Yes, this is for a homework assignment (sort of, it's a graduation project). No, I don't want anyone to do the work for me. Please don't delete.
The capstone/graduation project is preferentially a real project done at your real work. I work for a big bank, change management makes ancient Sparta look liberal and laid back, so that's not happening. Making something up is an option, but they want it to be 'indistinguishable' from a real project in the deliverables you end up with.
I know what I want to do: Design a campus network from the ground up. For their purposes, it'll be a brownfield deployment replacing a flat, poorly segmented college campus network with a proper hierarchical, modular network just like the Good Book CVD documents say they should be.
In the interests of keeping things 'real', though...how often are badly designed, mostly layer 2, nightmare networks still a thing you run across anymore? I had something to the vague effect of this simplistic layer 2 ring in mind, with each switch in its own building, a couple vlans stretched across the campus and IVR done really inefficiently RoaS style from a data center in the campus 'main building' basement.
Is that realistic? Ever seen anything like this in real deployment? Ever seen anything worse? What were some of the symptoms of the poor performance you saw on your monitoring tools/complaints from endusers if you did? What were some of the consequences you saw in networks like this one/worse than this one?
I can imagine some problems (wasted bandwidth from STP blocking one direction in the ring, poor STP failover time, easily saturating links for inter-VLAN traffic that has to hairpin up to the router and back out, lack of redundancy, etc), but I was interested in hearing about real world experience because my current professional experience is in telephony, not directly in the networking world yet.
Any war stories in this genre anyone would like to share?
No comments:
Post a Comment