Getting Technical

For technical information about FidoNet please see www.ftsc.org

Routing ERN, A Concept Driven View
21 April 2001
By Carol Shenkenberger

What is ERN and why is it so important to us all? Simply put, ERN is `Echomail Routed Netmail' and means to all that use it, that your netmail can be picked up from the same place you receive your echomail and files from, if someone desires to route a `private' message to you. Not all zones use ERN much, and some use it in a different pattern and don't call it `ERN'. Thats ok. The `concept' is the same. Some zones use a chart to maintain it (I know Z1 and Z6 do).

To start the process, lets go back a bit in history. At one time, in Fidonet, everyone got their mail from their `hub' (See Policy4 for a description if you've entered after this system was used- `Assistants to the NC'). The biggest `Hub' for mail delivery was called the NEC for Network Echomail Coordinator. That same path may or may not have been used to get your netmail to you. Some netsheld it at the NC's and the nodes would all poll the NC to collect it, or the NC would deliver it.

The NEC's would all poll the REC (Region Echomail Coordinator) and pickup all the echmail for the net. Some smart ones realized that Netmail for that net could also be `packed with the echomail' for delivery. In the early days, it wasnt called ERN.

The term ERN started to be used, when sysops started shifting the pattern to draw their echomail for their net, from sites other than the REC. George Peace, John Souvestre, and others were a part of this pattern.

The Sysop who was feeding another net, simply told their mail tossers to 'pack any netmail for zone#:net#/* to that net'.

To use it though, people had to know 'who connected to who' so they could pass the netmail the right way, ending up at the destination sysop's site.

That took the development of various different `chart' systems to track it. Initially, these were created by each REC for showing the routes (connections) for each net in their region.

Today, REC's still make these charts in Z1. In regions with no REC, the RC makes them. (My apologies to Z6, your method is slightly different and you seem to fit into a later set of this post and when I get there, I'll show what I've gathered is done).

In Z2, they do have a system but I am not able to address it fully here, because I lack sufficient information. I know they have a standard regional number used in most regions for `inbound routing' and that there are private charts of connections. They are not disseminated the same as in Z1 or Z6. I know nothing of Z3-5 routing systems, other than they have one.

Once nets started to connect `out of the standard pattern' the various charts in Z1 changed to fit the needs of the sysops. Initially it was probably much like the Z6 chart that they use today.

The Z6 routing chart is very simplified. It works like this:

Feed address   Nets that feed from that address
6:770/3774     6:654/0  6:654/*
3:774/605      6:701/36 6:701/* 6:653/* 6:654/*
6:770/1        6:751/321 6:751/* 6:750/* 6:758/* 6:64/*
6:770/3774     3:774/605

Thats a portion as retyped by me and slightly simplified.

The first node number listed, is like what Z1 calls a `Top Level Hub' (TLH) meaning `feeds many sites'. In Z1, the master chart shows who these sites are. In Z6, there is no chart but the information is directly shown in the links.

I'll get to Z1's method shortly but Z6's is easier to conceptualize in this instance.

What happens here is netmail destined for anyone in 6:751's net, is sent to 6:770/1. They connect to him, and pick it up. If 6:64/0 wants to send a netmail to 6:751/0, they send it to 6:770/1 (their uplink for mail and files) and 6:770/1 takes it, tosses it, and packs it for pickup.

This is the same as routing inside a region via the REC. Nothing fancy there as they all have the same 'uplink'.

Now lets say someone in 6:654 wants to route a netmail to someone in 6:701? Simple. 6:654 sends it to 6:770/3774 who then packs it to 3:774/605 (See, they have a link direct). 3:774/605, when they get it, repacks it to 6:701/36 for delivery.

This is called a `hop'. This is where a chart becomes needed so that the top end `uplinks' know where to send the mail to.

In Z1, there are many formats for the regional charts. The most common one is the only one I will try to explain here.

Conceptually it is the same as in Z6, but looks quite different due to being a more complex mesh of sites.

In Z1 as of my last chart handy (Routelst.105), there are 10 TLH's (Top Level Hubs) listed in a chart that shows how they all interconnect. The format is not the same. In Z1, our charts generally start with the net number, then the number of the node who is primarily connecting, and then to whom they connect. The last node listed, is one of the 10 TLH's.

A sample would look like this: (Note, this is for teaching and the connections shown are not accurate)

Top Chart:
                 A     B    C
(A) 1:10/3       -     X
(B) 1:140/1            -
(C) 1:275/100          X    -

Note that A connects to B but not C in this example. B connects to both of them. C connects only to B.

Now the route chart might look like this (again, not using real data, just sample views)

Net      Link       Hops
1:381    1:381/18   1:275/100
1:271    1:271/1    1:2600/1 1:10/3
1:202    1:202/15   1:10/3
1:2345   1:2345/33  1:140/1
1:2346   1:2346/411 1:140/1

Note that netmail for net 271 goes through 1:2600/1 before it gets to the TLH site. The others happen to have direct links so look shorter.

Now say 1:2345/33 wants to talk to a member of 1:2346? They send the netmail to their uplink (1:140/1) who packs it in an archive for 1:2346/411 to pickup.

This is the same as routing via the REC as they both have the same uplink. Nothing fancy here.

Say however 1:2345/19 want to route a netmail to 1:271/15? They send it to their uplink (1:140/1) who packs it to 1:10/3 who knows to send it to 1:2600/1 who then sends it to 1:271/1. This involves several hops. And if 1:381/18 wants to talk to 1:202/15, they would send it to 1:275/100 who would send it to 1:140/1 who would pass it to 1:10/3 (and from there to the site).

While this system sounds complex, remember, it's no more complex than what happens when you send an echomail reply. In fact, conceptually, it's simpler!

xxcarol
AKA
ITCS(SW) Carol Shenkenberger