5. Technical Components of Portals
A comprehensive portal solution incorporates a variety of internet and application-related technology components. Because the goal of the portal is to provide a single view to the end user of information coming from multiple sources, the list of possible technologies utilised within portals is endless. However, the following list is a good introduction to the most common technologies that are included in Portal solutions. Some of these technologies (such as Web services) are constantly evolving, while others (such as the portal database, which is usually built on a SQL database) are relatively stable technologies.

Application Server - Application servers are typically J2EE compliant and provide the underlying development and run-time infrastructure for the portal. Examples of application servers include iPlanet, BEA WebLogic, IBM Websphere, Oracle and Sybase Application Server. Many of the application server vendors are incorporating "portals" as add-ons to their base product. For example, IBM Websphere Portal Server, Oracle Portal, BEA Portal and Sybase Enterprise Portal are all built on top of the corresponding application server and in some cases are sold as one package. Several of the stand-alone portal products, such as Plumtree, Epicentric and Corechange have Java components or are Java-based and take advantage of an application server.
Web Server - The Web Server works in conjunction with the application server to provide the run-time environment for client requests. The web servers used with portals are standard HTTP web servers, such as Microsoft Internet Information Server (IIS), apache, etc. When an end user brings up the portal page, the web browser makes a request of the web server. The web server then passes the request to the application server. The portal (and its associated Portlets) runs on top of the application server.
Database - Most portals have an underlying database (such as Oracle, DB2, Sybase, or SQL Server) that they use to keep track of information specific to the portal - such as users, personalisation settings, available web services/Portlets and security. This use of the database is in addition to a transactional system's database (e.g. ERP, CRM or SCM system) that a portal might query to present application specific data to end users.
Taxonomy - A taxonomy is a classification scheme to organize a collection of information. Ideally, a taxonomy would take a group of documents and make it easy to browse, search or otherwise navigate information that the user is interested in. A taxonomy is analogous to a folder structure, with the additional functional components of metadata for document classification as well as the rules for categorisation. Most enterprise portals have a taxonomy for this purpose, as well as some may have automatically generated taxonomies generated based upon the metadata provided.
Crawler - A crawler is an automated process that reads, indexes and classifies documents at a pre-determined interval. A web crawler, for instance, would crawl target web pages periodically to determine if the content has changed. The content is then indexed into the taxonomy so that end users can easily find it. The crawler doesn't necessarily make another copy of the crawled document; rather it indexes it by creating a virtual card that describes the document. The card then lives in the portal index.
Metadata repository - A metadata repository contains metadata about the content within the portal and about the structure of that content. This includes the metadata about the taxonomy, as well as the metadata for the individual documents. For example, each of the documents placed in a folder called Clients might have a metadata field called "Client" which would have one or more values. The value of the Client field for a particular document is metadata about that document.
Portlet - A Portlet can be thought of as a "building block" of a portal. It is a user-interface for presenting data and functionality from multiple applications on a single web page. Portlets encompass the presentation layer and the business logic. They also tie into the back end data sources. Called different names by different vendors (Portlets, Gadgets, Blocks, Web Modules, Web Parts), many portal vendors have portlets for connecting to enterprise systems (such as SAP, Siebel, etc.) as well as for collaboration, news, and other functions.
Categorisation Engine - A categorisation engine is used for sorting documents into the folders of a taxonomy. The categorization engine may do this based on metadata in the documents, based on business rules, based on the content of the document, based on search criteria or filters, or some other scheme.
Filters - A filter is generally available in a taxonomy to restrict the documents that are admitted into a particular folder, or that are returned as part of a search. A filter can be word based (if a document has the word 'IBM"), concept based (if the document is like this other document), or rule based (if the field called CLIENT has a value of IBM).
Index - An index is a collection of information that allows for fast query and retrieval. Within the context of a portal, an Index is usually a combination of a full-text index and a meta-data repository for the documents/content that is included within the portal.
Virtual Card - Within an index or metadata repository, a virtual card is a description of a single document or piece of content within the portal. The card usually contains information about where the content physically resides, and contains the values of one or more metadata fields about that document. The card is the "placeholder" for the document within the portal.
