Portals

Digg This!

Enterprise portals are fast becoming the foundation of the Web-based economy thanks to their ability to give enterprises, trading partners, and customers global access to enterprise applications, back-office systems, and IT infrastructures. This ability has made enterprise portals appealing as the infrastructure of choice for enterprise IT organizations and has helped organizations justify the considerable expense of migrating from legacy systems.

However, what Web-based systems, such as portals, gain in global access and economy, they can sacrifice in stability and manageability compared to traditional closed, point-to-point systems. Portals are subject to the availability and performance problems that come with reliance on a Web-based infrastructure that is publicly shared, constantly changing, and ever growing.

This article discusses how performance management systems can be used to enable the development of critical, next-generation portals that are reliable and bulletproof.

Portals: Providing a Multi-Functional, Unified Structure
Despite the popularity of Web-based systems, Web technology is still in its infancy. The Web provides global desktop access to the corporate network, but without some organizing principle, this access can be merely an amorphous view of legacy systems and applications that have been hastily ported for Web delivery to a desktop already cluttered with any of a million other offerings available on the Web. By providing core business functionalities in a single integrated solution, enterprise portals such as BEA WebLogic Portal supply the organizing principle and technology needed to make sense of the desktop view of enterprise Web-based networks.

Companies typically deploy two types of portal solutions: external portals and internal portals. External portals provide organized access to a company?s products and services. Internal portals integrate the internal systems, applications, and information services of an enterprise by providing a common desktop interface and the ability to customize the interface to suit a particular department or role.

All portals require certain core functionalities such as a Model/View/Controller (MVC) paradigm and flexible controller implementation, user and group management, entitlements, personalization, security, and content management. Some portals also provide extendable rules engines, shopping carts, pricing systems, and targeted content engines.

Sophisticated, large-scale portal applications built with BEA WebLogic Portal provide customized access to content by user/group using complex, dynamic business logic and also maintain persistent session/transaction information, permissions/entitlements, and context. For example, to support a shopping cart or airline reservation, the portal must provide access to the appropriate internal databases and applications for the duration of an extended transaction. When the user clicks "submit" in one pane/tab, various options must be updated in other panes/tabs and customized based on the information submitted. The portal maintains this complicated context for user sessions and transactions. High-performing portals provide all the elements above to rapidly and reliably deliver useful information and services; enabling financial transactions, tracking internal processes, or allowing access to information systems and internal applications on a customizable basis.

Managing Portal Complexity
Increasingly, vendors such as BEA are providing "out-of-the-box" portal functionality that enables development teams to put together much more complicated and higher-functioning applications on the same schedule and budget previously allocated for simpler applications. By providing a common integrated application framework, portals make modular development possible (e.g., separate development groups might work on separate "tabs" of an application). But as the underlying IT system becomes more complex, managing, identifying, and controlling system failures can become more difficult. Portal-based applications usually run in the same environment as other Web applications and are given to the same teams to manage. However, when performance issues arise, the sheer number of components and transactions can overwhelm the IT staff members who must deploy and manage portal applications, fine tune them for performance, monitor the entire system for emerging trouble, diagnose problems, and rapidly solve them with a minimum of downtime.

Typically, these problems do not surface for the portal administrator until user complaints have come in; that is, after the problem has become severe enough to impair portal functionality. When problems do occur, their source is not always obvious. It may be in a portlet, a database, or any of the supporting back-end systems. Isolating and resolving problems can become a mind-numbing procedure of rebuilding and analyzing every portlet in its own "test" window after the failure has occurred. Portal administrators need portal management tools that let them "look under the hood" at internal portal processes and their interactions. Without such tools, isolating performance problems will become increasingly difficult.

To help them maintain high performance and 24x7 availability, portal administrators need effective management tools built specifically for monitoring not only individual portal components and workflow processes, but also the environment surrounding the portal. This includes connections to databases, transaction servers, mainframe systems, and other back-end systems. The tool portal administrators choose should:

• Monitor the complex, dynamic interactions taking place across the entire workflow and within individual processes
• Present the resulting data in a clear, simple display that highlights problems (and where they occur in the portal workflow) and allows the administrator to drill quickly down - to individual portlets and transactions, if need be - to the source of the problem
• Summarize overall performance as well as performance in the key portal workflow areas: portal servlet, control tree processing, JSP backing files, Java Pageflows, portlets, connections to back-end systems, and portal services

BEA WebLogic Portal: How It Works
BEA WebLogic Portal is one of the industry?s most comprehensive portal offerings and many enterprises are using it to develop and deploy critical portal applications. It contains a wide range of technologies for developing and maintaining sophisticated business-to-customer (B2C), business-to-business (B2B), and business-to-employee (B2E) portals. To better understand the requirements for monitoring WebLogic Portal, it will be helpful to discuss in general terms how it works.

As shown in Figure 1, the Portal Servlet Manager (1) fields the incoming user request, which in turn initiates Control Tree processing (2). The Control Tree represents all the structural elements in the portal. It also serves as the foundation for building a new portal page. During Control Tree processing, a new Control Tree is created (or drawn from cache if it already exists). The Control Tree is built throughout its life-cycle phases, coordinated by the Lifecycle Manager and the Control Tree Walker.

As a Control Tree is built, user Entitlements (3) are verified to determine which portal resources are available to the user. During the pre-render and postback data operations (the two primary operations during the life-cycle), the Presentation Context (4) is called. In the meantime, JSP backing files are also checked for custom code written by the development team. The Presentation Context, in turn, calls Java Page Flows and JavaServer Pages (JSPs; 5). The Control Tree then pulls data from individual portlets as necessary, depending on the type of user request. Portal services (6) such as personalization, content management, and user profiles are then initiated as needed. The final page is built and rendered as the processes described here are completed.

Resolving Complexity with Simplicity
In the workflow described previously, several areas are potential sites for problems that can affect portal performance and availability. An effective portal management system can help developers and portal managers identify the exact location and root cause of errors regardless of where in the environment the error occurs. When portal management systems are used throughout the portal lifecycle, developers can use the data to proactively monitor and manage session and transaction information, permissions, entitlements, and context in production, QA, and development environments. The result is a more stable and reliable portal. (See Table 1 for a list of the important BEA WebLogic Portal components and processes that administrators should monitor.)

Using customizable management UIs, administrators can rapidly assess the performance and availability problems described previously and immediately take corrective action. For example, by displaying key performance data in easy-to-use dashboards, portal managers can immediately view overall portal performance, as well as performance in key workflow areas. Traffic lights, time plots, and worst performing graphs plainly show administrators where to focus their troubleshooting efforts. By simply clicking on a tree view, administrators can drill down to the root cause, precisely isolating performance issues in individual components.

Administrators can use a portal management tool to react to problems more quickly. However, businesses have shown an increasing desire to use management tools in a more proactive manner. For example, a management tool can be used to anticipate problems by establishing baseline settings and border conditions for portal performance. Using this information, alarm thresholds can be assigned to target specific workflow processes so they can be proactively monitored (see Figure 2). Not only can administrators improve portal performance by monitoring and managing portals throughout the application lifecycle, but performance information can be used over the long term for trend analysis and capacity planning.

Conclusions
Enterprise portal solutions such as WebLogic Portal enable enterprises to successfully deliver Web-based IT services to customers, partners, and employees through a single, unified interface. However, as portals become more complex, so does the challenge of managing portal performance and availability. By adopting a portal management solution built specifically for monitoring and diagnosing portal performance, businesses can accelerate the resolution of portal-related errors and proactively deploy next-generation portals that deliver superior service and functionality.