Browse Prior Art Database

Using meta-model and change detection for optimised SNMP network discovery Disclosure Number: IPCOM000197746D
Publication Date: 2010-Jul-21
Document File: 2 page(s) / 63K

Publishing Venue

The Prior Art Database


Using meta-model and change detection for optimised SNMP network discovery

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 2

Using meta-model and change detection for optimised SNMP network discovery

Network monitoring tools have to first discover the target network in order to collect and report metrics. This discovery is performed via SNMP (which is a standard supported by a number of vendors and their devices). The results of the discovery are stored in a database for, among other reasons, change detection (for example, as described in

    Presently, the scope of this network re-discovery can be controlled horizontally (specific IP addresses) or vertically (specific technologies to be discovered on a device). For example, the network monitoring tools can be configured to discover devices on a range of IP addresses, a specific IP address, or to discover a specific manufacturers' QoS for a target device (at an IP address).

    However, on large devices this isn't enough where the effort to identify a change on the devices can both be time consuming and also put a lot of load on the device (in the form of SNMP queries). There is a need to minimise the impact for pinpoint / driving SNMP based discovery.

    The present invention is to build a meta-model complimentary to the change detection mechanism, and use it to drive the discovery process in an optimal way, reducing the amount of SNMP interrogation on the device.

    This approach would result in faster discovery of changes on an already discovered device, reducing the overall time in the provisioning process, as well as the CPU load on the monitored devices for the purpose of discovery.

A change detection mechanism is already present in most network monitoring tools (and described in some detail at These change detection mechanisms store data only for their use, discarding the rest.

A meta-model is built on top of the change detection mechanism to capture the current state of MIB tables in the target device on the network. This includes items such as:
1. For all MIB tables, arrays or maps are created indexed as in the MIB structure.

2. All indirect dependencies between MIB tables are resolved (such as those for Class Of Service, VLANs, etc), so logical items can be referenced directly. For example, runtime QoS Objects MIBs are indexed by cbQosPolicyIndex and cbQosObjectsIndex for each interface (therefore, 3 indexes to reference the actual runtime instance).

3. For each item in the tables...