Smart tools alleviate pain: managing more complexity with fewer resources is the challenge for today's IT managers. (Internet).

Today's IT environment is characterized by a variety of architectures, platforms and technologies. Within five years, hundreds of millions of people will be connected, using wireless and other devices, driving trillions of transactions and accessing large amounts of rich media and storage. For IT managers, this spells more complexity and greater challenges as they try to track, monitor, manage and troubleshoot the IT assets spread out over the network.

Compounding this challenge are shrinking IT budgets and a shortage of critical skills to manage the increased complexity. In short, IT managers are being asked to do more with less while, at the same time, maximizing system availability for their end-users.

The solution is not simply more "systems management" but smart systems management--the management of IT assets using tools with proactive, predictive and automated capabilities. Smart systems management can handle the repetitive operational tasks and allow the system administrator to manage proactively rather than reactively.

For example, configuring and deploying a server can be a time-consuming and complex undertaking. The server may be in a remote location where technical skills are minimal. The cost and complexity of keeping these remote systems current with the latest software and firmware can be a challenge.

Deploying smart systems-management tools that include versatile utilities can simplify the deployment and configuration of systems on the network and provide administrators the ability to remotely distribute an entire system image or firmware updates, saving time and travel. These tools should also be able to take advantage of intelligent and automated facilities, such as configuration wizards, scheduler and event-action planning.

AUTOMATIC DEPLOYMENT AND CONFIGURATION

Wizards make server deployment easier by automating the configuration of hardware settings. Once configured, profiles add convenience and time savings by enabling the configuration to be applied many times to other systems. Some solutions can be automated by including policies in the profile that give instructions to automatically deploy a software image according to the defined policy. Configuration and deployment is automatic, with no human intervention required.

Another tedious configuration task the IT administrator faces is installing PCI cards in the appropriate slot, while making sure input/output (I/O) throughput is optimized. Many administrators do this by trial and error, consuming valuable time. This task is magnified in environments with large numbers of I/O slots, such as those with remote I/O drawers attached to the network.

PCI cards should be managed with systems-management software with built-in intelligence that alleviates the pain of manually trying to match the PCI card to the correct slot to ensure optimal I/O performance. Such tools help optimize I/O performance by matching card characteristics to the PCI-X bus and offer guidance as to the best slots in which to install PCI and PCI-X cards.

Some cards, such as next-generation 10-Gbps (or higher) InfiniBand, Ethernet and Fibre Channel cards, will benefit from running in 133-MHz PCI-X slots, while other cards will run at top speed in 100-MHz, 66-MHz or even 33-MHz PCI slots. The software should also recommend whether cards should go in the server chassis or in remote expansion enclosures containing an additional six or 12 PCI-X slots.

Once a system is configured and running, another concern for the IT staff is determining if a system requires software updates and how to easily distribute them to remote locations with minimal disruption. To alleviate this challenge, tools are available that give proactive notification that updates are available for a particular server; some even categorize how critical the update is. These smart tools also automate the process of finding the update and scheduling its distribution.

Perhaps the most unbearable pain of all is the cost impact of downtime on servers running business-critical applications. Although there are many estimates regarding downtime costs, one fact remains constant: costs are incurred when a server goes down and these costs can include lost revenue, lost productivity and even damage to the company's reputation.

TOOLS WARN OF PROBLEMS

Reducing these costs can be accomplished by either avoiding the occurrence of a system failure or minimizing the duration of the outage. Although some types of system failures are unpredictable, many types are gradual and, over time, can snowball into a costly, unplanned outage. Smart tools can predict and provide advance warning of a brewing problem, and enable rapid recovery if a failure occurs.

Server resource-utilization and environmental parameters--like temperatures, voltages and fan speed--can be monitored to help predict failures. There are some tools that continually monitor the utilization of processors, disk drives, memory and network bandwidth to identify bottlenecks for an individual system or a group of systems--and make recommendations to alleviate diminished performance or downtime.

Some tools produce reports to quickly isolate historic problems and identify trends that indicate impending trouble. Once notified, alerts can be triggered to allow automated corrective actions, minimizing the involvement of the IT staff.

Just like hardware, software often exhibits an increasing failure rate over time when there is unbounded resource consumption. This is referred to as "software aging." Some tools exist that monitor software behavior trends and predict system outages due to this software aging phenomena.

Notification of a potential future outage can give administrators the opportunity to take corrective action and avoid unplanned downtime, or to automatically schedule a system restart during non-peak business hours. After the system restart, resources are reclaimed, and the software failure rate returns to its initial lower level.

Managing more complexity with fewer skilled people requires more than just systems management. Smart tools with proactive and automated features that help offload the valuable time of the IT staff are also needed. The time invested in implementing these tools is more than offset by the payback: more productive IT staff and cost savings from alleviating the pains associated with system deployment and downtime.

For more information from IBM: www.rsleads.com/308cn-259

Mazanec is a product manager at IBM in New York, N.Y.