Performance and Fault Management: A Practical Guide to Effectively Managing Cisco Network Devices (Cisco Press Core Series) - page 76

It is important to understand the interrelationships between these categories. Figure 9-1 describes the interrelationships between the categories of network management software that are described in this chapter.

Figure 9-1 will be referenced throughout this chapter to help you understand how the different tools need to integrate.

This chapter covers the following categories of network management tools:

• NMS frameworks

• Knowledge base

• Performance measurement and reporting

• Event and fault management

• Policy management

For each of these categories, this chapter will discuss the criteria you might use to determine which tool to use. Some of these criteria apply across all categories of software. These criteria include the following:

• Capability to integrate with other applications. We list the capability to integrate with other applications first. If you will be able to meld these various tools into a comprehensive network management system (NMS) for your network, it is very important that the tools you select can do tasks such as share data and launch one tool from the other.

• Support for the equipment in your network. Sophisticated network management is not useful if it doesn't support the equipment you have in your network. Checking the supported equipment against your network is an important part of the criteria for choosing your set of tools.

• Application reliability. Nothing is more frustrating than having a tool fail when you are trying to debug a serious problem. Application reliability may be hard to determine, but it is an important criterion in network management tool selection. Because the vendors of network management systems are not likely to advertise the reliability of their tools, you will have to evaluate that for yourself. You can start by getting references for the tool and checking them. It is very important to try the tools out on your network. Try performing all the functions you will need to use and evaluate their reliability for yourself.

• Capability to scale to support the number of network devices in your network. All applications have limits for how much they can support. Check the tools you are evaluating against the number of devices you need to manage. Note that the amount of devices supported by an application probably will vary, depending upon how you use the application. For example, an application that polls your devices will be able to support more devices at a slower polling frequency than at a higher polling frequency.

• Adequate performance. A related criterion is application performance. Although a tool may technically support the number of devices you need to manage and do what you want to do, it may not meet your requirements for speed while doing so. You'll probably also need to determine whether a tool provides adequate performance by evaluating the tools that otherwise meet your requirements.

• Ease of use and maintenance. Obviously, tools that are easy to use and maintain are preferred. Easier-to-use tools will also save on training costs.

• Operating system and platform support. Your company probably has selected a standard computer type and operating system. Tools that run in that environment will be easier for you to support and use. However, tools that have a web-based GUI may allow you to use tools, even if they must run on a non-optimal platform for your company—or even a mix of platforms. You must evaluate the increased support costs against the other criteria for each category of tool.

• The capability to work when the network is under stress. Your network management tools must help you solve problems on your network when your network is not healthy . If you must use the network to access your tools, you may have problems using them under certain conditions. Careful network design, including dedicated management networks or VLANs, and isolating the NMS components from other parts of your network may help your NMS deal with this. If your event and fault management systems are so overwhelmed with processing events during a network crisis that they can't supply you with correlated faults, they won't be of much use to you just when you need them. Stress testing the tools you select is an important part of the selection process.

Now, on to the tool categories.