DBMS UPDATE<@VM>Five vendors offer free database app trials
Multifaceted systems and powerful tools help meet the growing demands on government databases
By Edmund X. DeJesus
Special to GCN
Government database users face many new and difficult tasks these days.
They need to make data more accessible for analysis, even if it resides on nearly impregnable legacy systems.
They have to deliver interactive Web access to their data, even though there are colossal amounts of it.
They want to store nontraditional data types, including audio files and images that traditional databases can't handle.
And they want to get that data out to the people who use it, despite a dizzying profusion of destinations.
Fortunately, the latest heavy-duty databases have the ability to handle all those demands and more. While working to satisfy commercial customers, vendors have had to stretch the capabilities of database management systems. Commercial customers have been working on those databases for several years now, sparing government users the role of pioneer.
Vendors now can provide databases that satisfy stringent requirements for a variety of government needs. In addition, tools for creating and maintaining databases, and for analyzing data, have been tested in the same fire and are available.
Let's look at some of the new tasks you may be facing, and the options you have with today's databases.Data warehousing.
Government and commercial users share the concept of data as a working asset, not a piece of furniture. You want to use existing data for a variety of purposes, including decision-making, planning and analysis. One option is data warehousing. Although this would be a real chore with old databases, most modern databases include viable methods without your having to invent them from scratch.
Look for databases that can integrate data from many sources. 'Rewriting legacy apps would be prohibitively expensive, yet you want access to that data,' said Paul Kaplan, business unit executive for IBM Corp.'s federal division. He suggested looking for databases that could simplify data cleansing and populate the database automatically.
'Some government agencies got burned by pioneering data warehousing projects. It's safer now,' said Jon Wall, senior technology specialist for Microsoft federal systems.
|Tips for buyers|
'Integration: What must this database work with'other databases, tools, output formats'now and in the future?
'Data: What kinds of nontraditional data does this project require? How will that affect the size of the database?
'Analysis: What data do you want to access and what kinds of analyses will you perform?
'Web: How can this database help you with Web-oriented access tasks?
'Users: Considering your users, what steps should you take toward access methods and security?
You also want to be able to publish this data in whatever form your users want, including as spreadsheets or graphs.Relational technology.
Government databases are evolving to relational technology, a process that simplifies many tasks and the choices among support and analysis tools.
Carefully consider not only the database itself, but also vendor mechanisms for moving vast amounts of data from flat files and tape storage into relational structures. Vendors have had to face similar challenges from large commercial customers.Standards.
Investigate database support for other standards. Structured Query Language, of course, is the lingua franca of database queries. Open Database Connectivity is important, especially when working with Microsoft Corp. operating systems or databases'which is often, since you may well be shipping data to users' desktop PCs.
Joint Database Connectivity and SQL Java need your attention if your database is supporting Web-based access.Web access.
Establishing Web access to mainframe legacy data can be a Sisyphean task. Most systems use a modern database as a front end to stage data from the legacy system and provide simpler access via the Web. For government agencies, which must quantify and justify everything, the ability of a database to also monitor Web access'by hits, clicks, volume of data accessed or some other metric'is a major plus.
Extensible Markup Language.
|Check out this sampling of database vendors and products|
Angara Database Systems Inc.
Palo Alto, Calif.
Angara Data Server is a main-memory database that runs on multiprocessor servers under SunSoft Solaris, Microsoft Windows 9x and NT, HP-UX, AIX and Linux.
Computer Associates International Inc.
Jasmine TND uses neural network technology to supply decision support by monitoring and predicting patterns.
Monarch Data Pump utility provides a quick way to move, transform and propagate data.
Embarcadero Technologies Inc.
DBArtisan provides database administration for Sybase, Oracle, Microsoft SQL Server and IBM DB2, allowing management of multiple databases from a single console.
Santa Clara, Calif.
FileMaker Pro 5.0, for Win9x and NT, provides an easy way to build a database and put it on the Web.
DB2 Everywhere for Palmtops is a 50K database for Microsoft Windows CE and Palm OS, available as a free download from IBM's Web site.
