Open Source Globus!
Source: Carl Kessleman
From Telephone Conference Calls to
Access Grid International Video Meetings
Creating a Virtual Global Research Lab
Using IP Multicast
Access Grid Lead-Argonne
NSF STARTAP Lead-UIC’s Elec. Vis. Lab
Millions of Video Cameras
Are Attaching to the Net
• London Underground
– Initially 25,000 Video Cameras
– Expansion to 250,000 Possible
– British Transport Police Switch to Any Camera in 1 Sec.
– Source: Telindus
• British CCTV System
– Currently 2.5 Million CCTV Cameras Installed (NY Times)
– Average London Citizen is Seen by 300 Cameras Per Day
– Face Recognition Software Added in High Crime Areas
• Up to 6 Million Surveillance Cameras Across the USA
in 5-7 Years
– Privacy International Prediction
Integrating Wireless, Sensor and
Data-Management Technologies
Source: Graviton, a Cal-(IT)2 Partner
Prototyping Early Warning Systems and
Disaster Response Systems
• Three Tier System
– Wireless SensorNets Brings Data to Repositories
– Collaborative Crisis Management Centers
– Remote Wireless Devices Interrogate Databases
• Cal-(IT)2 Will Focus on High Performance Grids
– Analysis, Collaboration, and Crisis Management
– Broadband Wireless Sensornets
– Metro Optical Network Testbed
• Build a “Living-in-the-Future” Laboratory
– UCSD, UCI, and SDSU Campuses
– San Diego, Orange County, Cross Border
– Early Access to HW/SW from Industrial Partners
Prototyping a Data Grid for the Geosciences
IGPP, SIO, UCSD
Cal-(IT)2 Multi-Megapixel Displays
for Seismic and Geosciences Monitoring
Cal-(IT)2 / SIO / SDSC / SDSU
Linking Analysis and Visualization Facilities
by Optical Networks is the Next Step
• Multiple Uses
– Data Analysis
– Crisis Management
– Governmental and Private Sector Virtual Enterprises
• Distributed Collaboration
– Multiple Video and Large Image Streams
– Software to Enable Real Time Group Decisions
• Data and Compute
– PC Clusters
– AI Data Mining
• Driven by Data-Intensive Applications
–
–
–
–
Safety of Civil Infrastructure
Subsurface Oil and Gas Discovery and Management
Intelligent Transportation Systems
Distributed Medical Facilities
Multi-Sensor Data Fusion Control Rooms
are Needed for Managing Crisis Response
• Create Through Telecommunications and IT
–
–
–
–
–
–
Situational Awareness
Common Operational Picture
Local Data Warehouse with Remote Data Access
AI Data Mining of Distributed Databases
Spatial Data Analysis
Consequences Assessment Tool Set
Source: Panoram Technologies
Integrate Remote Reconnaissance
with Satellite Imagery and Ground SensorNets
HOTLINKs for Damage Locations
using GPS and Digital Photos on GIS
Maria Feng
UCI Civil & Environmental Engineering
GIS Based Crisis Management Software Tools-SAIC’s Consequences Assessment Tool Set (CATS)
• Prediction of Disaster Evolution
– Hurricane, Earthquake, Terror
•
•
Casualty Distribution Probabilities
Response Management
Source: SAIC
Common Portal Architecture
Customized for Crisis Management
Web Browser - Portal Interface
State Values
Portal Engine
Analysis Tools:
- Pattern Recognition
- GIS Queries
- Data Mining
- Multi-Sensor Fusion
Data Gather
XML
User Preferences
HTML
Legacy and Problem Specific
Databases, Collections, & Literature
Applications:
- Epidemiology
- Transportation Systems
- ...
Built on Prior SDSC and NCSA Work
Data Mining Across Agency Stovepipes
Is an Essential Next Step
INFORMATION
SOURCES
CRIMINAL
JUSTICE
EMERGENCY
RESPONSE
Federal
State
Local
INTEGRATED
INTELLIGENCE
SYSTEM
PUBLIC
HEALTH
Cal-(IT)2 is Partnering With SAIC
On Emergency Preparedness
SECURITY