Grid and Pervasive Computing

1
Grid and Pervasive Computing

• David De Roure

2
Overview

• What is Grid?
• What is Pervasive?
• The Grid and Pervasive vision
• The Grid, Pervasive and Semantic vision
• Some projects

3
A short history of the Grid

• Science as a team sport
• Grand Challenge Problems of the 80s
• Gigabit Testbed program
• Focus on applications for the local to wide area
• FAFNER
• Factoring via Network-Enabled Recursion
• I-Way at SC 95
• First large-scale grid experiment
• Provided the basis for modern grid infrastructure
  efforts

(Fran Berman, SDSC)
4
Datagrid perspective
www.griphyn.org www.ppdg.net
www.eu-datagrid.org
5
1995 2000 Grid Computing

• Important infrastructure and middleware efforts
  initiated
• Globus, Legion, Condor, SRB, etc.
• Grid book gave a comprehensive view of the
  state of the art
• 2000 Beginnings of a Global Grid
• Evolution of the Global Grid Forum
• Some projects evolving to de facto standards
  (e.g. Globus, Condor)

6
The Grid Problem has moved on

• Resource sharing coordinated problem
  solvingin dynamic, multi-institutional virtual
  organizations

Foster, Kesselman, Tueke
7
Open Grid Services Architecture

• Anatomy vs Physiology
• Present Grid Architecture is a services
  architecture
• Implemented using Web Services Technology
• OGSA provides
• Naming /Authorization / Security / Privacy
• Higher level services Workflow, Transactions,
  Datamining,Knowledge Discovery,
• Exploiting Synergy Commercial Internet with Grid
  Services

8
New Books

• Grid Computing Making the Global Infrastructure
  a Reality, (eds. Fox, Berman, Hey), Wiley and
  Sons 2003.
• The Grid Blueprint for a New Computing
  Infrastructure, (eds. Foster, Kesselman), Second
  Edition Morgan Kaufmann 2003.

9
Grid vs e-Science

• Grid computing is the infrastructure for
  e-Science
• e-Science is about the new science we can deliver
  using this infrastructure
• ECS majors on Grid
• Also applicable to e-Business, e-Health,
  e-Learning, e-Social-Science, e-Anything
• (or is that e-CS? ?)

10
Pervasive computing

• Essentially, devices everywhere
• e.g. Environment
• e.g. Clothes (wearables)
• e.g. Everyday artefacts
• e.g. Sensor arrays
• Also known as Ubiquitous Computing (terms are
  pretty much interchangeable)
• The term ubiquitous comes from Xerox (e.g.
  Weiser 1991) and the Xerox vision emphasises the
  computing being in the background

11
Weiser

• Ubiquitous computing is not virtual reality, it
  is not a Personal Digital Assistant (PDA) it is
  not a personal or intimate computer with agents
  doing your bidding. Unlike virtual reality,
  ubiquitous computing endeavours to integrate
  information displays into the everyday physical
  world. It considers the nuances of the real world
  to be wonderful, and aims only to augment them.
  Unlike PDA's, ubiquitous computing envisions a
  world of fully connected devices, with cheap
  wireless networks everywhere unlike PDA's, it
  postulates that you need not carry anything with
  you, since information will be accessible
  everywhere.

12
Weiser

• Unlike the intimate agent computer that responds
  to one's voice and is a personal friend and
  assistant, ubiquitous computing envisions
  computation primarily in the background where it
  may not even be noticed. Whereas the intimate
  computer does your bidding, the ubiquitous
  computer leaves you feeling as though you did it
  yourself.
• See also Weiser, Mark. The Computer for the
  Twenty-First Century. Scientific American.
  September 1991. pp. 94-104

13
CACM
14
Ambient Intelligence
From MST EU Programme News
15
From ISTAG Scenarios for Ambient Intelligence in
2010 Final Report, Feb 2001, EC 2001
16

• Embedded intelligence services
• Dynamically and massively distributed device
  networks
• Communications
• Trust and confidence enabling toolsEncryption/IPR
  /Dependability
• Cross Media Content
• Multi-modal and adaptive interfacesHeavy
  Content/tagged/mark-up
• Dialogue goal-oriented negotiation (agents)
• Electronics (Micro/nano) scaling down
• Displays
• Power source

