TMN (Telecommunications Management Network) is an ITU standard which defines
a Network Management Architecture. (see also OMNIPoint, X.700, etc).
Originally most standards defined the interface to the customer, network management
standards define internal workings of the network. The effect should be to allow
us and our competitors to mix network elements from different manufacturers
and buy off the shelf management systems to manage them all. In other words
to be systems integrators and to add value on top of that.
The TMN is the architecture currently accepted by the industry and therefore
if we want to buy off the shelf management systems then we have to use this
TMN Architecture - Segmentation
It seems to me that Business Management Systems are less closely coupled to
the network and more linked to the business process such as HOMS, FMS, NESS
and NBS. If we loose these systems there may not be an immediate effect on customers
unless they have a fault or wish to change their service. The middle column
are more tightly coupled to the physical network and more realtime, such as
polling, monitoring alarms, displaying network status, checking customers configuration,
monitoring protocols, loading configuration changes, etc.
Examples of Business Management issues:
- How can we reduce our product development interval?
- What can we do to improve customer satisfaction?
- How can we take unit cost and still do more?st out of the business?
- Are we making the most of our capital assets?
Billing & Reconciliation (no usage billing yet)
Support for Helpdesk, Trouble Ticket, etc
In some cases router management is done by individually configuring and faultfinding each box. This is very inefficient and needs lots of highly skilled people with a detailed knowledge of the network.
Examples of Element Management issues:
- Whats the errored seconds count?
- Why haven't the routing tables been updated?
- Why is the switch blocking calls to that area?
- What's the transaction rate threshold for this processor?
- Alarms and polling
This is the level of management that we have on our X.25 networks, although this is based on proprietary and DIY systems.
Examples of Network Management issues:
- Have we got adequate capacity to handle traffic from point A to point B?
- Can we recover in the event that we loose that switch? Are all the routing changes made to show the new number range?
- How should we alleviate congestion?
- Configuration and loading
- Alarms/polling Event correlation
- Design & Simulation Systems
- Automatic Diagnosis
- Network Inventory & software releases
- Security management
This it the management of the end-to-end service regardless of what networks
the service goes through and needs a knowledge of virtual network such as what
terminals connect to what hosts. Customer stats, monitoring QOS and contract
Examples of Service Management issues:
- Are we meeting our service level agreements with our customers?
- Customer stats
- Can we deal with on-line requests for additional capacity?
- In the event of a failure, can we reconfigure to put highest-priority customers back in service?
- Do we know about a problem before our customers call us?
- Can we investigate end-to-end problems ? where both ends can PING correctly but there are high level protocol problems?
- Do we have inventory of equipment used for this customer?
- Can we see a map of the customes virtual network?
- What Terminals connect to what host?
- Customer Network Management
- SNMP feed
- Customer Reports
- Service Level Agreement monitoring
- On-line subscription changes/management