A Discovery of the Network is required to be able to create Inventory/ITIL and Network Mapping.
The Network Discovery Tool automatically creates a detailed Inventory of Software and Hardware without using Agents. Updated and tested over 17 years of development, the field proven discovery engine, discovers large and small networks automatically.
The results are stored in a MySQL DBase facilitating direct customer access, it discovers old and recent devices using its extensive network protocol repertoire and huge device library.
Multiple different networks can be discovered with historical versions, stored and accessed individually.
After a Network Discovery these multiple databases can be selected in the Maps and ITIL features. A user library of discoveries, including historical discovery versions, can be created and used by the Inventory and Map features.
Any number of unique Named Discoveries can be created, critically with Historical Versions being created on Re-Discovery of the same named network. These versions can be compared in the Web Maps to see additions and missing devices per Historical Discovery per named Database.
The discovery engine covers both new and old devices in its more than 20,000 device library. Any number of subnets can be defined to exactly control the Discovery Domain – a critical feature out in the field.
One or more Subnets can be entered to control the scope of the discovery. This is typically important for large networks.
Note. the subnets can be pre-checked by the user, using the system in a SNMP/Ping scan, together with a full SNMP MIB browser.
The discovery is started by clicking a button, and a Summary Report produced.
Discovery is also used to automatically setup the Monitoring of the Network as it learns what devices and links are present. That allows the Discovery to be used to automatically poll and monitor the network. Other major features such as Alerts and Ticketing know which IP addresses are associated with which devices adding detail and organises Alerts internal (such as CPU Overload, Slow Ping Reply, etc) and external (Syslog and SNMP TRAPS).
Discovering the Network is fundamental to Mapping and Monitoring the Network as it learns what devices and links are present. That allows the Discovery to be used to automatically poll and monitor the network.
It also means setting up and displaying Netflow and all other features like Alerts know what IP addresses are associated with devices.
Note: only the LAST Discovery is used to setup Monitoring and do Device Drilldowns, this may not match a Map View created from a previous Discovery.
The Discovery can be edited to assign specific Icons to devices and change Location for instance, together with assigning Classes that can be used in Reports. Modifications are preserved across new Discoveries.
The Discovery Engine uses a variety of techniques to discover devices, such as inspection of ARP tables and controlled scanning techniques. This overcomes a limitation of many existing approaches, which need to know what to discover in order to draw a network. Once discovered, devices are queried using SNMP for MIB 2 and current vendor MIBs. The Discovery system has a stored device list database of most current and many old generation equipment types.
If this is a new Network then select Discover New Network and enter a unique Network Name (FYI will create a MySQL database with this name). Otherwise to re-discover an existing network select Use Existing Discovery, the system will tag this with a new discovery History Number.
The user can either start a completely new Discovery or Re-Discover an Existing Discovery.
Enter a memorable name for your network. The name will be used to reference the discovery in maps, reports and file-names, so do not use the following characters in the name :/\*?”<>|.
Note: If you plan to run the discovery process more than once on the same network this name is used as a cross-reference, so it is recommended that you keep Network names consistent.
The discovery operation will always include a Ping scan for Class C networks.
You have a Tick Box option to extend the scan range to cover Class B networks. Each Class B network found during discovery takes approximately 1 hour to ping scan.
The Ping Scan facility is not used on any Class A networks found.
Use the Slider to quickly setup common options to control discovery detail versus discovery speed.
Alternatively, simply tick the tick boxes to control the discovery exactly as required.
You can import a range of known IP addresses into the Discovery facility, i.e., effectively create a list of seed addresses that subsequently provide a means of speeding up the discovery process, as the Discovery Engine will bypass the early discovery stages and immediately activate SNMP operations for the listed addresses to obtain the additional information it needs to complete the autoMap process.
Experience has shown supplying a Seed List to Discovery greatly speeds up the process and produces a better discovery, so is HIGHLY RECOMMENDED.
There are two ways to create a seed list:
1) Manually Add ranges one at a time using the Add IP Address Range button, good for a couple of ranges.
2) Specifying a text list using the Cut/Paste Seedlist button. Use this to Paste in saved text file seed lists, good if there are many seed list ranges.
Note. The two systems are compatible with each other and can be used together.
You can check that the start parameters are ok, prior to starting the discovery. The check will provide confirmation that the address will :
Click on the Scan button to initiate a test to an IP Range. The results will appear on the righthand side window as below. Note, it can take a while for the test to complete.
You can also use the Test Seed List as a quick means of checking that your managed devices are going to be able to respond to the SNMP requests associated with the discovery process.Instead of starting the discovery, just temporarily change the Seed address to the IP address of the managed device, add the SNMP Information and click on the Test Seed List button.
Optionally, to send an email on discovery completion fill email address To:field.
To automatically setup Polling/Monitoring after discovery tick Setup Polling when Done tick box. Note: these Polling settings can be revised at any time under the Device Monitors tab.
To start Discovery click Start Discovery button.
