Day one comes to a close

It's been a long and interesting day. I learned a few things both technically and about myself.

My goals for today were to complete a baseline clonable topology for GNS3. I soon realized that a full scale topology was not needed as the individual labs in IE Workbook v1 only use a few devices each time. So far in the Frame Relay (more on that later) book I have used the same topology over and over again, but with only two routers. I found it very easy to create a base config that only included basics (such as router name, turning off domain lookup (I HATE DOMAIN LOOKUP!)) and write mem it. Config up a lab, but don't write it. When it works, just stop the routers and restart them, back to the base config in 30 seconds. So did I accomplish this goal? Technically no, but I found a more efficient way to perform labs consecutively.

The second goal was to begin with the EIGRP labs. Again, failure, but for good reason. I thought EIGRP would be a good starting ground, but I have a severe lack in knowledge in Frame Relay and several EIGRP labs require working Frame Relay. So I've started there. Was able to get through all two-router scenarios without cheating. Will complete the Frame Relay workbook tomorrow which includes 5 labs designed around Hub-and-Spoke.

What did I learn about Frame Relay? It's not as intimidating as I once thought. It's complexity lies in it's many ways to do the same thing, namely how to identify a distant router by a layer-2 address (DLCI) and then how to make a layer-2 to layer-3 mapping either statically or dynamically. Ultimately it's just an interface that can receive many different layer-2 frames from different senders. It's how you want the router to interpret those frames and make routing decisions based upon that.

As such I can see the need for both point-to-point and multipoint initially. Point-to-point makes life easier to understand as a DLCI only points to one distant router eliminating complexities such as split horizon, but required the use of exceedingly more IP addresses, even using /31 addressing. Multipoint allowed one IP address to talk to all distant routers in the cloud, but had several shortcomings, again with split horizon and OSPF concerns. Now with the use of ip unnumbered, I cannot see any valid reason to ever use mutlipoint. The main drawback, IP space, was removed and made even better as you can utilize the loopback address further reducing IP utilization.

Finally, as for personal growth information, I've found I start to go cross-eyed at the 7th to 8th hour. A quick nap and a few more labs made for an interesting, if relaxing first day. If all days go like today, I'll be ready before the end of June...

Anthony