Configuration Management

IACT 418/918 Autumn 2005

Gene Awyzio

SITACS University of Wollongong

2

Overview• Configuration Definitions

– OED: Arrangement of parts or elements in a particular form or figure

– OED: The way the constituent parts of a computer system are chosen or interconnected in order to suit it for a particular task or use; the units or devices required for this

– Websters: 1 a : relative arrangement of parts or elements: as

• (1) : Shape

• (2) : functional arrangement <a small business computer system in its simplest configuration>

– Webster: 1 b : something (as a figure, contour, pattern, or apparatus) that results from a particular arrangement of parts or components

3

Overview• Configuration Management is the process of

– Obtaining data from the network

– Using that data to manage the setup of all network devices

• Involves

– Gathering information about the current configuration of the network

– Using that data to modify the network configuration of devices

– Storing the data

– Maintaining an up-to-date inventory

– Producing reports on that data

4

Benefits of the Configuration Management Process• Enhanced control over configuration of network

devices

– Rapid access to config data

• Speeds up re-configuration• Examples:

– Configuration data holds the current setup of each device

– If you wanted to add more interfaces to the device you would need to know

• Current number of interfaces in that device

• Network addresses assigned to those interfaces

– Config management would hold that data and enable rapid access

5

Benefits of the Configuration Management Process• Also assists in modification of network devices

– Remote deactivation etc

– Remote changes

• Inventory control can be used to

– Negotiate vendor contracts

– Examine vendor quality

6

Accomplishing Configuration Management• Configuration Management steps

– Gathering information about the current configuration of the network

– Using that data to modify the network configuration of devices

– Storing the data

– Maintaining an up-to-date inventory

– Producing reports on that data

7

Collecting Data Manually

• Obtaining data often begins with a manual collection

– Need to record the devices

• Serial number

• Address assignment

– Store data in a spreadsheet, database etc

– This can be tedious and error prone for large networks

8

Collecting Data Manually

• Data can be collected automatically

– Using network management protocol

– Autodiscovery

9

Autodiscovery

• Can be implemented using ICMP echo (PING) to every possible address

– When device answers record details

– Advantage: will discover every working device on network

– Disadvantage: wasted bandwidth and time querying non-existent devices

10

Autodiscovery

• Could also find one device and then query it to discover what other devices it has communicated with recently

– All network devices discovered in a breadth first search manner

– Advantage: works quickly

– Disadvantage: May fail to find a device that has not communicated with the network recently

• Can also help produce a graphical map of the current network

11

Autodiscovery

12

Autodiscovery

13

Modifying data

• Once configuration management information has been collected it will usually need to be updated and maintained– With a 5000 node network even if 1% of those machines

required a change once a week that would be 50 changes per week

• Addresses are only ONE of the parameters that need to be tracked

• Manual system are inefficient and error prone– What happens if an engineer forgets to record the changes

– Thames vs Thomas• Configuration management systems can record these

changes automatically

14

15

16

17

Storing Data

• Configuration management should also provide information storage

– Centralised storage provide the network engineer with efficent access to configuration data

18

Storing Data• Currently a DBMS is considered the most efficient

manner to store this data– Advantages

• Enables large amounts of data to be stored on a single computer

• Fast searching

• Automatic sorting of data

• Restoration of lost data

• Data relationships

– Disadvantages• Complex administration procedures

• May us its own language

• Tends to be platform or OS specific (lack of portability)

19

Configuration Management on a Network Management System• Simple

– Provide central storage of network information

• Network addresses

• Serial numbers

• Physical location

– Include an autodiscovery device for finding all devices on network

20

Configuration Management on a Network Management System• More Complex

– Automatic comparison of current to stored configuration

• Automatic update???

– Automapping of network

– Graphical display of configurations

• Physical

• Logical

21

Configuration Management on a Network Management System• Advanced

– Use a relational database

– Automatic generation of reports

• Inventory reports

• Change reports

• Vendor control reports

• Warranty reports

• Repair reports

– Provide for evaluation

• Duplication of addresses

• Duplication of function

22

Generating Configuration Reports• Reports should typically include

– For devices• Name

• Network address

• Serial number

• Manufacturer

• Operating system

• Local person responsible

• Could also list– Circuit numbers

– Vendor contact name

– Physical location

23

Generating Configuration Reports• Reports should typically include

– Summary of recent changes

• Listed by category

• Include

– Who made the changes

– When changes were made

• Categories could include

– New devices

– Replaced devices

– Modified devices

24

Generating Configuration Reports• Reports should typically include

– Summary of network inventory

• All details from device report

• Length and type of warranty

• Complete upgrade history

• Service history …