Stay Alert

This otherwise unembellished data monitoring and alert product is friendly to business users.

By Ganesh Variar      

The evolution of enterprise computing from mainframes to client/server architecture to, most recently, the Web has caused business data to spread across various data sources. Mergers and acquisitions have also contributed to the increasing number of data sources. Business users need to access these myriad sources and monitor the state of the data, which is vital for realtime decision-making in the Web-enabled economy.

Metagon Technologies' DQagent 1.0 is a tool that monitors enterprise systems for user-specified events, which in turn trigger user-specified tasks. DQagent works in conjunction with Metagon's other major product, DQbroker, which can connect to various data sources and make them appear to users as a single data source.

DQagent lets users create agents that use DQbroker to connect to one or more data sources and monitor the data for specific conditions. When these conditions are met, the agent can initiate several responses, such as notifying the user with a pop-up message or email, running an application or another agent, and printing, saving, or browsing the result set.

Business users can create their own agents without relying on the IT department too much, as DQagent's desktop user interface is simple and intuitive. Agents can be scheduled to run periodically and can automate routine decision-making. For example, executives may create agents that monitor gross margin percentage daily and emails them when the percentage falls below or rises above a certain threshold value. In this example, the gross margin percentage may be computed based on costing and sales revenue figures that reside in two different databases on disparate platforms.

The Competition

Metagon has been active in the data integration market for more than five years with the DQpowersuite set of tools. This package centers on DQbroker, which is a well-established tool. DQagent is the latest addition to Metagon's tool set, and it uses the power of DQbroker to connect to most common data sources.

The database monitoring market consists of tools as diverse as simple alerting systems in ERP packages to sophisticated neural network-based agents like Computer Associates' Neugents and Searchspace's Sentinels for the Intelligent Enterprise Framework, which use artificial intelligence. Other products in this space include WorkWise Inc.'s Data Agent Server (DAS), which can monitor as well as modify the data source, and Categoric Software Corp.'s Xalerts, which has a more sophisticated application program interface and support for mobile messaging like WAP. (See Stewart McKie's review of WorkWise DAS, "Easier Database Monitoring," in the June 5, 2000, issue.) DQagent's strength lies in its simple interface, which makes it intuitive for business users, and in its ability to access diverse data sources and present a single logical view of the enterprise data.

Creating an Agent

You install DQbroker on the server where it runs as an NT service or Unix daemon. DQadmin is a graphical user interface for administering DQbroker, and you can install it on the administrator's machine. DQagent is typically installed on the business user's desktop. Installation is very straightforward, and each product takes less than five minutes to install.

To test the product, I gave myself the task of accessing sales revenue tables from two different databases (Oracle and Microsoft SQL Server). One contained the revenue from Web-based sales; the other, revenue from the more traditional brick- and-mortar channel. I used DQadmin to connect DQbroker to the two data sources. DQbroker uses SQL*Net for a native connection to Oracle and an ODBC connection to SQL Server. I created a catalog and assigned the two data sources to the catalog. Finally, I created a user and assigned permissions to the catalog for the user. Setting up the data source interfaces is simple enough for a junior IT person to do. DQbroker controls only the data source-level access, leaving the detailed data-level security to the database itself.

I created an agent that would monitor the two data sources and notify me when the revenue generated by a product through the Web channel exceeded the revenue the product brought in through store sales. As Figure 2 shows, the DQagent editor dialog box has four frames: Query, Action, Schedule, and Informational. The Create Query button in the Query frame opens up a Query Builder. It lists all the tables available to the user along with their data source names and lets the user interactively create a complex query that can join tables from multiple data sources and filter, group, and sort the results.

Business users should feel fairly comfortable with the Query Builder. The Display Data button runs the query and returns the results in a browser. The Action frame lets the user define one or more tasks that DQagent needs to execute after completing a query. A query may produce some data, produce no data, or fail. The user can define one or more tasks for each result. The Schedule button opens up a dialog box that lets the user schedule the agent to run at a specified time or interval. The Informational frame lists the steps the user must perform when creating an agent.

The Bottom Line

DQagent is an easy-to-use product that achieves the strictly defined goal of monitoring enterprise data. But you can get further value from its underlying DQbroker product, which lets any other ODBC-complaint client application connect to it (using the DQODBC add-on that ships with DQbroker): The client application gains a single logical view of enterprisewide data sources. However, in today's e-business environment, DQagent would be more valuable if it could access HTML and XML data sources and provide a Web-based end-user interface.

DQagent is for you if you have multiple data sources and need a simple monitoring tool that can provide an integrated view of the data to your business users.

Ganesh Variar is a consultant with Saama Technologies, a Silicon Valley-based consulting firm. He has seven years' experience in managing and designing business intelligence solutions.