17
Users
OWL
Adaptive infrastructures
Grid
IPv6
Devices
18
Building threads across groups
19
Grid and Pervasive
Grid
Pervasive
20
Grid-and-Pervasive
Grid and Pervasive
21
G PvC Relationship

• Pervasive Computing is the means by which the
  digital world of the Grid couples into our
  physical world

22
G PvC Relationship

• Both involve similar computer science issues
• Large scale distributed systems
• Service description, discovery and composition
• Issues of availability and mobility
• Peer-to-Peer

23
G PvC Relationship

• Real Grid computing applications have pervasive
  aspects
• e-Scientists in smart laboratories
• collaborative environments such as Access Grid
• patients using wearable computing technologies
• wherever the Grid meets people!

24
G PvC Relationship

• Real pervasive computing applications have grid
  aspects
• sensor arrays increase the temporal and/or
  spatial density of data, requiring Grid for
  significant computational tasks
• Many PvC RD projects are small scale (due to
  small numbers of devices or small numbers of
  users) how do they scale?

25
G PC Relationship

• Both are visions for the future of computing
  they are the world we are moving into
• We need to understand the relationship between
  the two!

26
And Semantic

• There is a third vision!
• The Semantic Web
• Web is server-to-human user via browser
• Semantic Web is machine-to-machine
• Primarily provides interoperability, needed to
  assemble our Grid and Pervasive apps (preferably
  automatically!)

27
And Semantic

• Semantic Web technologies can also be applied in
  the Grid and Pervasive infrastructure
• e.g. semantic web services
• Pushes the envelope in the Semantic Web research
  community as we are not dealing with static and
  persistent resources

28
Resource Description Framework
29
DAML-S Service Profile
Non Functional Properties
Functionality Description
30
Semantic
Grid
Pervasive
David De Roure
31
Semantic Grid
Semantic Web
Richer semantics
Classical Web
Classical Grid
More computation
Source Norman Paton
32
Knowledgeable Devices
33
Agents
Agent
Interactions
Organisational relationships
Environment
Sphere of influence
34
Ambient Intelligence

• Ambient Intelligence is the manifestation of the
  Semantic Grid in the physical world through
  pervasive computing

35
Some Projects

• These are some examples to illustrate this talk
• Grid
• Pervasive
• Grid and Pervasive
• Most have semantic aspects too

36
Comb-e-Chem
Video
Simulation
Properties
Analysis
StructuresDatabase
Diffractometer
X-Raye-Lab
Propertiese-Lab
Grid
37
myGrid - bioinformatics

• Imminent deluge of genomics data
• Highly heterogeneous
• Highly complex and inter-related
• Convergence of data and literature archives
• Database access from the Grid
• Process enactment on the Grid
• Personalisation services
• Metadata services
• Grid Services Ontologies

38
Geodise
Geodise will provide grid-based seamless access
to an intelligent knowledge repository, a
state-of-the-art collection of optimisation and
search tools, industrial strength analysis codes,
and distributed computing data resources
39
Industrial applications
40
IT Innovation Projects
41
Technical innovation in physical and digital life

• Henk Muller (Bristol), Matthew Chalmers
  (Glasgow), Adrian Friday, Hans Gellerson
  (Lancaster),Steve Benford, Tom Rodden
  (Nottingham), Bill Gaver (RCA), David De Roure
  (Southampton), Yvonne Rogers (Sussex), Anthony
  Steel (UCL)

42
Limited DigitalEnvironment
Mainframes
FTPShared InfoStores
Multi User Machines
Conferencingand GroupwareSystems
NetworkedPCS
Growing Presence of the Digital in the Physical
World
Increasingly RichDigital environments
Web andVirtual Worlds
Mobile DevicesWearablesNovel Displays
Grid Computing
Fully Converged Digital and Physical Environment
?
Seamless Meshing of Digital and Physical
Interaction
43
Ambient Wood
44
(No Transcript)
45
(No Transcript)
46
MOHICAN - Cooperative Problem Solving

• Using multiple cooperating agents
• Such agents should have the expertise and
  resource to achieve the target (call
  establishment)
• An agent decides to select an agent to pass a
  request to forward a call
• Effective modelling of agent capabilities
  updated knowledge about network conditions
  facilitate such decision making
• Cooperation (see Figure)
• Agents communicate individual (partial) views of
  local network conditions to diagnose/detect
  network load/faults
• Agents share individual resource/expertise to
  route calls

