Community Webs (C-Web): Functionality and Architecture Issues V. Christophides Computer Science Department, University of Crete Institute for Computer Science - FORTH Heraklion, Crete

1
Community Webs (C-Web) Functionality
andArchitecture IssuesV. ChristophidesCompute
r Science Department, University of
CreteInstitute for Computer Science -
FORTHHeraklion, Crete
2
What is C-Web ?

• Set-up methodologies and infrastructure for fast
  deployment and easy management of
  knowledge-intensive Web applications in
  communities requiring
• effective knowledge assimilation, elicitation,
  ...
• efficient query answering

3
C-Web Main Idea Virtual XML Warehouse
Knowledge
Single Point of Access

• Main Goal to provide a generic platform for
  describing, organizing and querying various XML
  resources according to a concept taxonomy shared
  by a specific community

4
C-Web Objectives

• Reuse existing knowledge structures (e.g.
  ontologies, thesauri)
• Integrate easily heterogeneous XML resources
  (e.g. data, documents)
• Provide an intelligent information access (i.e.
  conceptual querying and browsing)
• Support collaboration facilities expertise
  management (e.g. annotations)
• Enable automatic generation of new information
  resources (e.g. e-books)

5
C-Web Functionality Current Status

• Support for creating community conceptual models
• integration of existing ontologies and thesauri
• definition different viewpoints
• Support for describing and integrating resources
• resource content description metadata (CDM)
• resource structure mapping metadata (SMM)
• Support for conceptual browsing and querying
• High level property-centric queries to resources
• Querying both conceptual schemata and related
  instances
• Support for collaborative resource annotation
• Support for intelligent information publishing

6
The main C-Web Requirement Interoperability

• Heterogeneity is not a drawback, but a feature of
  autonomous information resources in large scale
  distributed systems
• Interoperability the ability to uniformly share,
  interpret and manipulate data and documents from
  heterogeneous resources

Semantic Structure
vocabularies
Syntactic
viewpoints contexts
query language dialects
abstraction aggregation details
data formats
Functional System
domain data models
communication protocols
transaction processing
security policies
7
C-Web Design Principle Repository Independence
Domain Model
Pointillism
Museum
Artifact
Artist
Neo-Impressionism
Conceptual
Impressionism
Painting
Fine-Art
Archeological
Sculpture
Sculptor
Painter
Source 1 XML enabled DBMS
Source 2 XML Repository
ltelementType nameArtWork"gt ltsequencegt
ltelementTypeRef name"Title"
minOccur"1"/gt ltelementTypeRef
nameCreator"
minOccur"1"/gt lt/sequencegt. lt/elementTypegt
lt!ELEMENT MusArtifact (Name, Event)gt lt!ATTLIST
MusArtifact material CDATA IMPLIED
size CDATAIMPLIEDgt lt!ELEM
ENT Event (Person,Place,Date)gt lt!ATTLIST Event
nature (creationacquisition)gt lt!ELEMENT Person
(Name, Nation, Life?gt ...
Logical
XML SQL XSQL Servlet
XQL Xpath Servlet
Physical
8
C-Web Related W3C Standards

• Semantic Interoperability Content Description
  Metadata Standards
• ontologies (e.g. ICOM/CIDOC), thesauri (e.g.,
  ULAN, TGN, AAT), metadata element sets (e.g.
  CIMI/Aquarelle Z39.50 profile)
• Resource Description Framework (RDF) for
  expressing semantics
• Structural Interoperability Schema languages for
  specifying logical structure of Web resources
• DTDs, XML Schema
• Syntactic Interoperability Markup languages for
  exchanging (semi-) structured data over the Web
• XML, XLL, ...
• Functional Interoperability Data Manipulation
  languages for (semi-) structured data over the
  Web
• XPath, XQL, XSL, ...

