Midrange Stuff, OS/400 Edition--Back-Office Automation Pays Off for Newell Rubbermaid

The success of a product is determined by its ability to deliver on a promise every single day. The longevity of Newell Rubbermaid's brand legacy is testimony to this principle, and has helped the company to achieve global recognition for a vast number of successful consumer products under the names Paper Mate, Parker, Waterman, Rotring, Graco, LittleTikes, Pyrex, Swish, and Rubbermaid. Behind those trusted names, Newell Rubbermaid's business relies on the smooth operation of its OS/400 environment.

The daily promise of high availability is extended to a fast-growing user network across Europe from Newell Rubbermaid's data center in Sunderland, United Kingdom, via four critical iSeries machines running on a single logical-partitioned iSeries Model 890 server, which runs all the key applications, including Intentia's Movex ERP system, Manhattan Associates' PkMS warehouse management application, SAA Consultants's REIMS EDI system, and payroll. Helping the data center team to manage this busy iSeries environment is a trio of real-world systems management solutions from CCSS, including QSystem Monitor, QMessage Monitor, and QRemote Control.

A recent move of the Data Centre to its present location gave the team an opportunity to invoke a new culture of management that would free their time, allow them to be more proactive, and add value to their sector of the business. An impressive bank of huge screens, at the front of the data center, now project graphic representations of their centralized management of all system-performance parameters and messages to all personnel.

Rob York, Newell Rubbermaid's European data-center manager, explains the shift this has brought about in the way that the team works and how they are perceived by the company. "The screens have helped to eradicate a defensive approach to problem solving," he says. "In the past, if a call came in, the first 20 minutes would be spent by individuals finding evidence to justify why it wasn't their particular problem. Now we can cut straight to the point. A problem occurs, and we all just have to look to the screen; there’s no time wasted, and people are more forthcoming in offering their help to resolve it. The users also now have a different perception of what we do. It's very obvious when something is going wrong. Problems have no place to hide, so users are aware of how visible their needs are to us and how responsive we can be."

QSystem Monitor tackles performance issues through a real-time online monitor and helps the team to quickly identify problems before they have an impact on user activity. A recent incident with a couple of runaway jobs illustrates the efficiency with which such typical problems can be resolved.

The jobs in question had brought the machine to a virtual halt, and at the same time, the data center screen was showing that there was a problem with CPU and response time. Without the phone ringing with users telling them of the situation, the team was able to pinpoint the problems and to correct the situation. The total impact on the users was reduced to around five minutes, and the team was able to remain proactive and to prove the new approach could add value to the business by eliminating problems before they became a threat to availability.

"Finding the root cause of problems is now so much easier," says Glen Foster, Newell Rubbermaid’s technical analyst. "The CCSS solutions help us to go back and find out why the machine was running slowly at a specific time, for example. It means that not only has the fault resolution time improved greatly but also we're getting fewer incidences as time goes by, because we have had an opportunity to go back and iron out any minor problems."

QSystem Monitor can monitor standard performance parameters, like response time, CPU, and auxiliary storage, but also offers 12 user-definable bars and 10 text-definition bars, which can be tailored to each organization's individual needs. The data center team have allocated the text-definition bars to monitor the status of critical jobs and subsystems. Maintaining the active status of a series of "auto jobs," which run in the background to generate dispatch notes, invoices, and lists, is vitally important to business operations; if they should fail, the process would stop and prevent users from progressing with their work.

Now the team need not scour subsystems for the rogue job following a call from a user. Instead, problems are brought to their attention with a flashing red bar on the data center screen. User bars are devoted to monitoring job queue activity, and any stagnant jobs can be identified before they impede other jobs from completing.

Historically, gaining access to, and subsequently explaining the results of overly technical reports, failed to satisfy at a management level. Now Foster and York can use the summary information from QSM in their Cognos application, used by the company as a global standard for all their reporting requirements. This helps to integrate IT's role as a justifiable and cost efficient component of the company to managers looking for a broad but all inclusive business summary. E-mail compatibility means that these reports can be delivered on a scheduled basis directly to the inbox of key personnel.

Before installing QMessage Monitor, Newell Rubbermaid, like many iSeries shops, experienced a serious messaging incident that has since forced comprehensive message escalation to the top of their agenda. In this instance, a job started to loop early in the evening and began to deposit replica spool files into a non-printing queue. Before long, hundreds of thousands of unwanted pages began to build up and because the job restarted each time, it was not immediately obvious that this was the problem job. At the same time, the operator overseeing this was diverted away to an equally pressing situation and as a result, missed a critical message informing him that system storage was at 85 percent. The situation soon escalated to the point that the system was at 95 percent and IPL’ed immediately, and as a result, the manager had to drive to the office in the middle of the night to restore the system. Apart from the evident inconvenience to the manager and the operator, the team struggled for two days to find the cause of the IPL with only a painstaking investigation of the history log to assist them.

Foster explains how the system would react with QMessage Monitor in place, given the same circumstances. "Firstly, those types of critical messages are now defined so the system will let us know about them immediately," he says. "It flashes red and the speech software will literally shout out the problem to the entire data center! Even if by some kind of miracle that were to go unnoticed, the message would then begin an automatic escalation procedure that would page me directly. If I ignored it, it would page another member of the team and so on until it was resolved. If we had QMM running at the time of the incident we would have been fully protected, without a shadow of a doubt, but because we didn’t, we were exposed and the impact was huge. Fortunately, we have taken measures to ensure that the system is never that vulnerable again."

It seems no iSeries shop is without its share of daily problems, be they relatively minor or ones that could have a more devastating impact on the business. What is clear is that valuable lessons can be learned from experience. Implementing solutions that make best use of time, resources, and intelligent automation, means iSeries managers can enjoy the success of delivering a high availability promise to their users, every single day.

Greer Hahn is a freelance IT writer who has worked with several software vendors, including CCSS Ltd. E-mail: greer.hahn1@btopenworld.com