Enterprise Network Management  The Good, The Bad, and The Ugly

RFG believes enterprise environments contain elements of network management thatare good, sometimes bad, and maybe, just plain ugly. Technology complexity and continuouschange in the network infrastructure provide an ongoing challenge for any networkmanagement system. Consequently, management system updates and house cleaning is requiredas part of an annual checkup program. CIOs should review the annual checkup results toidentify and eliminate the Ugly, upgrade or eliminate the Bad, and fully utilize thefunctionality of the Good elements in the enterprise.

Business Imperatives:

• Good management systems will clearly demonstrate effective fault and performance management that ensures reliable user access to business applications. CIOs should review network performance and business requirements to determine if the network management systems are contributing to network availability and reliability.
• In network management, fewer are better and integrating disparate network management solutions into one or two management platforms is beneficial. CIOs should review all enterprise management systems and determine their contribution to network uptime and reliability. Then, an integration or replacement strategy should be prepared that includes expected integration benefits, complexity, and costs involved.
• An annual checkup is critical to insure the enterprise network management infrastructure continues to proactively meet user access and business application requirements. CIOs should direct network operations to perform annual health checkups on each network management system to maintain high levels of network performance.

Good management systems provide an unmistakable view of everything important occurringin the network at a given time. This includes concise alarm information, identification ofpotential data bottlenecks, quick outage identification, and non-disruptive planning andmodeling tools for adding or moving equipment and services. Moreover, they also provideusers with reliable access to the enterprise business applications. Without qualitynetwork management, IS and network administrators are at the mercy of impatient users whocould claim they lost hours of productivity and gigabytes of crucial data, when inreality, the network connection was only down for a few minutes. CIOs should verify thatpresent network management systems are meeting network demands by asking users and ITsupport staff how the network is performing.

Some management systems provide terse alarm alerts and cryptic notifications thatrequire operators to perform time consuming research to identify problems and to determinewhat actions to invoke. Others may require cryptic command line interface (CLI) entriesthat are time consuming and non-user friendly. CIOs should consider upgrading or replacinga system that exhibits a poor user interface with a user friendly GUI (graphical userinterface). Likewise, CIOs should ensure alarm correlation, thresholding, and probablecause capabilities are installed and operational because these basic functions aremandatory requirements to achieve improved network availability and reliability.

Moreover, enterprises do not have a shortage of network management systems anddiagnostic tools. Typically, operation centers are jammed with terminals and printers in achaotic attempt to manage all the network elements. Still, the challenge remains toquickly derive meaningful alarm information from the various management systems thatimproves network availability, reliability, and maintains user satisfaction. Manymanagement systems are effective at diagnosing troubles with their specific managedelements but this is no longer enterprise compliant. The best enterprise managementsystems can identify the affected users and impacted business applications while manyvendor-specific point solutions fail to provide this important information.

Likewise, integration tools are mandatory to import and export information from thevarious enterprise management systems and to utilize these feeds as input for correlationand probable cause functions. Not all network management applications are tightlyintegrated with the system itself and as they say with children’s toys and barbecues”some assembly is required.” The integration objective is to consolidate andcorrelate information from management systems when possible and to reduce the number ofterminals and associated alarms operators must deal with. Point solutions and homegrownproducts that cannot integrate may require new software or replacements. CIOs shoulddetermine which network management systems and functions are appropriate for integrationand calculate the cost, complexity, and timetable required for completion.

To maintain management systems in peak condition, the network operations staff shouldconduct annual health checkups for each management system and vendor point solution, asshown in the table below. While not all encompassing, the checkup provides a quick-glanceprofile that positions a management system into the Good, Bad, or Ugly category. Many Badcategory systems can probably be upgraded to the Good category; however, most Ugly systemsare technically difficult to upgrade because of hardware and software limitations.Nevertheless, some Ugly systems, especially homegrown internal varieties, may stillprovide significant value and should not be eliminated or replaced unless their specificfunctionality can be provided by another system.

As a final point, the management system architecture should be evaluated forstandards-based protocol support and open APIs. This includes four key areas for openmanagement systems: interoperability, operability, portability, and scalability.

Leading management systems shine in operability and provide easy access to associatedapplications and system users along with excellent administrative capabilities. One keyaspect of operability is consistent behavior at the user interface (GUI) where managementinformation is displayed and acted upon. Access to management information repositories anddatabases are also part of the operability picture and should not be overlooked. CIOsshould recognize that the best management systems provide intuitive GUIs and a logicaluser workflow to help complete any task with minimal user confusion and time delay.

Good systems also exhibit excellent portability including easy and consistent access tomanagement services, strong development tools for third party application support, andeasy porting of management applications to different management platforms. Portabilityincludes not only horizontal and vertical applications but also core applications, whichare part of the management platform framework. Java and CORBA technology are now appearingin the architectures of many network management systems because of their excellentcapabilities in this area.

Interoperability is another important area and it allows management platforms totransparently exchange information with other peers and software applications. The supportfor open interfaces, RFCs (request for comment), and popular communication stacks such asSNMP and TCP/IP is critical. Likewise, development tools that directly supportinteroperable platform interfaces are also an essential part of interoperability. The Badand the Ugly systems will earn low marks in interoperability and CIOs should plan toeither upgrade capabilities in this area or replace the systems with better products.

Finally, scalability is the ability to easily and cost effectively expand managementcapabilities over larger systems with more managed entities. This includes the totalnumber of managed elements, user administration workstations, simultaneous applications,and administrative capabilities needed to slice and dice management system functionalitybetween different platforms and network operators. CIOs should review the annualmanagement system checkup results and initiate an action plan to integrate, upgrade, orreplace management systems as required to meet enterprise business requirements.

RFG believes CIOs should direct network operations to evaluate existing andproposed network management systems on an annual basis to determine where they each fit.Good systems should continue to be fully utilized, the Bad should be either upgraded orreplaced, and the Ugly should be evaluated for replacement with newer technology. CIOsshould also review the enterprise management integration status and verify maximumconsolidation occurs in order to reduce overall network management costs and improveoperational efficiency and performance.

© 1999 Robert Frances Group. All rights reserved.

John Stehman is the Robert Frances Group’s Principal Analyst. He can be reached at 203-291-6900 or jstehman@rfgonline.com