9
From RDF Schemata to XML resources
r
rdfsLiteral
d
artistSculpture
artistsculpts
artistArtist
d
RDF schema
r
d
s
s
r
artistmaterial
artistPainter
artistSculptor
artistlives_in
artistMaterial
t
t
t
RDF/XML metadata
August_Rodin
artistsculpts
Paris
The Burghers of Calais
Iron
artistlives_in
artistsculpts
The Gates of Hell
artistmaterial
about
www.artist.gr/august_rodin
s rdfssubclassOf
XML Resources
ltARTISTgt ltNAMEgtAugust Rodin ltLIVESgtParis
ltWORKgt ltTITLEgtThe Gates of Hell
ltMATERIALgt Iron ..
t rdftype
d rdfsdomain
r rdfsrange
10
XML Resources and C-Web Metadata
XML Structure
C-Web Schema
of period
Period
title
Museum Object
String
isa
has_style
Artifact
Natural Object
Style
Material
consists of
ltrdfRDF xmlnsrdf"... xmlnsrdfs"..."
xmlnssmycweb.forth.gr/..."gt ltrdfDescription
aboutwww.artist.gr"gt ltsmappingsgt
ltrdfBaggt ltligtltDescription aboutArtist.Work.Ti
tlegt ltsmap rdfresourcestitle/gtlt/li
gt ltligtltDescription aboutArtist.Work.Materialgt
ltsmap rdfresourcesMaterial/gtlt/ligt
lt/rdfBaggt lt/smappingsgt
ltARTISTgt ltNAMEgtAugust Rodin ltLIVESgtParis
ltWORKgt ltTITLEgtThe Gates of Hell
ltMATERIALgt Iron ..
ltARTISTgt ltNAMEgtAugust Rodin ltLIVESgtParis
ltWORKgt ltTITLEgtThe Gates of Hell
ltMATERIALgt Iron ..
C-Web Resource Description Interface
ltARTISTgt ltNAMEgtAugust Rodin ltLIVESgtParis
ltWORKgt ltTITLEgtThe Gates of Hell
ltMATERIALgt Iron ..
11
The C-Web Architecture
Artist
Painting
Client Tier
Museum
Query Browsing Interface
Painter
Virtual Document Render
http
Schema Editor
XML/XSL
RDF/XML
CWEB/APP Server
Middleware APIs
Session Manager
Logical Middle Tier
Metadata Store
RDF/XML Loader
XML/XSL Processor
Query Engine
URL Resolver
RDF/XML
XML
XML
XML
Resources
Artist
http
XML Wrapper

URL
Well-formed
XML enabled DBMS
Other docs
XML docs
Resource Description Interface
Resource Annotation Interface
on the Intranet
on the Web
e.g. mails,
news, reports
12
C-Web Middleware Main Features

• Genericity capture any XML structure (various
  DTDs), any form of XML semantics (DTDs, XML
  Schema), any XML access interface/protocol (XQL,
  XLL)
• Scalability w.r.t. the volume of XML resources,
  the number of XML repositories, network and
  server load, etc.
• Extensibility evolution of XML resources
  semantics and structure does not affect the main
  processing components and interfaces
• Openness rely on standards APIs allowing to
  plug and play the same components services in
  various applications, domains, etc.

13
The C-Web Architecture Pending Issues

• Schema Editor Standalone application or client
  of the Middleware?
• Where are stored large thesauri/ontology before
  their integration?
• Who is responsible for Schema Validation (from
  scratch vs. integrated)?
• What communication protocol we need with the
  C-Web Middleware?
• Resource Description Interface Loose or Tight
  coupled with the Middleware?
• How C-Web Schema browsing/querying is
  implemented?
• Where we can find the XML DTDs/Schemata of
  resources?
• Who is responsible for Resource Description
  Validation?
• What communication protocol we need with the
  C-Web Middleware?
• Metadata Store What persistence support we need?
• What is an efficient RDF storage model (indexing
  clustering)?
• Did we also need to support updates/versions
  (versioning model)?
• What are the authentication/security policies
  (RDF/XML with signatures)?
• What is the result form of a C-Web query
  (triples, statements or objects)?

14
C-Web Communication Protocol 3 Alternatives
Client
Server
Query
Reply
Synchronous a blocking query waits for an
expected reply
Query
Handle
Client
Server
Next
Reply
Next
Reply
Server maintains state replies sent individually
when requested
Subscribe
Client
Server
Reply
Reply
Reply
Asynchronous a nonblocking subscribe results in
replies
15
Towards a C-Web Physical Architecture
Resource Description Interface
Query Browsing Interface
Schema Editor
RDF/XML Schema
RDF/XML Descriptions
RDF-QL
Internet
CWEB/APP Server
Session Manager
RDF/XML Parser Loader
Persistent Namespace Service
RDF/XML Query Engine
Ethernet
NM3 C-Web Schema Instances
Metadata Store
Metadata Store
Metadata Store
NM1 Ontologies
NM2 Thesauri