The Most Universal Communication Vehicle Is Still: TEXT

10. Web Services

Web services are software programs that use XML to exchange information with other software via common internet protocols. They enable the programmable (not just interactive) Web.

· The most universal communication vehicle is still: TEXT

Ø More exactly: (semi-structured structured) text: XML

Basic goals

· Expose services to other processes

Ø Interconnection over internet or intranet

Ø Distribution and integration of application logic

Ø Component-like, reusable black boxes

Ø Loosely coupled, fine grained services

· Supported by many platforms; Java, the .NET Framework ...

Ø JAXP, ASP.NET Web Services

· Based on open standards

Ø HTTP, XML, and SOAP

· Overcome of some shortages of available solutions

Ø CORBA or DCOM scale wrong, because of tight coupling

Ø DCOM is limited to Windows

Ø Component based architecture required

§ CORBA is “remoting”, the component model is hardly implemented yet

§ EJB is restricted to Java

Ø ASPs provide typically large, coarse grained interfaces

Basic Technology

· WSDL (Web Services Description Language) – WHAT

· UDDI (Universal Description, Discovery and Integration) –WHERE

· SOAP (Simple Object Access Protocol) – HOW

Ø Soap messages are texts – in XML format

Ø Interoperability relies on XML – that’s all

Web Services Infrastructure

An Example

· ZipCode Resolver

Ø Accepts a (maybe partial) address as input and returns the postal code

Ø Methods

§ string FullZipCode(string accessCode, string address, string city, string state)

Given a valid accesscode, street address, city, and state, returns the ZIP code in NNNNN-NNNN format. If an error occurs, returns 00000-0000 instead. Use accessCode of '0' or '9999' for testing.

§ string ShortZipCode(string accessCode, string address, string city, string state)

Given a valid street address, city, and state, returns the ZIP code in NNNNN format. If an error occurs, returns 00000 instead.

§ . . .

EraServer.NET ZipCodeResolver Web Service Test Page

Testing .NET SOAP Proxy

.NET SOAP Return Value:

1 MICROSOFT WAY, REDMOND, WA, 98052-6399, 98052

Address......
City......
State......

The Web Services Technology Stack

· On the bottom level: well-known Internet and Web protocols

Ø TCP/IP and HTTP

· Core XML level

Ø XSL, DTD; XML Schema

· Horizontal XML Vocabularies

Ø Functionality across industry boundaries

Ø Developer and industry consortia create consensus and standards

Ø E.g. ebXML for e-business

· Web Services Technologies build on top of the core and horizontal (if available) vocabularies

· Vertical languages are XML-based processes for single application domains

· E.g. RosettaNet defines business processes for IT, electronics and semiconductor industries

Vertical Language / Vertical Language / Vertical Language / Vertical Language
Web Services Technologies: SOAP, WSDL, UDDI
Horizontal XML Vocabularies: ebXML …
Core XML Processing: XML, XML Schema, XSLT
Web Framework: Internet Protocols; HTTP, TCP/IP

10.1. XML (Extensible Markup Language)

· Language specification for describing data (meta language)

Ø Syntax rules

Ø Syntax & Grammar for creating Document Type Definitions

· Widely used and open standard

Ø Defined by the World Wide Web Consortium (W3C)

· Designed for describing and interchanging data

Ø Data is logically structured

Ø Human readable, writeable and understandable text file!

Ø Easy to Parse; Easy to Read; and Easy to Write!

· Metadata

Ø Data that describes data; data with semantics

· Looks like HTML…but it isn’t!