Informix Software Inc.
Menlo Park, Calif.
Cloudscape 2.0 is a Java database designed for mobile or downloadable applications.
Red Brick Decision Server is optimized for Internet-enabled analytic applications.
Informix-SE is a Structured Query Language database engine for small to midsize applications, working with Unix database servers.
Microsoft Data Engine is a subset of code from the company's SQL Server that enhances Microsoft Access as a database for mobile use.
Magic Software Enterprise Inc.
Magic 8.20 for Linux is a tool for creating both traditional and Web applications.
Redwood Shores, Calif.
Oracle8i Lite for portable and embedded systems is Java software, requiring 50K to 750K of memory, capable of replicating data and Java applications from a database.
Pervasive Software Inc.
Pervasive.SQL 2000 comes in three versions: a small, 8K database for Java smart cards, an embedded systems database of about 50K for Internet appliances and other devices, and a database with an engine using less than 400K of memory for mobile devices using Windows CE and Palm OS.
San Mateo, Calif.
PointBase is a Java database in server and mobile editions that manages network data, targeting multiplatform capabilities for electronic commerce applications.
Quest Software Inc.
Newport Beach, Calif.
Data Manager is designed to improve project release schedules using graphical user interface wizards to extract data to create small test databases.
Sun Microsystems Inc.
Palo Alto, Calif.
Java Database Connectivity 3.0 allows presentation of database responses as Extensible Markup Language documents and translates XML documents into Java to interact with relational data.
Sybase SQL Anywhere Studio is a low-cost system for running a mobile database on remote systems, working with Win9x, NT and CE, and Linux.
TimesTen Performance Software Inc.
Mountain View, Calif.
TimesTen 3.5 is an in-memory database that provides a performance boost for voice over IP and other data and networking functions.
San Jose, Calif.
Web-Now DBMS utility pulls in data from a variety of databases, including Oracle, Informix, Sybase, IBM DB2 and Microsoft SQL Server, for Web use.
Log Analyzer converts Internet server logs to reports on Web activity, with graphics. For Win 9x and NT.
WhiteLight Systems Inc.
Palo Alto, Calif.
Analytic Application Server helps build applications that automate the decision-making process.
Part of the appeal of XML'the next Java, some analysts are calling it'is the ability to tag data from different sources to keep apples with apples. This is one example of the use of metadata: cross-database information that makes sure everyone is referring to the same entity.
For many managers, the most attractive use of XML is as a transport technology. 'XML can make it easier to exchange data between databases,' said Karen Watterson, a consultant and principal of the Watterson Database Group of San Diego. Using XML, you can integrate data from different sources and stage it for Web access using simple protocols such as the Hypertext Transfer Protocol.Analysis.
If analysis is your goal, consider some special-purpose database features. For example, using data mining and online analytical processing to find important patterns in data are high on many users' lists, especially users working with health care and financial data.
It is both possible and desirable to integrate historical data for trend analysis with current data for pattern analysis. With year 2000 work successfully completed, 'demands for data mining may well swoop in to take up the extra resources,' Microsoft's Wall said.
Commercial users find it beneficial to stage their data in nontraditional hypercube structures that facilitate slicing-and-dicing and rotation manipulations. To maintain this parallel-processing architecture separately from your transaction database, you should quiz vendors about their ability to populate hypercubes automatically.
Some vendors claim their products can support both analysis and ordinary use. But get proof.Integration.
You know how it is: A did project X with database Q, B did project Y with database R, and lucky you have to assimilate the whole alphabet into one soup. In the past, you would have chosen one database, then laboriously ported in everything.
In many cases, this would be impossible, either because of the sheer amount of data or the outrageous cost of re-engineering legacy databases.
Commercial merger mania has created a demand for quickly integrating large and incompatible databases. Look for databases that can communicate with other databases, including their competitors. Then use that single database to import information from those sources.
In fact, some databases can grab the data that you need in real time'even on the fly from transaction systems'and present it for your use as if it were stored permanently in that database. This technology bridges the separate data stores.Nontraditional data.