47
MOHICAN - Competitive Market-based Agents

• Wireless network calls originate from any node.
• Problem limited node capacity.
• Aim maximise call throughput.
• Entities sellers/buyers/auctioneers
• - Seller agents one per node own and lease out
  their node relaying capacity.
• - Buyer agents one per node purchase node
  capacity on behalf of callers.
• - Auctioneer agents conduct auctions and
  allocate resources to buyers.

48
IPv6 Projects

• 6NET Large-Scale International IPv6 Pilot
  Network
• Euro6IX European IPv6 Internet Exchanges
  Backbone
• 6WINIT IPv6 Wireless Internet IniTiative

49
ASAP Advanced Specialization and Analysisfor
Pervasive Computing
EU IST FET Programme Project Number
IST-2001-38059 Keywords Program Specialization,
Program Optimization, Pervasive Computing,
Program Analysis, Cost Analysis, Resource
Consumption Analysis, Abstract Interpretation,
Global Analysis, Logic and Constraint
Programming, Declarative Programming. Partners
University of Madrid, University of Southampton,
University of Bristol, University of Roskilde
Objectives Automate as much as possible the
development of software systems for pervasive
computing platforms. Develop a novel method for
system development for pervasive computing,
based on the automatic generation of specialized
systems e.g., from general, already existing
ones. Develop an integrated tool and make it
available as open source code.
50
Trusted Software Agents and Services for
Pervasive Information Environments (TSA)

• ECS (mjb, mal) QinetiQ
• Investigate trust issues in the context of
  software agents and services in pervasive
  information environments
• Develop strategies to assure trust in these
  systems, underpinned by formal validation
  techniques
• Scenarios / small-scale demonstrators to help
  understand issues and allow for validation of
  developed approach

51
The Floodnet approach

• High resolution data to calibrate models, tune
  real-time model runs, guide emergency services
  and reduce false positive/negative warnings
• Rich data (with depth, velocity, floodzone
  dynamics etc) to guide contingency planning and
  real-time response
• Scenario testing by the sensor network to
  identify likely cause/pattern of the flood

52
More sensors enable higher spatial and temporal
resolution
Using Explicit Intelligence
53
SECOAS System Architecture (BT)
Using Implicit Intelligence
Surface moored buoy, VHF comms (shore and/or
other buoys), u/w comms and control of sensor
packages, data storage.
Basic disposable sensor packages (e.g temp,
turbidity). Limited storage and power. u/w comms.
Enhanced disposable sensor package (e.g. temp,
pressure, turbidity, current). Increased storage
and power, u/w comms.
54
Amorphous Computing

• An amorphous computer consists of a very large
  number of distinct processing elements, randomly
  distributed the elements are programmable, they
  execute asynchronously and each element can only
  communicate via unreliable local broadcast. We
  explore this through simulation.

55
Grid AND Pervasive Projects
The Grid
IPv6
Devices
e.g. Model of floodplain e.g. Medical signal
processing
Can we shift the computation towards the devices?
56
Medical Devices

• Providing medical information onto the Grid
• Focus on combining medical information with
  motion information
• Information reported remotely to mobile devices

57
GRID-based Knowledge Services for Triple
Assessment

• Ontology Services covering concepts and
  processes in TA.
• Enrichment and Annotation Services enrichment of
  TA content, annotation of images using ontology
• Medical Images and Signals Services dynamic
  registration and fusion of images and their model
  based interpretation
• Grid Reasoning Services classify, cluster and
  retrieve images, based on concepts or features

58

• CoAKTinG will provide tools to assist scientific
  collaboration by integrating
• intelligent meeting spaces
• ontologically annotated media streams from online
  meetings
• decision rationale and group memory capture
• meeting facilitation
• issue handling, planning and coordination support
• constraint satisfaction
• instant messaging/presence.

59
(No Transcript)
60
Conclusions

• Grid and Pervasive Computing are important
  visions for the future of electronics and
  computer science
• GPC aims to consolidate current activities and
  facilitate strategic development of the combined
  area