A report is produced in the righthand panel of the GUI. It contains counts of devices discovered organised by device category.
To terminate the Discovery Engine, click on Abort Discovery button – this can take a long time to complete in some cases.
When the user clicks on Abort Discovery button during a Network Discovery, the user is now asked to confirm the Abort as above.
To get an update of the Discovery process click Refresh Report.
Only devices that respond to SNMP will be identified and classified. Unknown devices can be mailed to your reseller and included in a release.
Windows has an option to make Devices ‘Discoverable’, ensure this is enabled. Otherwise, they will just be discovered as IP addresses – if at all.
Note. workstations do not usually have SNMP enabled. The key to getting the WMI to work is to make sure the machine discovery is running and allowed to ask the workstations for WMI information. Multiple sets of WMI credentials can be entered into the Discovery setup, similar to SNMP community names, please refer to your network administrator for help on this.
Tick the Detailed Logging tick box to collect detailed Logging of the discovery which can be used by Spider Support to diagnose and issues associated with the discovery. The discovery may be slowed down slightly.
Under the Discovery Report Option Buttons there is also a Visio Diagrams Download Facility. It is essential that Pop-Ups are Allowed to Download the Visio File (check browser settings).
The current discovery can be updated throughout the Toolbox features for device Location and device Icon.
Additionally, up to 2 user defined Classes can be assigned to Devices that can be used in features to filter/group devices.
Updates can be Reversed at any time.
Grid tick boxes are used to specify individual devices for modification.
Columns that are being edited have a green background.
Click on the Modify Loaded Discovery button to access the Modify Discovery tabbed grid. This pop-up has extensive instructions provided.
There is Merge Discoveries function when rediscovering a network.
Automated Network Discovery
This feature has the capability to automatically schedule and execute Network Discoveries using the product’s web GUI.
Multiple discoveries can be run sequentially.
Probes can be used to run parallel Discoveries to accelerate total discovery time.
The Enterprise Inventory Explorer can be used to combine the Databases or several discoveries.
First an existing Discovery is selected, so a Discovery must be completed before this Re-Discovery feature is available.
The user can set repeat discoveries on the following criteria:
Below is the Re-Discovery Interface, launched by clicking on the Discovery Scheduler button as below:
The Rediscover options are set as below per discovery:
Note that continuous mode means a new discovery starts when the discovery engine if free. There can be multiple discoveries set to continuous mode, in this case they are done in turn.
A typical inventory/discovery run can take between five minutes and many hours. The rules set out below will assist in determining the time an inventory/discovery will take.
WARNING – Platforms running Discoveries must have hibernation/standby disabled.
Each IP address is tested individually for Ping, SNMP v1, SNMP v2c or SNMPv3, and optionally WMI, NetBIOS and SIP Queries for a total of 15 seconds.
If that IP address is an SNMP device then it is fully processed. The processing of a simple device such as a printer takes 15 seconds. The processing of a high-end router or switch with hundreds of interfaces and VLANs takes several minutes.
You also need to be careful in a situation where the Ping Scan is applied to Class B networks, that process will slow the discovery and if it is not required should be disabled.
The average time to process one IP address is 30 seconds.
This average, combined with the number of IP addresses, and the number of parallel processing threads, can be used in the formula below to calculate an approximate maximum time for the inventory duration.
Maximum Duration (Mins) = (Total IP Addresses)/(2 * (number of parallel processes)) A network can include many subnets of different sizes:
Number of IP Addresses
Subnet Class – Number of IP AddressesA – 16,000,000B – 65,000C – 254
The total number of IP addresses in this context is the sum of the number of IP addresses in all subnets. Not all IP addresses are used, however, hence our calculation is an upper limit on inventory/discovery duration.
If the total number of devices is actually known, then it should be used instead. This would give a much more accurate view of the duration of the inventory/discovery.
Number of Parallell Processes
The Discovery Engine can process multiple IP addresses in parallel as a means of speeding up the inventory. The user can adjust the bandwidth using the Discovery Speed slide bar provided in the Discovery Setup.
A facility has been added to the Toolbox to allow any IP Address to be investigated with a SNMP Browser, this is especially useful when debugging Discovery Issues.
The Browser is launched from the Discovery Tab Panel called Setup and Test IP List– as can be seen in the picture below. Clicking the Open Browser button opens the Browser window.
The Browser opens in a new dialog box as below, this is called the MIB Tree:
To Browse a Device such as a Server or Switch one enters the IP address of the target device as in the picture above.
Next open a MIB Tree leaf like system above then click on sysDescr to show the MIB System Description. After a few seconds an analysis appears for, in this case for sysDescr.
The Toolbox Web GUI sends a SNMP request to the Toolbox server which then polls the requested MIB information and returns it to the Toolbox Browser.
The sysDesr example above is a simple SCALAR value, however by clicking on a MIB Table like ifTable, a full list of rows like Interfaces, can be examined too.
Note. the Grid can be searched, sorted, filtered and also exported to a CSV file or alternatively printed.