The Buzz: "Enterprise Information Portal".

Want soundproofing? Read on.

One of the loudest buzz phrases in the computer industry today is "Enterprise Information Portal" (EIP). Much of the buzz is based on a study by Merrill Lynch, which famously predicted a $14.8 billion market for EIPs by 2002 . . . without any primary market research at all behind it. The study simply took a number of different technologies such as business intelligence, enterprise resource planning, document management, and the like and squashed them together into an amorphous blob labeled ETP.

Now, everyone, it seems, is talking about EIPs. Some fifty or sixty vendors have planted their flag in this new world, some simply by re-labeling existing business intelligence or document management products with the word portal

somewhere in the new name or marketing description. So, an emerging market slouches toward any number of "portal" conferences, leaving a trail of FUD behind it.

To be perfectly fair to the analysts at Merrill Lynch, what they did was analogous to what an EIP is supposed to do, give users a unified view of all corporate knowledge assets using the new universal interface, the web browser. There is little doubt that this represents a real trend in corporate computing: attempting to combine through a single, easily-mastered interface various information systems and resources such as databases, document management systems, data warehouses, imaging systems, business intelligence, and ERP and CRM applications. The goal is to empower knowledge workers with the information they need to do their job effectively, enabling them to share information and to get information that they would not otherwise know about.

Yet in the rush to capitalize on this new trend, there's the danger of overlooking some of the alternatives to an enterprise information portal--other ways of exploiting the simplicity of the web browser interface. These may be way stations on the road to a full EIP or they may, in themselves, be the 80/20 solution to the problems an EIP supposedly addresses--e.g., they may get you 80% of the way there for only 20% of the cost of a full EIP. So I'm going to take a look at the offerings of two vendors with an interesting approach to "portals" to illustrate the other paths to information sharing available.

Look, Ma, No Middleware

Some enterprise information portals are based on a complex middleware layer that takes care of gathering information from the various data sources, reformatting it for the front end interface, handling updates, and so forth. This is especially common in portals that purport to handle all kinds of disparate information: structured (e.g., transactional and RDBMS data), unstructured (e.g., text files, email, images, etc.), and commercial (e.g., newsfeeds and the like). While an EIP may need to grow to handle all this, the folks at BroadQuest Inc. (San Jose, CA, www.broadquest.com) point out that, right now, the main need for a portal today lies with companies that need quick access to the structured data that is the mainstay of business activity. Of particular interest to many companies, as our research indicates, is Customer Relationship Management (CRM), which aims to take the transactional data resulting from day-to-day interactions with customers and turn it into knowledge that can be acted upon to increase profits by better satisfying customers.

This is precisely the area on which BroadQuest is focusing. The first implementation of their technology is called myEnterprise: Sales, which provides account managers with a single view of live customer information from multiple back-end systems. The company points out that about 20% of an account manager's day is spent tracking down customer information--where's that order, when was it shipped, why hasn't this trouble ticket been taken care of--and that 70% of all transaction systems contain customer data. The myEnterprise: Sales application connects account managers to that information.

The architecture of the BroadQuest EIP is remarkably simple. It consists of three parts. The BroadQuest Broker, based on the Common Object Request Broker Architecture (CORBA), a standard supported by hundreds of vendors, manages updates from connected databases on a schedule determined by the system administrator. The Broker gains access to these databases using BroadQuest "Wrappers," a non-invasive connector object that can be written using a wrapper API. Support is included out of the box for databases such as Oracle, Sybase, Lotus Notes, and even flat files.

Key to the rapid response the system is capable of is a remarkable architecture that I like to call a Redundant Array of Inexpensive Computers (RAIC). Analogous to a RAID array of disks, the BroadQuest RAIC, which is where the Broker lives, maintains all the data gathered from the connected databases in main memory, yielding real-time access to huge data sets. The system runs on Windows NT, among other platforms, giving users the benefit of the price/performance advantage of Wintel systems. According to BroadQuest, implementations of the system at Novell and Wells Fargo yield sub-three second access to over 30 million records for thousands of users.

The BroadQuest Portal is a secure and scalable CORBA service that gives users access via their desktop browser to an interface featuring Navigation, Summaries, Alerts, and Reports. The user and the administrator can customize this interface. It includes predefined applets for business performance charts, information access tools such as search and a context driller, and a workflow feature. The BroadQuest Administrator is the control panel for the application, enabling not only highly technical operations such as performance tuning, but also control of users and groups by non-technical users.

By the time you read this, BroadQuest will have launched myEnterprise 3.0, along with an application development environment called myToolkit, which will enable companies to build their own portals by creating new applets for the user interface. The company intends to support other forms of data, unstructured and commercial, through partnerships with leading vendors in these fields. The system is not cheap. BroadQuest's Rapid Success program costs $99,000, but it promises a working system in 30 days: connection to one database for 50 users with all development training included. The company is looking to partner with integrators and their focused business model makes it likely they'll be around for the long haul. So integrators with the expertise to apply this technology should definitely be in touch with BroadQuest.

Abandon Sloth, All You Who Enter Here

ThoughtStar Inc. (Sandy, UT, www.thoughtstar.com) is at the opposite end of the portal spectrum. Their product, QuickTeam 99, is a Java-based collaborative application designed for small teams, although the company claims it will handle up to 2000 users. This is an aspect of portal design that many companies, if not the majority, have overlooked: enabling and enhancing collaboration between ad hoc groups of users.

Although QuickTeam is available as a free, company-hosted application on the web, the real power lies in the full version, which can be run on a company's own server or hosted by the company's ISP--thus tying ThoughtStar into another burgeoning trend, the Application Service Provider market. Yet that's another column.

QuickTeam 99 consists of a bundled web server and database that enables dispersed teams of workers to collaborate using their desk-top browsers. The user inter-face, a Java applet, gives users access to features like threaded discussions, chat, paging and a whiteboard, voting and surveys, team news and events, contact management, calendaring, project scheduling, task management, Gantt charts, and more. It's easily customized, both by an administrator and by the users, allowing team members to quickly construct a shared workspace that meets their needs and one that can quickly adapt as those needs change. It can also work with existing groupware products, so it enables adopters to leverage existing IT resources.

QuickTeam isn't perfect--one hole in particular is the lack of any integration with email programs, forcing users to cut and paste. This is likely to be especially annoying, given that the program seems to do so many other things so well. Yet it's a good first step, an example of a capability that's needed in any portal, and worth a look by any integrator looking to cash in on the portal craze at the low-to-middle end of the market.

Dave Trowbridge is a senior analyst at Survey.com, a market research firm specializing in data-intensive reports on information technologies, where he monitors operating systems, portals, and various aspects of business intelligence.