Web Service - A web service is a program that accepts and responds to requests over the Internet. Typically, a web service accepts requests in an XML-based format. The actual format of the request and the response depends on the XML standards that are being used. One such standard is SOAP. There are public registries and languages - such as UDDI, WSDL - which are used to catalog the different available web services. A calling program can query the registry (UDDI) to find an appropriate web service, then use WSDL to figure out which parameters the service needs, and finally use a calling protocol and XML standard like SOAP to actually make the call to the Web Service.
Development Standards and Protocols - A very important component of any development project is to understand what the current industry standards are for developing Portal Solutions and how they relate to each other. A brief summary of the most common is provided below.
- XML - Extensible Markup Language. XML is a language used to represent almost any type of data. XML is similar to HTML (they are both descendents of SGML, a generalised markup language). Whereas HTML is used to tell Web browsers how to show information to the end user, XML is more typically used to send information between programs. The XML files usually do not have information about the display of the information - the program that receives the data, often by using an XSL style sheet and XSLT, usually handles this separately. The structure of an XML file is usually defined by its DTD (document type definition) or XSD (XML Schema Definition).
- XSL, XSLT - Acronyms for Extensible Stylesheet Language and Extensible Stylesheet Language Transformation. While XML documents contain data, XSL documents contain rules for "transforming that data" into a presentation that the user can understand. This presentation format might be HTML for web browsers or it might be WML for wireless devices or PDF for printing out the information.
- DTD and XSD - Document Type Definition and XML Schema Definition. Both DTD and XSD are ways to define the structure and layout of XML documents. DTD and XSD become important for validating that an XML document is in the right format for passing information between different systems, or for passing information from a back end system to the portal.
- WSDL - Acronym for Web Services Description Language. WSDL allows a Web Service to describe what actions it supports. A "stock quote" web service, for example, might have two actions that other programs can call - getStockQuote, which takes a ticker symbol and returns the closing stock price, and getTickerSymbol which takes a company name and returns one or more ticker symbols. WSDL is an XML based language that allows both calling programs and Web Services to describe legal ways to invoke the program. WSDL is important for portals because portals will typically aggregate information from multiple web services onto a single screen and so need to communicate with each one in the appropriate format.
- SOAP - Acronym for Simple Object Access Protocol. SOAP is an XML based standard for making function calls across the Internet to another application. SOAP provides the underlying calling protocol (which can be used as an alternative to HTTP GET/POST), a wrapper so that the calling application can send parameters to the program it is calling, and a method for getting results back from that program. Because SOAP is XML based, it is completely platform independent. SOAP is quickly becoming a leading protocol for invoking and getting results from Web Services.
- UDDI - Universal Description Discovery and Integration. UDDI refers to a specification for finding web services and a public registry where Web Services can publish information about themselves. UDDI can be used to get back XML based "descriptive information" about Web Services. This descriptive information might be in an XML format such as WSDL. UDDI has broad support from all segments of the Internet industry.
- WSUI - Web Services User Interface. WSUI is a specification for standardising the display of Web Services to end-users. This extends the traditional web services model, which is used to get and retrieve XML data, by providing a framework for how that data will be displayed to end users. WSUI is akin to a standard way to describe Portlets. In the WSUI model, a Portlet makes a call to a web service, gets back XML, and then users XSLT to transform that XML into HTML, which can then be displayed within the portal.
User Profiles - Each Portal contains a profile for each of its users. This profile is used for customisation and personalization. Each of the Portlets in a portal has access to this user profile and can use it to store preference information about a user or a class of users. This profile is also how the user "configures" the home page of a portal and chooses which Portlets show up and what information they should show.
Content Management System - Most enterprise portals contain a Content Management System, which allows approved end users to submit information into the portal. There is typically an approval process that eventually results in the content becoming available in the correct part of the portal's taxonomy. A Content Management System can deal with documents in their original formats (Microsoft Word, PDF, etc.) or might contain Web Editing features to allow end users to author web pages. (Note: this definition than the one in the prior section)
EAI (Enterprise Application Integration) - EAI serves as the umbrella term for all software and services meant to integrate enterprise applications with one another. Given the complexities of each type of application (sales, manufacturing, service, HR, purchasing, etc.) this can be a difficult and expensive proposition. A number of vendors have released software that makes this effort much simpler - including Crossworlds, WebMethods, Tibco, NEON, and MQ Series, etc. EAI impacts the portal because the portal ideally will show consolidated information from multiple back end systems. An EAI layer is needed so that the queries can be coordinated and the results consolidated.

Back to Index

Want to find out more?