As we transition to a new editor at NetworkPerformanceDaily.com, we’re going to be reprinting some of the best articles from our archives for a little while. We’ll have new content up shortly.

Originally Published November 29, 2006

by Ed Tittel

The ISO/OSI Network Management Reference Model is usually rendered as FCAPS: Fault management, Configuration management, Accounting management, Performance management and Security management.

This model fails to give full weight to the impact of performance. Performance drives perception, which means that, from a user's standpoint, the source of poor performance doesn't matter as much as the fact that performance is, in fact, poor. According to Denise Dubie at Network World, network managers and engineers are being increasingly tasked to prioritize performance and user experience:

"Distributed IP networks and complex real-time applications have forced a change. Now network managers need to be in the know from the start about application performance, helping developers understand what will work on a network, spotting poorly performing applications before users feel the effects and delivering LAN-like performance over the wide area to remote and branch offices."

In other words, it's not just about monitoring devices anymore. It's about delivering services, at a reasonable cost, in a reasonable amount of time, where users are increasingly asked to decide what's reasonable, time-wise. (For more, see Network World's "User experience is key".)

A performance-first approach (see whitepaper [PDF]) to network management turns FCAPS into PFCAS -- or rather, PCAS, given that fault may be considered merely the most extreme expression of bad performance. The performance-first paradigm inverts the traditional, bottom-up device-monitoring approach and begins with top-down visibility into overall performance of applications running over the network.

Infrastructure availability and utilization aren't the only gauges of network health. Why focus the entirety of network management efforts on the small fraction of network issues caused by hardware or software infrastructure failures?

The fundamental purpose of the network infrastructure is to transport data from one end of the system to the other as quickly as possible. The more efficiently data flows at the transport layer, the better applications perform. Hence, end-to-end response time measurement is the best measure to use when deciding how to optimize the network, plan new infrastructure rollouts and upgrades, and identify the severity and pervasiveness of problems.

This approach recognizes that, between the limits of the network and application infrastructure being “up” or “down,” performance can--and does--vary widely. It is not uncommon for availability status indicators in the Network Operations Center (NOC) to be “all green” even while the help desk phones are ringing off the hook with users complaining about slow response times.

By focusing on the performance of key applications running over the network, IT organizations can focus on what's most important: making informed infrastructure investments to support business demands; delivering consistent, acceptable end-user response times; and quickly resolving business-critical problems. IT organizations that successfully make the transition to a performance first approach typically -- and deservedly -- receive high marks from the business lines they serve.