Enterprise Problem Solver

As seen in the December 2000 issue of Showcase

Technology Insight:
Enterprise Problem Solving in the Internet Era

by Simon Murrell

These days, IT problems can come from anywhere. Applications are being continually improved, integrated, and extended. In the Internet era everything is so interconnected that when one application fails, the whole enterprise can come tumbling down, meaning lost business and a lot of work for you.

Effective problem solving is now more important than ever before. When problems do occur, you need to know about them immediately. You need technology that allows you to collect all the information needed to diagnose problems, you need to be able to solve them...and then prevent them from happening again.

Technological advances allow AS/400 technicians new and valuable capabilities to monitor, diagnose, debug, and solve problems on the fly, locally, or over the Internet. By detecting and solving problems as they happen, mission-critical applications can be kept running smoothly around the clock.

Monitor Error Messages

>It was never possible to monitor all interactive jobs on your system for error messages. You would have to make changes to each program, which means anticipating errors. At best you could be alerted when errors occur in applications being tested, but errors in production could easily go unnoticed, corrupting databases and in turn causing "cascade" errors through your system. Many companies rely on non-IT-literate end users to report problems. Sometimes those users don't even report problems, and when they do, the information they give is rarely useful. Solving a software problem with little clues to indicate the cause can be very time-consuming and unproductive.

Now your whole AS/400 network can be monitored for all critical messages without the need to make any changes to your programs. This means that you can monitor all applications, new and old, for any error message with minimal CPU use. Even interactive error messages can be monitored, so you no longer need to rely on end users. Programmers get all the information that they need to solve problems immediately so they don't need to waste time deciphering end-user reports, and that means that they can solve the problem faster and move on to the next task.

Full message information can be presented graphically on any PC, sent by email or by pager, or integrated with other system management packages such as Tivoli, CA-Unicenter, etc. For example, you could configure a Simple Network Management Protocol (SNMP) trap to be sent in response to certain AS/400-related error messages, improving your current system management suite by providing more accurate AS/400 information.

Now, with effective application error monitoring, any error can be caught by filters and sent to the people who need to know immediately, before it becomes critical. If the person experiencing the problem is attending a customer then you can give them instant remote support — before the customer heads off in the direction of your competition.

Historical error message reports can now be used to isolate problem areas and audit applications. Once you have discovered which programs or applications cause most frequent problems or user confusion you can take steps to correct the situation. Targeting your efforts on those areas that cause most problems is the most efficient use of your resources.

Systems operations managers using Tango/04's Enterprise Problem Solver have reported how the product has uncovered recurring errors that they never knew existed. Users had not bothered to mention that erroneous data was being produced, creating inaccuracies further down the line. By collecting and filtering information on all messages that occur, you can effectively audit your applications and ensure that nothing is slipping through the net.

Collect Error Information Instantly

With your resources pushed to the limit, programmers can waste a lot of valuable time researching errors and recreating them in a testing environment. And nobody likes examining job logs and program dumps to find out what went wrong after a job crashes.

The best solution is to instantly capture job information as errors occur. Enterprise Problem Solver uses technology that reveals all live internal job information, including the end user's screen (for interactive jobs), the job log, the local data area (LDA), QTEMP, the call stack, and more.

It can even get down to the source code level while the supported job is still running. Technicians can start a full graphical debug session, taking control of the program execution, stepping through the program, and seeing and changing variable values. When problems occur in production, debugging may not be possible because source code is often not available on production systems. This is not a problem anymore — programmers can now map source code from any remote system, so they can debug and identify problem source areas on the fly, saving more valuable time.

Applications are constantly changing so a common source of confusion is the lack of fresh or clear documentation. An advanced GUI can be used to display graphical trees of your programs, data, and application structure, which can be refreshed in real time. Technicians can browse source members, program objects, and applications just by clicking on visual branches or hyperlinks.

It is often easy to see how to solve a problem — if only you could take control of that job. That used to be out of the question. However, now you can execute any command on any running job, even batch jobs. Customers have been able to save critical batch processes, simply by editing the LDA while the job was running, or by adding a library to the library list. There is no need to cancel the job so no work gets lost, and in many cases disasters can be avoided.

Remote Operations

Technicians need to be able to connect to your AS/400 when they are at home or traveling, especially when problems occur. But what can you do if you don't want to put your AS/400 on the Internet for security reasons? Enterprise Problem Solver's Internet Security Extension allows AS/400s that are not connected to the Internet to be remotely serviced in a safe and efficient manner, with no risk of intruders.

Effective Testing

Prevention is better than cure. Testing your applications effectively before putting them into production is fundamental to project success. Powerful debugging tools improve programmer productivity and help you to ensure that no bugs reach production.

Enterprise Problem Solver includes a powerful graphical debugger, giving programmers a wide range of functions to help find bugs.

Reduce Costs, Increase Availability

Time spent solving problems can be reduced because Enterprise Problem Solver lets you know immediately when problems occur. Programmers can concentrate on doing what they do best — developing great software.

Simon Murrell is the former director of business development at Tango/04 Computing Group.