Cisco Cisco MGX-FRSM-HS2 B Serial Frame Service Module Libro blanco

Deployment Guide

Page 5 of 35

importance of these metrics is closely related to your specific network; for example, an enterprise with a large

number of branches and a small main campus might focus on WAN utilization, whereas an enterprise with one

main campus and one large branch with users that use collaboration tools across the two will likely focus on

application performance metrics such as response time measurements.

2. Create a baseline of current network performance metrics: The NAM can help document a variety of these

baseline metrics including applications, bandwidth per application, top conversations and hosts, QoS values

used in the network, unrecognized protocols, and current server and end-to-end response time measurements.

These measurements might meet or exceed your expectations in step 1. It might be worthwhile to revisit the

expectations set in step 1 and check whether some refinements are necessary (for example, 80 percent

utilization on the WAN link may be quite acceptable, whereas the real reason behind application delays seems

to be bursts of unrecognized traffic. In this case, one might be lenient on WAN link utilization and focus more on

QoS-related issues).

3. Enforce policies using alarms, syslogs, traps, and other alerts: NAM can provide alerts by email, FTP, and

other traditional methods like syslogs and traps. These tools must now be configured such that the normal

functioning range of the network is demarcated. If any of the tracked metrics show values that are outside this

normal range, then the NAM can be used to send alerts as appropriate. The information stored on the NAM is

openly available to applications. It is recommended that any enterprisewide network management tools and

monitoring applications be configured to receive alerts from NAM. The NAM is then able to act as a network

sentinel and warn proactively about a host of issues and also provide access to rapid troubleshooting when

problems occur.

Again, simply stated, the goal is to remain within the normal ranges of all important network metrics. But knowing the

normal range of the network is a constant learning process, and as the network evolves and grows, it can be a

moving target. Therefore the lifecycle described above is a continuous process of fine-tuning the network and the

metrics that are most important to normal behavior.

Figure 1. The Network Performance Management Lifecycle

Places in the Network Where NAMs Are Deployed

Because NAM is available in various form factors, it allows significant flexibility in deployment. At the same time, the

available NAMs must be deployed in locations that are most effective in helping you monitor, measure, and report on

the network’s health. Any location that is the ingress or egress point of a logical network boundary (aggregation

layer, core, campus edge, and so on) can offer valuable insights into the network activity within that partition.

Therefore, such boundary locations are usually good choices for NAM deployment. Figure 2 shows various possible

locations at which NAMs can be deployed. The access and distribution layers, the data center, WAN edge, and