Government agencies do not live by text and numbers alone. Geographical map data is a prime example. You may also have images'documents, pictures from satellites and so forth'about which a traditional database is clueless.
Newer databases are more adept at handling such nontraditional data, with the capabilities to handle sound, images and video. If you have special requirements, make sure the database has extensibility that lets you define your own data types.
Accessing nontraditional data is something to take into account also, using, for example, IBM's Query by Image Content technology.Storage.
If you are storing images, sound or video, your storage requirements are about to sharply increase. Even more ordinary database goals may produce an avalanche of data.
Look for size-scalability features in your database, to be able to increase the size of the database without affecting performance. Ask vendors for examples of scaling in their products.
'You can even prototype your system with artificially constructed data sets to test how well they'll work in your situation,' IBM database specialist Hunter Ambach said.Scalability.
Another aspect of scaling involves multiple processors or multiple machines. Your goal is to make access transparent to users, so databases perform heterogeneous queries across multiple sources. Some vendors claim their database works on everything from a palmtop computer to clustered Unix boxes. Check out those claims.English-language queries.
As more users require access to data, it is important to simplify the means of access. 'SQL expertise should not be a requirement,' Wall said.Mobile users.
Many agencies must support field personnel who require local data access. Investigate a system's ability to copy subsets of data to roaming machines, then accept modifications back into the database.Tools.
Make your life easier: Look for tools to accomplish certain tasks, either from the database vendor or a third-party vendor.
Tools from your database vendor are meant for use with the database and have obvious performance benefits you cannot ignore. The downside is that these tools probably only work with a single database. Third-party tools work with many databases and often excel at their particular task, providing more functionality.
Tools can simplify the creation and maintenance of databases. 'You often have to reverse-engineer legacy databases that lack documentation,' said Fred Pierce, a database administrator subcontractor at NASA's Goddard Space Flight Center in Greenbelt, Md. He uses Embarcadero Technologies Inc. tools for rapid development.
Tools from companies such as Embarcadero Technologies can help you rapidly deploy database applications and manage database operation.
They can gather information on a range of database factors, such as disk usage, in a proactive rather than reactive way'that is, 'You're okay now, but start thinking about adding more storage,' rather than 'You just blew your last disk, hope you made backups.'
Tools also ease report creation and analysis. Companies such as Cognos Corp. of Burlington, Mass., specialize in the analysis and reporting of data. Their products usually are more sophisticated than what you might find from a database vendor.Long-term planning.
Thirty years ago, no one thought that some systems would still be running today. 'Long-term viability is an important consideration for commercial customers and should be for government customers,' Watterson said.
Choose technologies'and vendors'that will last. You want to go with vendors who will be around for a while because your database almost certainly will.Edmund X. DeJesus of Norwood, Mass., writes about information technology.
|DB2 6.1||NT, AIX, Linux, Solaris, Win9x (Personal Edition only), OS/2, Unixware 7, HP-UX, OS/390, VSE and VM, AS/400||60-day trial: www6.software. ibm.com/dl/db2udbcd/db2udbcd-p; Personal Developer's Edition: www6.software.ibm.com/dl/ db2pde/db2pde-p||$2,000 up|
Menlo Park, Calif.
|Dynamic Server 2000||NT, Solaris, HP-UX, SGI Irix, AIX, Linux||Trial download: www.intraware .com/trialsoftware||$1,500 per seat|
|SQL Server 7.0||Win9x, NT||120-day trial: www.microsoft.com/ midatlantic/offers/sqloffer.htm||$1,400 up (five users)|
Redwood Shores, Calif.
|Oracle8i Release 2||NT, Solaris, AIX, SCO, Unix, Linux, HP-UX, Digital Unix||Trial downloads: technet.oracle.com/ software/products/oracle8i/ software_index.htm (password required)||$4,000 up (five users); $100 per power unit|
|Adaptive Server Enterprise 12.0||NT, AIX, Solaris, HP-UX, Win9x (client only), Compaq Unix||Tools: softwarexpress.sybase.com/ esd2/esd.stm?sname=softwarexpress. sybase.com (password required)||$1,200 per seat|