Service Oriented Architectures for R

1
Service Oriented Architectures for RE
networksGoogle Mashing everything

• Bill St. Arnaud
• CANARIE Inc www.canarie.ca
• Bill.st.arnaud_at_canarie.ca

2
Google mashing

• Google as developed a web service for Google Maps
  that allows users to overlay any geographical
  data
• A powerful example of SOA and web services
• No more using maps as GIFs or JPEGs, or using
  proprietary mapping software
• Users can also create a workflow of their
  geographical data overlaid onto Google Maps and
  offer that as a web service to others

3
Todays Network OS
The network is subservient to the computer
The application is tightly bound to the OS
Network
Application
Application
User
User
OS
OS
The network is a mechanism for applications to
communicate with each other
Data
Data
4
SOA Network
Application and data exist on the network and are
uncoupled from any specific machine or location
SOA
SOA
Network
OS
OS
The computer is subservient to the network
Application and Data
SOA
SOA
SOA
SOA
OS
OS
OS
OS
Data
Data
Data
Data
5
SOA (Web 2) versus Web 1

• Web 1
• HTML is the composition language of Web 1
• Its power is the ability to incorporate links to
  other web pages and in turn be linked to by
  others
• Frontpage (and others) allow HTML editing
• Human grammar and sentences provides the semantic
  structure of a web page between the various
  elements including hyberlinks
• Apache (and others) convert HTML script into
  working web page accessible via HTTP
• Web 2 (SOA)
• XML is the composition language
• Its power is the ability to incorporate links to
  other web pages and in turn be linked by others
• BPEL provides the semantic structure between
  various web services
• Resulting BPEL script is also a web service which
  can be linked to by others
• Apache/Axis (.Net, Wepshere) convert XML into
  working web services accessible via SOAP (mostly
  via HTTP)

6
The big picture
Integrative Science
E-Science or E-Research
Cyber-infrastructure SOA (web services,
workflow, security, etc)
Networks
Grids
HPC
Databases
Instruments
7
Science drivers for SOA for RE networks

• Big Science
• CERN, eVLBI, Ocean Observatories
• Integrative Science
• Increasing interests by researchers into
  multi-disciplinary science as opposed to
  reductionism
• Need to link sensors, instruments and databases
  from different fields to extract new knowledge
• Examples
• York University is connecting smog sensors along
  freeways and correlating with large population
  health data to predict consequence of traffic
  congestion on public living near the freeway
• Neptune undersea network to investigate algae
  blooms that precede major undersea earthquakes

8
New Integrative Science
Source Office of Integrative Activities NSF
9
More Diversity, New Devices, New Applications
Personalized Medicine
Picture ofdigital sky
Knowledge from Data
Wireless networks
Instruments
Source Larry Smarr??
10
SOA and networks
11
GENI-Network Virtualization
Source Network Virtualization web site
12
GENI SOA UCLP
APN
Instrument WS
Parent Lightpath WS
Substrate Router
Substrate Switch
GMPLS Daemon WS
Child Lightpath WS (may run over IP Ethernet,
MPLS, etc
Virtual Router WS
Wireless Sensor Network
Timeslice WS
13
GENI is a subset of UCLP

• Parent or root lightpath substrate link
• Child lightpath (SONET, MPLS, IP tunnel)
  virtual link
• Router substrate router
• Virtual or blade router virtual router
• APN virtual end to end system linking processes
  (time slices), instruments, storage, etc
• No equivalency to switch or virtual switch in
  GENI
• SOA Web service can represent time slice,
  instrument or other process
• UCLP allows user to configure their own APNs
  using BPEL
• Change topology, bandwidth etc
• APNs can be made up of layer 1 to 3 virtual links
  connecting instruments, routers or switches

14
Extending the network into the application
APN extends into computer to specific processes
zzzz41001
Instrument Web service or software process
User A
xxxx41001
xxxx41004
Virtual Router WS
xxxx41002
Single Computer or WS instance of an orchestration
Routing daemon Web service
xxxx41003
xxxx41005
DWDM Network
Web service or software process
Interface Card or port
yyyy41001
VPN Links
User B
15
Similar initiatives at Cal-IT(2) UCSD
A real-time data grid system Multi-disciplinary
data being integrated Multiple Sensor types being
adapted Real-time data virtualization
enabled Discovery access through metadata
supported

• (Laboratory for the Ocean Observatory
• Knowledge Integration Grid)
• Integrate Instruments Sensors
• (Real Time Data Sources)
• Into a LambdaGrid
• Computing Environment
• With Web Services Interfaces
• New OptIPuter Application Driver
• Gigabit Fibers on the Ocean Floor
• Goal Prototype Cyberinfrastructure for NSF
  ORION
• www.neptune.washington.edu

