Routing 2

Print version

10. Routing 2

Packet through Morse node – enlarged description.

The routing packet entering the CU carries a quadruple MORSE address – source, SRC, destination, DST, node address from where it is coming, FRM, and a node address to where it is directed, TO. The procedure of processing is represented in the diagram Block circuit of MORSE node and includes the content of the mentioned addresses.

The access module, represented in the left part of the diagram is assigned by configuration to some of the nodes. The module only accepts packets whose TO address is the same as the address of that node. From there the packet continues with only the SRC and DST addresses.

Packet incoming to the node, contains it’s target address DST. The address of immediate next step NN is determined in the node according to address DST.

The address DST is compared with address of the node (NODE ADDRESS), where the process runs. If they are equal then the packet is in the target node and will be sent, according to its type, to the user or service port (including the SRC and DST addresses) or to the service processing. If they are different, then the routing tables processing follows. Into address NN is at first inserted the address DST, which is later rewrited by address found in respective routing table.

Node block diagram

Fig. 10.1: Node block diagram

You can get the diagram in pdf format.

The packet address can be divided to the 4 parts called (from left) global, net, wide and local. The global part represents a big unit where the address belongs (country), part net determines more exactly the region inside the country (geographically or according to other points of view). Next specifying is in part wide and part local defines finally the actual address of node. At the choice of routing table the addresses DST and NODE ADDR. are compared. The parts global, net, wide and local are compared gradually. If the difference is found (for example in net part), then in the respective table (net) will be the address of next step found. The differences in lower parts (here wide and local) are then not important..

The routing table choosed in this way, can be in the CU (Communication Unit) prepared in the versions 1 to 4. This version is used, which number is written in the menu (N)odes (e)dit in the row of respective node and in the column according to table type (here net). If in this place is not the number 1 to 4 written then on the place of address NN remains the address DST and the packet will be sent directly to this address. This connection will be successful only rarely, because the node having address DST is located usually out of the area of direct audibility.

The routing table contains the pairs (p)ath and (n)ode. The (p)ath represents the compared part of the address (here net) and (n)ode is address of node NN where leads the next packet step on it’s way to DST address. The tables wide and local contains in (n)ode the lower address part only. The item (n)ode is the resultant information found in the routing tables and is inserted into address NN. More this situations can happen in the routing tables:

  • In the local table there are not displayed the equal items, e.g. (p)ath 5E (n)ode 5E. This we can employ at removing of the unhelpful items in this way, that we put in (n)ode the same item like is in the (p)ath.

  • It is not necessary to put in the local table the trivial path that is the ways to nodes, which are directly audible and which address is different in part local only. (In this case is to=dest and (n)ode=(p)ath, see last note)

  • In the tables global, net and wide are not displayed the items, where is (n)ode = 0. In this way we can remove the unhelpful items by putting 0 in the item (n)ode.

  • The next result is, that the zero items are not used in the net and wide parts. In the global part only there are used the addresses beginning by 00 for the zero node N0 in each CU. These addresses (production numbers) are unique and unchangeable for each CU produced. The are created in different hierarchy then standard addresses and are not suitable for the work with routing tables. However it is possible create the connections via this addresses using the Path packets.

  • If the table doesn’t contain the item (p)ath equal to compared part of address, then the packet is lost. The exception see next article.

  • In the tables global and net there it is possible to use the item (p)ath, which is equal to respective part of address of node, in which the process runs (NODE ADDR). Then the relevant item (n)ode is the default address, where are sent the packets, which haven’t own item (p)ath in the table.

The packet equipped now by the NN address goes to the next processing according to diagram “Block circuit of MORSE node”. If the NN address equals to NODE ADDR, then the packet is sent to the link output L. The actual routing of the link output is written in the menu (N)ode (e)dit, in the column L. Some of serial channels or output to the Ethernet is used. The link output is used above all for the connection between two CU working on different frequencies. The way of packets can go through nets with different frequencies and can be fully described in routing tables. On the other end of wire line there is next node, which accomplish the assignment of next packet step according to address DST, which the packet carries all time.

In next step is the address NN compared with addresses of other nodes in own CU. If the concordance is found, then the packet is sent through inner bus to the respective node. It is not necessary to have other setting in menu Ne or SIe. In this node there begins the packet processing from start point again include using (different) set of routing tables.

Before sending to the net output the setting of flag link/net to user addr is checked. It is contained in the menu Ne in column M (Multiaddressing). If it is not set, which is the common case, then is the column M empty. By its setting the latter N appears here and then all packets, originally going to the net output, are sent to the user output. The similar opportunity is to change routing of packets going originally to link retranslation output to the user output. Both options are used in special cases of routing only.

If the packet is no diverted to the link output nor to inner bus nor by multiaddressing, then it is sent to the net retranslation output N. It could be the radio or ethernet output assigned in menu Ne in column N.

Behind the network or link output of the node the output part of the access module continues in the connected communication channel. Here the packet (amongst other things) is again given a quadruple address. SRC and DST addresses remain unchanged, the value NN is inserted into the TO address, and the address of the node, in which the routing process took place, is (generally) written into the FRM address.

In this way the packet is sent to the net medium and for it’s next processing takes charge this node, which address is equal to TO address contained in the packet.

This description is the more detailed version of the article Routing 1. The example showing the routing possibilities can be found in the article Routing examples 3.