HKU CSIS DB Seminar: Web Services Oriented Data Processing and Integration

1
HKU CSIS DB SeminarWeb Services Oriented Data
Processing and Integration

• Speaker Eric Lo

2
Presentation Outline

• Introduction on Web Services Related Work
• Web Services
• SOAP
• WSDL
• UDDI registry
• Web Services Oriented Data Processing and
  Integration
• Introduction to WSIPL-Web Service Integration
  and Processing Language
• Application Examples

3
What is a Web Service?

• public class Exchangepublic float
  getRate(String country1, String
  country2) return 123.45
• 1) Compile it to Exchange.class
• 2) A Web Server (e.g. Tomcat) a SOAP server
  (e.g. Apache SOAP)
• 3) Put Exchange.class to a directory that the
  SOAP server know
• It is already a Web Service!

4
What is a Web Service?

• Too simple? What is the Web Service usage?
• It is simple! So simple that seems no big deal?
• Clients connect to the Exchange services can get
  back 123.45
• Identical to traditional Remote Procedure Call
  (RPC) in distributed computing
• Question? Why not use current Web Application?
• E.g. Clients connect to a Web Server with some
  HTML form page to feed in input data on the form
  and get back a HTML page with 123.45 (the Web
  application can be a Java Servlet, CGI program or
  .ASP script)
• The WS paradigm is to let different Web
  ServiceS to Integrate and Cooperate together

5
What is a Web Service?

• If more than one Exchange Web Services, e.g. one
  Exchange WS is for answering exchanging rate
  between US and HK dollar (e.g. provided by HK
  Bank), another WS is for answering exchange rate
  between US and Euro dollar
• 2 Web Services can integrate together to provide
  a new Web Service

6
The heart of Web Services-SOAP

• SOAP is the standard protocol for XML messaging
  and RPC communication on all platform, all
  languages.
• Stands for Simple Access Object Protocol
• Is actually ride on existing communication
  protocol, e.g. HTTP, SMTP because XML is also
  just a text message

7
What is SOAP?

• Is a communication protocol
• Is for communication between applications
• Is a format for sending messages
• Is designed to communicate via Internet
• Is platform independent
• (of course?XML?text)
• Is language independent
• Is based on XML
• Is simple and extensible
• Is a W3C standard

8
Life Before SOAP

• B4 SOAP, applications communicate using CORBA,
  they are RPC, application A call B methods and
  return something to A. Even its work, they are
  not for http (Internet). IIOP born for COBRA and
  RMI for Internet, but still need to fight with
  firewall, proxy, security and compatibility.

9
Clients invoke SOAP

• Java client need to import SOAP stuffs this time!
• E.g.
• Import org.apache.soap.Import
  org.apache.soap.rpc.public class client
  url http//soapserver call.seturl(url)
  OUTPUTcall.getRate(HK, US)

10
SOAP Message at a glance

• If capture the underlying SOAP message, can find
  format like this
• ltSOAP Envelopgt
• ltSOAP Headergt
• SOAP body
• lt/SOAP Envelopgt

11
Captured SOAP message
12
Web Services Description Language (WSDL)

• Describing the web services you provided
• Usage?
• Let clients know which operation you provided,
  how they bind with, what you have used as
  protocol via SOAP (http, ftp, smtp)
• If the operation of your web service fulfill
  predefined sets of roles (types), you can then
  compete with other similar service providers
  (e.g. you provide function with getRate() and
  transfer(), then you can be type of BANK)

13
WSDL

• WS code (server side java code) ?? WSDL ? Client
  stub ? Client code
• Where do the WSDL files locate?
• As a file in the service site
• UDDI registry

14
UDDI-Universal distribution, discovery, and
interoperability

• Similar concept to yellow page
• New services register to the UDDI registry
• By writing code (import UDDI library, uddi4j)
• Or simple done it on web
• Clients can search it
• By writing code again(uddi4j)
• Or by web
• Returning the WS information and the
  corresponding WSDL file or the URL to that WSDL
  file

15
Web Services Oriented Data Processing and
Integration

• Business applications often require intensive
  human programming efforts on data processing and
  integration
• E.g. A company may want to enquire the total
  inventory level of a particular stock in
  different regions, while each regional office
  maintain its own inventory database

16
Before and After XML

• Before - Need low level programming
• Network programming
• Database query
• Not flexible (business requirement varies time to
  time)
• After
• Network programming is easier (handle XML text
  only)
• Simple handle of data transform (e.g. 2 databases
  has different field name)

17
After XML

• However, flexible problem is persist
• Very tedious to write a new application for
  different purpose with minor changes
• Introduce new efforts on handling XML parsing,
  transforming (e.g. XML ? HTML)

18
Our approach

• A novel approach for data integration and
  processing
• WSIPL (stands for Web Services Integrating and
  Processing Language)
• Key Requirements
• Web Service Orientation
• Coordination
• Data Processing and Integration

19
Web Service Orientation

• All data sources and processors are exposed as
  Web Services
• Reusable
• Ready for incorporate and coordinate under
  standard
• For legacy (existing) operations, add a Web
  Service Interface (by wrapper)

20
Coordination

• WSIPL provides constructs
• W3C standard compliance
• Execution
• IF-THEN-ELSE
• FOR-EACH
• SWITCH-CASE
• EXCEPTION
• same as XSLT and XPath, thus can push all
  efforts to a XSLT processor
• Parallelization
• If no dependency between web services, executes
  in parallel to increase efficiency

21
Data Processing and Integration

• Validation
• Validate with XSchema
• Transformation
• E.g XSLT
• Handling data field problem
• Composition
• E.g. summary of inventory data

22
A WSIPL example

• ltwsiplscriptgt
• ltwsiplsource nameUSDollar/gt
• ltwsiplvariable nameUSDollar
  select/data/namevaluepair_at_namedollars/gt
• ltwsipltask nameHKBankgt
• ltwsiplcall urlhttp//hkbank.com/opensevic
  e urnCurrencyServices operationExchange
  styleRPCgt
• ltwsiplparam namedollars
  valueUSdollars/gtlt/wsiplcallgt
• lt/wsipltaskgt
• ltwsipltask nameEuroBankgt
• ltwsiplcall urlhttp//Eurobank.com/opensev
  ice urnCurrencyServices operationExchange
  styleRPCgt
• ltwsiplparam namedollars
  valueUSdollars/gtlt/wsiplcallgt
• lt/wsipltaskgt
• ltwsipltask nameIntegrategt
• ltwsiplinclude nameHKBank/gt
• ltwsiplinclude nameEuroBank/gt
• lt/wsipltaskgt
• ltwsiplresponse nameIntegrate/gt

23
WSIPL system reference architecture
24

• End