Ø Uses tags to delimit data and create structure

Ø HTML is designed

§ for presentation

§ with a fixed set of tags

Ø XML does not specify how to display the data

§ The tags can be used for any purpose, e.g. search

Ø XML has few predefined tags, it is extensible

Ø XML is case sensitive (HTML is not)

Components of an XML Document

· XML Processing Instruction

Ø Declares the document as an XML document

<?xml version = “1.0” encoding = “UTF-8”

standalone=”yes” ?>

Ø Version information

§ Currently always 1.0

Ø Encoding type

§ UTF-8, UTF-16, ISO-10646-UCS-2, etc

Ø Standalone declaration

§ Indicates if there are external file references

Ø Namespace declaration(s)

Ø Processing Instructions (for applications), etc

· Document Type Declaration (DTD) or XML Schema

Ø Internal DTD

<!DOCTYPE CustomerOrder [

Ø External DTD reference

<!DOCTYPE CustomerOrder

SYSTEM "http://www.myco.com/CustOrder.dtd">

· Document Instance

Ø This is the XML document instance; the “XML-ized” data

Validation of an XML document

· Syntax check, or checking for being “well-formed”

Ø Check against basic XML syntax rules

· Validation

Ø Check against Document Type Declaration or

against XML Schema

Example customer order

<?xml version = “1.0” encoding = “UTF-8” ?>

<! DOCTYPE CustomerOrder

SYSTEM “http://www.myco.com/dtd/order.dtd” >

<LName> Smith </LName>

</Person>

91 Park So, New York, NY 10018 </Address>

Hauptstrasse 55, D-81671 Munich </Address>

</Customer>

</Orders>

</CustomerOrder>

· Document Root Element

Ø Required if a document type declaration exists

Ø Must have the same name as the declaration

· Elements

Ø Can contain other elements and have attributes assigned to

Ø May or may not have a value

· Attributes

Ø Properties that are assigned to elements

Ø Provide additional element information

DTD: Defines document syntax and semantics

<?xml version = “1.0” encoding = “UTF-8” ?>

<!DOCTYPE CustomerOrder [

<!ELEMENT CustomerOrder (Customer, Orders*) >

<!ELEMENT Customer (Person, Address+) >

<!ELEMENT Person (FName, LName) >

<!ELEMENT FName (#PCDATA) >

<!ELEMENT LName (#PCDATA) >

<!ELEMENT Address (#PCDATA) >

<!ATTLIST Address

AddrType ( billing | shipping | home ) “shipping” >

<!ELEMENT Orders (OrderNo, ProductNo+) >

<!ELEMENT OrderNo (#PCDATA) >

<!ELEMENT ProductNo (#PCDATA) > ]>

XML parsing

· The parser checks the validity of the document with the help of the DTD or XML Schema document

· It usually builds up a DOM (Document Object Model) tree in main memory

Ø Elements and values are the nodes of the tree

Ø Attributes are values assigned to the nodes

· The application can use the procedures of an XML parser to access individual elements, values and attributes

· The parser may also generate call-backs on tags (SAX)

XML Schema

· Similar goal as a DTD, both are XML document definition languages

· XML Schema is written in XML

· Unlike DTD’s, XML Schema are Extensible – like XML!

· More verbose than DTD’s but easier to read & write

· Defining datatypes

Ø The simple or primitive datatypes

§ Based on (or derived) from the Schema datatypes

Ø Complex types

· Facets

Ø Constraining values, e.g.

</restriction>

</ simpleType >

· Declaring data types

· Two kinds of datatypes: Built-in and User-defined

· Built-in

Ø Primitive Datatypes

§ string, double, recurringDuration, etc.

Ø Derived Datatypes

§ CDATA, integer, date, byte, etc

§ Derived from the primitive types

§ Example:

· integer is derived from double, time is derived from recurringDuration, CDATA is derived from string

Ø User-defined

§ Derived from built-in or other user-defined datatypes

Examples

</sequence>

</complexType

XSLT: Stylesheets & Transformations

· Offers the ability to transform one XML document into another

Ø It is typically used to control the presentation of an XML document, i.e. to generate an HTML document

