SPECIAL FOCUS:
NETWORK MANAGEMENT

From the May 2007 
issue of Communications News

The new face of network monitoring

The proliferation of edge devices brings the need for better management solutions.

by Jamie Lerner

While reliable, profitable IT operations require the right IP asset-monitoring solution, the right systems and a network-monitoring solution are equally important. Bringing sophisticated IT capabilities to devices beyond the network and to all IP-enabled elements (i.e., edge devices) can produce competitive results. With this in mind, an automated network-monitoring system approach can assist in managing the proliferation of networked devices–from mobile phones to HVAC systems to slot machines.

Until recently, only a limited number of specific-function devices were network connected or connectable. These devices were generally limited to servers, storage, switches and routers. As long as these were the only devices attached to the network, they could be monitored without the need for expensive, special-purpose tools.

Now, however, the number of unique devices is increasing exponentially. Whether these devices are point-of-sale terminals, robotics or security cameras, they can now all be viewed from the network. Understanding this shift showcases that monitoring only data center servers, storage, switches and routers is not enough for today’s environment, where there might be 20 or more new networked devices for every traditional networked device.

IT directors generally expect modern solutions to be easy to use, as well as to be presented with meaningful dashboards, data representations and reports that provide real information to help desk administrators and executives alike. As networks change, the common view is that the monitoring solution will grow in functionality to meet the specific needs of unique applications without an increase in complexity. The litmus test for modern network-monitoring solutions should not only be how well they monitor and report on data center equipment, but how quickly and how effectively a network composed of distributed IP-enabled devices can be detected and monitored, especially in an automated fashion.

Today’s network extends everywhere. Yesterday’s relatively stable data center environment, with a reasonably finite number and type of devices to monitor, was generally owned by a single entity, and effective monitoring of servers was in the interest of both the data center and the server manufacturer. Networks containing a wide variety of new IP-based edge devices do not present such a clear line of demarcation.

So what happens when new edge devices are added to a network? Manufacturers of these devices have been forced to provide their customers with unique monitoring capabilities, usually in the form of limited functionality tool kits, with no easy points of integration to other monitoring tools. Years of effort might be necessary to provide even limited monitoring for a single new device or application.

In addition, visibility is an issue. Most network-management solutions are not designed to monitor equipment and applications at hundreds, and potentially thousands, of locations. As a result, performance issues often go unreported or are not identified until they have impacted operations and revenue.

To successfully operate global network environments, organizations need a new network-management strategy. As networked devices proliferate to include everyday devices, IT organizations need tools that will automatically discover the devices, interpret key services that should be monitored and start tracking performance and availability. When there are problems, IT staff will be able to visualize all of the interconnections in the network to simplify problem resolution. When any node is added, removed or relocated, system-management tools will know about it and adapt monitoring in real time as changes happen.

For example, a national retailer opening a number of new networked stores across the country might not have a system in place for monitoring each new store as it is being rolled out. By implementing an effective and reliable network-monitoring system, however, the company can be assured it would have the capacity to monitor the status of each store every step of the way. A legacy network-management system would not be able to handle this kind of demand and could leave the retailer in a vulnerable position.

An effective monitoring system should perform the following functions:

  • discover new and existing devices on the network;
  • identify the configuration of devices, including operation system version and IP addresses;
  • place devices into predefined categories and measure uptime and performance of the network;
  • provide predesigned graphs for a particular node that show performance statistics and latency;
  • provide response information on each pollable service;
  • poll real-time statistics such as CPU, storage, software, users, processes, services and memory;
  • automatically build historical performance graphs for statistics for the network, servers and applications;
  • provide preconfigured thresholds for all network nodes and applications;
  • dynamically create topology maps to diagram the relationships between networked nodes and their dependencies, used for root-cause analysis;
  • automatically send notices or descriptive messages to a pager, an e-mail address or both when important network events occur; and
  • route and escalate notifications to groups and users based on duty schedules and scheduled outages.

The most effective network-monitoring system is one that is designed to respond to the changing face of the networked world, bringing sophisticated IT capabilities to edge devices and to all IP-enabled elements. Monitoring solutions should detect and help to solve system problems, and provide the metrics that enable sound executive decisions.

Jamie Lerner is CEO of CITTIO, San Francisco.

For more information:
www.rsleads.com/705cn-254