User's Manual

Doc. No.

Rev.

Date

DRAFT

1.01

2012-01-27

Damm Cellular Systems A/S, Denmark

TetraFlex

7.5 Manual - IP Backbone network layout and

configuration

TetraFlex® 7.5

Manual

2-151

2.16 IP BACKBONE NETWORK LAYOUT AND CONFIGURATION

2.16.1 Multi node router considerations

From release 7.40 it is deemed mandatory to implement routers for realizing the

interconnection. In fact it is not in relation to changes introduced in 7.40 that the routers have

been mandatory. In fact the nature of multicast – when connected over switched networks

(layer 2) can make problems with broadcast like behavior of the traffic. Meaning that a

network which is shared with other users may be flooded with traffic from TetraFlex nodes

While saying that routers are mandatory – it is also a fact, that the rendezvous point (RP)

router is mandatory. It may lead to the misconception that RP is single point of failure. In the

basic configuration of the router network this might be correct – as the RP is sitting as hub for

all other routers, i.e. all traffic passes here. However – there are router techniques available –

making it possible to introduce load sharing and redundancy setup for RP’s.

Recommendations for this are under preparation. Cisco acronyms for this are MSDP

(multicast source discovery protocol) and Anycast RP.

2.16.2 Network layout principle

As opposed to the virtual network between the TetraFlex

sites – which in its nature is a

mesh network, the routers connecting the sites are connected in a star type network. The

connections through the internet, which the routers use as their backbone, are all radiating

like spokes in a wheel from one router that is announced as the rendezvous point.

The routers then facilitate mesh type communication around the network, through routing

tables built inside the network.

In reality the multisite Tetra network from Damm Cellular Systems A/S is two or more release

7 TetraFlex

nodes – working together through an IP network. All parties are equal and no

node is single point of failure.

The subscriber relevant information is contained in two registers. Inside the network, these

registers are distributed and present on all nodes at all times. The distribution of the registers

is done through multicast routing in the connecting router network. With respect to the

registers and housekeeping of the same – a node is pointed out to contain the master mode

checksum mismatch on one or more nodes, the master register will initiate a download and

overwrite of the relevant register(s).