In a recent Info-Tech survey, fully one-third of the 115 respondents stated that their organization does not use any type of Network Management System (NMS) to proactively monitor the LAN and WAN. This is a disconcerting statistic, as a NMS is a critical component of infrastructure management, and should be a staple in every organization that has a network infrastructure.
Why a NMS?
A NMS is critical to properly managing an enterprise network. The following is a partial list of key NMS functions that are a necessity for organizations of all sizes:
- Proactive monitoring and alerting. The network management staff needs to be aware of network outages or performance issues as they happen. It is no longer acceptable to wait until users report a problem before reacting. When the CEO calls the CIO to find out why e-mail is not working, the CIO had better be able to respond that IT is already aware of the problem and are working to fix it as quickly as possible.
- Traffic insight. It is critical to know how much bandwidth is being used, and what it is being used for. To gain this critical insight, a NMS is essential. Network management should be able to proactively plan for network augmentation or expansion before performance begins to suffer.
- Root cause analysis. In order to understand why problems occurred, what systems, applications, and users were affected, and what network events were correlated to the root cause of the issue, a NMS is absolutely necessary. A good NMS will identify correlated events and help identify the root cause of a problem much more quickly than one step at a time troubleshooting.
- Drilling down to the offending device, service, or application. Often the first thing that happens in the event of a network issue is finger pointing. The server management team blames the applications team, who blame the network team. A NMS can monitor servers, services, and even applications to help identify where responsibility lies, and what team should be working to resolve the issue.
In most organizations, the network is business critical infrastructure that must be highly reliable and available. If the network goes down, so does employee productivity, customer interaction, and sales activities, while the CEO’s blood pressure goes up. Implementing an enterprise NMS will identify issues sooner, help troubleshoot problems, and inevitably improve network availability.
How Much Is This Going to Cost?
The cost of NMSs is widely variable. There are excellent open source NMSs available at no cost, while other large enterprise solutions can run into six figures, and every price point in between. Some examples of products and pricing include:
- Open source (Nagios, OpenNMS). These open source products are free, and run on Linux, BSD, and Unix-like OSes. They are full-featured and quite simple to get started with.
- Low Cost (SolarWinds, Ipswitch, LAN/WAN vendor solutions). These products are full-featured and easy to implement and manage. Solutions start at roughly $2,500. Some LAN/WAN vendor solutions are even free, such as HP ProCurve’s PCM (ProCurve Manager).
- Mid-range (NetScout/Network General, NetQoS, LAN/WAN vendor solutions). These products tend to be very capable and more specific to the needs of larger organizations. Solutions typically start at over five figures, and can reach six figures.
- Large enterprise (HP OpenView NNM, IBM Tivoli, CA Unicenter). These solutions can become quite complex and often run into six figures for a complete implementation.
Recommendations
Every enterprise with a network should have a NMS. With costs starting at nothing, there is no excuse for not having one. Consider the following when evaluating solutions:
1. Start small. If there is currently no NMS in place, start by implementing a free or low cost option. If the network is large, consider monitoring and managing a section of the network first to get comfortable with monitoring, alerting, and troubleshooting.
2. Put some processes in place. Having the software monitoring the network is a good start, but largely useless if no processes are defined for who does what in the event of a problem. Define who will be alerted to problems when they occur, how long they have to respond, when the problem gets escalated (based on time and severity), and what troubleshooting steps should be taken.
3. Don’t forget about people. If the intention is to monitor the network 24x7, make sure that there are capable people who are willing to be on call. It will also be necessary to define on-call and response compensation.
Bottom Line
Many enterprises do not have a NMS in place to monitor enterprise LAN and WAN infrastructure for faults and performance issues. Enterprises that lack a NMS also lack insight into network, server, and application performance. With a plethora of low cost or free options available, there is no reason not to have a NMS in place.
