5.16.3.1 Discovering Networks

 



 

Overview

 

The active discovery process goes about enumerating or sweeping the whole address range of the network (based on the class of the network), individually discovering each IP address. In order to discover all the active (or alive) nodes in the network first, the discovery process maintains a list of IP addresses that are possibly active. The list gets populated by various ways, which are listed below:

Then, for each IP address in the network as taken from the list first and then the other sources, the discovery process will perform tests to confirm whether the corresponding node is active or not. Among the tests performed, the first test is to call the user defined discovery test. User defined tests are specified in the userdiscovery.conf file, where the user-class to be invoked and other specific parameters are given to perform user defined discovery process. If the discovery process succeeds in performing the user-specific test, the nodes or objects are considered to be active and the objects specified by the user will be added to the topology database. Please refer "User Defined Discovery" section for more information.

 

This is followed by SNMP, ICMP ping process and TL1/CORBA testing using the entries in tl1seed.file/corbaseed.file. For SNMP ping, an SNMP GET request for the System group OIDs (of the RFC-1213 MIB) is made and a response got would be taken as the corresponding node being active. If there is no response for the SNMP ping, an ICMP echo packet would be sent (using the 'ping' system command available in the underlying operating system), and a response got for it would again be taken as the node being active. When the SNMP or the ICMP ping succeeds, then the corresponding interface and node objects will be added into the topology database using the topology module.

 

The objects discovered by the above process, before being added to the database, are passed through user written discovery filters. This is done to allow users have control over the objects and its properties.

 

 

This is how the discovery engine goes about discovering external networks. Once the entire network address range has been swept or enumerated, the tasks will wait for a configurable amount of time and then again starts the discovery of the network (rediscovery process). This whole process is repeated for ever, for each IP network, unless the discovery process is manually stopped.

 

The following points gives you a summary of the processes performed during a network discovery:

Note: NMS discovery engine will start the discovery of Networks that are already configured. If no Networks are specified, by default, the discovery engine will start with the Local Net discovery.

 

Back

 

Discovering Other Networks

 

Networks that are not part of the Local network and not explicitly configured for discovery are treated by NMS as 'other' networks. The existence of other networks is found during the discovery of routers in configured networks.

 

There are two parameters that define NMS discovery engine's behavior w.r.t discovering 'other networks'. They are

These parameters take true/false value and can be configured as arguments to DBServer Process in NMSProcessBE.conf. For information on the DBServer process please refer BE Startup Options for Topology.

 

MANAGE_OTHER_NETWORKS

 

For unmanaged Networks, operations like status polling, discovery will not be performed. Depending on the MANAGE_OTHER_NETWORKS value configured in the NmsProcessesBE.conf, the other Network discovered via a Router discovery will be managed/unmanaged.

 

If the MANAGE_OTHER_NETWORKS parameter is set to 'true', Web NMS Server will add the other (sub) networks identified during the discovery of routers as managed.

 

DISCOVER_OTHER_NETWORKS

 

Discovery of other Networks will be started only when MANAGE_OTHER_NETWORKS and DISCOVER_OTHER_NETWORKS value is set to 'true'. Discovery of other Networks added as managed can be stopped by setting DISCOVER_OTHER_NETWORKS value to 'false' in NmsProcessesBE.conf.

 

The following table gives you a matrix of enabling and disabling Managing other networks and discovering other networks and the status of the net discovery (started or not).

    1. MANAGE_OTHER_NETWORKS DISCOVER_OTHER_NETWORKS

      managed

      discover

      NetDiscovery

      false

      false

      false

      false

      not started

      false

      true

      false

      false

      not started

      true

      false

      true

      false

      not started

      true

      true

      true

      true

      started

'managed' and 'discover' are properties of Network Object. When managed property of a network is set to 'false', discovery and status poll will not be performed. When managed is set to 'true', status poll for the network object will be started and discovery/rediscovery of network will be started based on the 'discover' value. When discover is set to "false", no discovery will take place, and when it is set to "true", discovery will be started provided the network is managed.

Back

 

Subnet Discovery

 

During the discovery of a network, the discovery engine will try to discover all the Nodes in the Network unless the discovery of a subNet of this Network is started as a separate process.

 

For example,

 

Network A contains subnets B and C. Let us take two IPs i1 belonging to B and i2 belonging to C. The seed.file is configured to discover Networks A and B. The two Network discovery A and B is started as two separate Net discovery process and during the discovery of devices in A, care is taken not to discover devices that belongs to the subnets of Network A that are added into the database.

 

So i1 will not be discovered during discovery of Net A discovery. Also if Net C is added into the database, i2 will not be discovered in Net A discovery.

Back

 


Copyright © 2013, ZOHO Corp. All Rights Reserved.