Ø Can be used for any kind of transformation

· XSLT is an XML programming language

Ø Can have rules, evaluate conditions, etc

Using XSLT to create a formatted table

10 / 100
10 / 200
20 / 501

An example XSLT definition to create a formatted table

<?xml version='1.0'?>

<xsl:stylesheet
xmlns:xsl=

"http://www.w3.org/TR/2000/CR-xsl-20001121/">

<xsl:template match="/"> <!—Match the entire documet -->
<HTML>
<BODY>

<xsl:for-each select=“Customer/Orders/OrderNo">

<xsl:for-each select=“Customer/Orders/ProductNo">

<TR>
<TD> <xsl:value-of select=“OrderNo"/</TD
<TD> <xsl:value-of select=“ProductNo"/</TD

</TR>

</xsl:for-each>

<TR</TR>

</xsl:for-each>

</TABLE>

</BODY>
</HTML>

</xsl:template>

<xsl:stylesheet>

10.2. SOAP (Simple Object Access Protocol)

"Lightweight mechanism for exchanging structured and typed information between peers using XML."

· Simple Object Access Protocol 1.1 (1.2 is coming)

Ø Object: any active entity that can accept and respond to messages – NOT in the usual oo sense

· Designed both for messaging and RPC

· Specifies three parts:

Ø SOAP envelope for message framework

§ “Envelops” XML “letters” (e.g. BizTalk standard)

Ø SOAP encoding for standard, interoperable marshaling

Ø SOAP RPC as remote procedure call convention on the top of HTTP

§ HTTP is a connectionless request/response mechanism

§ GET is the access protocol for resources

§ POST is the application protocol for interaction

§ Calls are directed to a URI

§ Headers specify options

§ Response codes indicate success/failure/status

· A simplified example

HTTP/1.1 POST /soap/mycomponent 
Content-Type: text/xml 
SOAPAction: MyComponentMethod
<SOAP:Envelope>

<SOAP:Body>

<m:MyMethodArgumentsRoot>
<MyArgument>1</MyArgument>
</m:MyMethodArgumentsRoot>
</SOAP:Body>

</SOAP:Envelope>

SOAP as RPC Protocol

· Marshals procedure calls via XML

· Transport can be HTTP, SMTP, etc.

Ø Any text based protocol can transport SOAP data

Ø “Anyone” can implement it easily

Ø Designed for existing Internet infrastructure

· No expensive middleware required

· Default wire-protocol for .NET Remoting

Ø The .NET framework, JAXP etc. provide built-in support

SOAP as Messaging Protocol

· Descriptive container for any XML data exchange

· Transport can be SMTP, X.400, Queues etc.

· May rely on a message provider: in this case asynchronous

· Mail-like: Similar to Exchange Forms idea

· Basis for BizTalk Framework

· Interoperable and platform agnostic

The SOAP Protocol Layers

· SOAP Envelope

Ø Uniform container for XML messages

Ø Defined by XML Schema

§ http://schemas.xmlsoap.org/soap/envelope/

§ this name space must be referenced to

Ø SOAP:Header for handling instructions

Ø SOAP:Body for XML Payload

Ø Specifies order, cardinality for structural elements

Ø Defines SOAP:Fault for communication errors

HTTP/1.1 POST /soap/myservice
Content-Type: text/xml
SOAPAction: MyInterface#MyComponentMethod

<SOAP:Envelope>
<SOAP:Header> 
<m:myHeader xmlns:m=“myURI“

SOAP:mustUnderstand=“0“> HeaderInfo

</m:myHeader> 
</SOAP:Header>

<SOAP:Body>
<m:MyMethodArgumentsRoot xmlns:m=“myURI“ >
<MyArgument>1</MyArgument>
</m:MyMethodArgumentsRoot>
</SOAP:Body>

</SOAP:Envelope>

· SOAP Encoding

Ø Uniform type and encoding system for data in SOAP Envelopes

Ø Encoding is defined by XML schema

§ http://schemas.xmlsoap.org/soap/encoding/

Ø Usual options of format / encoding type

§ document / literal (usually default)

§ rpc / encoded

Ø Defines rich type system:

§ Simple Types, Enums

§ Compound Values, Structures and References

§ Arrays (incl. Sparse Arrays, Partial Arrays)

· SOAP RPC

Ø Interface/Service identified by URI

Ø Method invocation is modeled as structure

Ø Parameter order for [in/out] in response must be the same as in request.