16
APN Resource List CreationView by CANARIE staff
Lightpath Object Creation
CANARIE ONS Network Resources
1
Toronto
Edmonton
Chicago is hidden
2
CANARIE OME Network Resources
ONS
Montreal
ONS
ONS
Vancouver
BCnet
Toronto
Ottawa
Edmonton
Chicgao
3
MAN LAN HDX
4
New York
Chicago
Seattle
Toronto
STAR LIGHT HDX
5
Amsterdam
Ottawa
Vancouver
Edmonton
Montreal
To Fermi
Victoria
New York
Geneva
New APN Resource list composition
To Brookhaven
17
CANARIE provides APN resource list to TRIUMF
1G Interface WS
URI http//canarie_apns/triumf_apn.ws
5G Interface WS
10G Lightpath WS
1G Lightpath WS
Toronto
Amsterdam
Ottawa
Vancouver
Edmonton
Montreal
To Fermi
Victoria
New York
Geneva
NOTE This resource element is actually an
aggregation of several elements on CANARIE
network. The exposed WS may actually be a BPEL
composition of the underlying WS elements
To Brookhaven
18
TRIUMF GUI harvests other APNs from UoVic, UoT,
etc
TRIUMF Tier 1
1G Interface WS
UoToronto Physics Tier 2
5G Interface WS
UBC Physics
UA Physics
UoT Physics
10G Lightpath WS
External links or APNs
UoVictoria Physics Tier 2
UdM Physics
TRIUMF APN
UoT APN
Carleton Physics
Toronto
Amsterdam
Vancouver
Edmonton
Montreal
UoV APN
Ottawa
Victoria
CAnet 4
New York
Geneav
Chicago
FERMI Tier 1
Note Typical View on TRIUMF UCLP GUI
CERN Tier 0
Brookhaven Tier 1
19
TRIUMF/HEPnet Lightpath Object Composition GUI
UBC Campus CWDM Lightpath Object
UoVic Campus 802.11 Lightpath Object
TRIUMF APN
Toronto
Amsterdam
Ottawa
Vancouver
Edmonton
Montreal
To Fermi
Victoria
New York
Geneva
UoVic
TRIUMF
To Brookhaven
Vancouver
Victoria
Lightpath Object for 2 Gbp Tiier 2between
TRIUMF and UoVic
Composition Window
20
UoVic Physics UCLPv2 GUI or workflow tool adds
Router WS to lightpath object
UoVic Physics router resource CLI interface
exposed as a WS
Resource Window
UoVic
TRIUMF
Vancouver
Victoria
Lightpath Object for 2 Gbp Tiier 2between
TRIUMF and UoVic Created by TRIUMF/Hepnet
UoVic
TRIUMF
Vancouver
21
DRAC/UCLP Demo Network
Nortel DRAC
The Power of Web services
Canarie UCLP
Montreal
Ottawa
App
App
Halifax
Toronto
22
SOA Applications
23
CANARIEs i-Infrastructure program

• To adapt Service Oriented Architectures (SOA) to
  process control, instrumentation systems and
  sensor networks
• Applications include manufacturing, oil and gas,
  power systems, water, building management
  systems, environmental control systems, etc
• Built upon CANARIEs initial work on User
  Controlled LightPaths (UCLP)
• Start with large science research facilities such
  as Neptune, Canada Light Source and then expand
  into industrial applications
• www.canarie.ca/ccip

24
Typical Large system today
VPN
USER
Firewall
Process
Process
Process
DMAS
Process
Process
SONET/DWDM
Instrument Pod
SONET/DWDM
Layer 3 switch/router
Layer 2 switch
Sensor
Sensor
Instrument
Instrument
Sensor
25
Service Oriented Architectures
VPN
WS
HPC
WS
USER
Process
Process
Data Management System
WS
WS
Process
Process
LAN
WS
Instrument Pod
LAN
Web service Interface CANARIE UCLP New web
services
WS
WS
Layer 2/3 switch
WS
Sensor
Sensor
Instrument
Instrument
Sensor
26
Science user perspective
WS
CANARIE UCLP
WS AAA process
WS
WS HPC Process
WS
WS
New Web service
Lightpath
WS
WS
New development
ONS15454
WS
NLR or CAnet 4
USER with WSFL binding software
Log Archive Process 2
WS
DMAS
Log Archive Process 1
WS
LAN
WS
UDDI or WSIL service registry
Science Pod
LAN
WS
Sensor/Instrument
WS
User defined WSFL bindings
27
VESPERS Beamline at the Canadian Light Source
1. E-gun Linear Accelerator

• microanalysis with unprecedented sensitivity

4. Beamline End Station
3. Storage Ring
Courtesy of CLSI
28
UCLP-Enabled Virtual Design Studio 3D digital
construction of the Salk Institute
Michael Jemtrud Konstantin Privalov James
Hayes Nicolas Valenzuela Carleton Immersive
Media Studio Carleton University , School of
Architecture, Ottawa (Canada)
29
SOA for Participatory Design Studio

• Service provides are
• network resources (UCLP)
• devices (cameras, displays, rendering computers)
• software (MAYA)
• Provisioning for a PDS session requires
• finding a configuration of network resources,
  devices and software that meets the users needs
• SOA will monitor session
• Does not transport high definition signal
• Demo illustrates how end users can establish
  UCLP connections without knowing details

30
Other SOA Network Projects

• Design Service-Oriented Architecture (SOA) and
  build Web Services for linking research data to
  scholarly publications
• Web services control of undersea HDTV camera
  Neptue
• SOA for military real time simulation

31
Amateurs discover most Supernovas
 http//www.nytimes.com/2002/11/07/technology/cir
cuits/07astr.html?todaysheadlines Nasa and
amateur scientists nightly harvest about 1,000
images, which are shared with other amateur
astronomers over the Internet. Together, they
analyze the pictures for previously undiscovered
supernovas, the remains of collapsed stars.
Over 58 supernovas have been discovered
While most amateur astronomers use computers
to enhance a hobby, the advances in technology
are also blurring the distinctions between
professionals and sophisticated amateurs.
32
Sloan Digital SkyServer

• http//skyserver.sdss.org/en/
• Large database of astronomical data and images
• Available to scientists, students and public
• XML and Java web services interfaces

33
Conclusions

• SOA Cyber-Infrastructure will fundamentally
  transform science and IT
• Better get prepared and learn as much as possible
  and learn about CI and SOA
• Web services
• Resource discovery and consumption
• Publishing services
• Workflow and orchestration
• SOA platforms OGSA, .NET,. Websphere
• Commercialization potentials of integrative
  science and CI are significantly greater than
  with traditional science