On the surface, many network management products seem the same. But any network management system claiming to be a solution will, at a minimum, maintain an inventory of the network devices, understand their interconnectivity, monitor (and alert on) each device’s response (availability), monitor each device’s performance to some degree, and may note configuration changes. But beauty is only skin deep. How, how well, and how deeply an NMS does these makes all the difference.
Let’s take a look at how this plays out with respect to discovery and inventory (I’ll cover other major functional areas in future posts). There are degrees of what systems provide for inventory management.
Rudimentary: A very basic NMS requires you to manually load the device list into the system, which may not be complete if you forgot a device or two. Once the list is loaded, the NMS performs some basic level of pinging or SNMP polling to see if the device is responding. This level of functionality is usually found with the provisioning tools of a network device vendor and is specific to that vendor’s products only. If this tool isn’t free it should be. A vendor proprietary toolset may seem appropriate until a merger results in a heterogeneous multivendor scenario.
Basic: A slightly better inventory system automatically discovers the network inventory for you. It also performs basic availability testing and alerting. A frequent weakness at this level is that discovery is performed once at the beginning. Unless you manually “re-discover” inventory, it can soon be out of date (and providing misleading management status and information).
Intermediate: An improvement from basic one-time discovery is automated continual discovery—the solution continues to monitor the network for changes to the inventory, notifying you of additions and removals. With this system, the inventory list is now the “inventory of record.” At this level, inventory should be multi-vendor, and cover more than routers, switches and hubs. Modern network equipment also includes firewalls, load balancers, application accelerators, virtual switches, virtual servers, and more.
Advanced: Advanced inventory systems go deeper than top-level MIBs. Most low-end NMSs discover only the basic information about a device and gather only basic information, such as system name, location, and contact name as set by the administrator. Some may go further and report information such as CPU and memory utilization for the most basic and common of devices. But SNMP polling can garner much more detailed and interesting information about a device or its components and sub-components.
The deeper a monitoring solution can dig, the more insight you get into the health of the network. That information is invaluable for other business initiatives beyond keeping the network up. An accurate inventory, for example, allows maintenance contracts to be renewed accurately so you’re not paying for maintenance on devices you no longer use. It also lets you keep tabs on spare capacity (“Do we have enough?” “Do we have too much?”).
Now that’s a beautiful thing.