Ø Behavior on error (return SOAP:Fault)

· Example for calling a Web Service from Java

import java.io.*; import java.net.*;

public class JavaClient { // Calls GetTime of TimeService

public static void main(String [] args) {

try {

URL url = new URL("http://localhost/time/TimeService.asmx");

// Build up HTTP connection to service

HttpURLConnection conn = (HttpURLConnection)url.openConnection();

conn.setRequestMethod("POST");

conn.setDoOutput(true); conn.setDoInput(true);

// Create SOAP request to call GetTime

conn.setRequestProperty("Content-type", "text/xml;charset=utf-8");

conn.setRequestProperty("SOAPAction", "http://tempuri.org/GetTime");

String msg = "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n" +

"<soap:Envelope" +

"xmlns:soap=\"http://schemas.xmlsoap.org/soap/envelope/\">\n" +

"<soap:body>\n" +

" <GetTime xmlns=\"http://tempuri.org/\" /> \n" +

"</soap:body>\n" +

"</soap:Envelope";

byte [] bytes = msg.getBytes();

conn.setRequestProperty("Content-length", String.valueOf(bytes.length));

OutputStream out = conn.getOutputStream();

out.write(bytes); // Send SOAP request

out.close();

BufferedReader in =

new BufferedReader(new InputStreamReader(conn.getInputStream()));

System.out.println("SOAP-Response:");

String inputLine = in.readLine(); // Take SOAP response

while (inputLine != null) {

System.out.println(inputLine);

inputLine = in.readLine();

} // while inputLine != null

in.close();

} catch (Exception e) {

System.out.println("error: " + e.getMessage());

} // try-catch

} // main

} // JavaClient

After execution we get the following result:

SOAP-Response:

<?xml version="1.0" encoding="utf-8"?>

<soap:Envelope

xmlns:soap = “http://schemas.xmlsoap.org/soap/envelope/”

xmlns:xsi = “http://www.w3.org/2001/XMLSchema-instance”

xmlns:xsd = “http://www.w3.org/2001/XMLSchema”>

<soap:Body>

</GetTimeResponse>

</soap:Body>

</soap:Envelope>

10.3. WSDL

(Web Service Description Language)

· Is a kind of "TypeLib for SOAP"

· Usually automatically created

Ø Nevertheless, sometimes must be read by people

· Uniform representation of services

Ø Transport Protocol neutral

Ø Access Protocol neutral (not only SOAP)

· Describes:

Ø Schema for Data Types

Ø Call Signatures (Message)

Ø Interfaces (Port Types)

Ø Endpoint Mappings (Bindings)

Ø Endpoints (Services)

· Uses XML grammar, defining:

Ø Services and ports that communicate via messages

Ø Binding

§ Specify a protocol or a data format for a message or a port

§ Extensions for SOAP 1.1, HTTP GET/POST, and MIME

· Public description of a Web Service and its content

Ø WSDL contract

· Example

10.4. Web Service Discovery

DISCO

· Dynamic Discovery of Services by Microsoft

· Is a kind of “bookmark”

Ø Contains URIs to WSDL documents

· Easy discovery model for HTTP

1. Call endpoint with GET

2. Check whether this is a DISCO file or a redirect

3. DISCO file contains link to WSDL

4. If redirect goto 1

UDDI

(Universal Description and Discovery Interface)

· A search engine for WebServices

· WebService-Programmable "Yellow Pages"

· Advertise Sites and Services

· May point to DISCO resources

Distributed Systems, 10. Web Services László Böszörményi and Harald Kosch