boxley wrote:
Come on Rick have you ever, never met a client who refused to listen to you because of cost and you could not make them understand?
What an utterly bizarre, non-sequitur question.
Of course I have. What's your point?
I'll even give you a sort of example. (There are probably ones that fit more closely, but this one comes immediately to mind.) One client was a California architectural engineering firm with branch offices in Costa Mesa, Portland, and Phoenix. Each office had a LAN that was NetWare-based with an increasing number of NT 4 fileservers. There were WAN links among the offices -- but, every week or so, the routing tables would slowly become badly screwed up, and stay that way until all the routers had be reset. They asked my advice:
Me: "Put in static routes."
Mgmt: "But that would mean we'd have to have someone revise the routing tables whenever we bring onboard a new branch office."
Me: "That's true. But your routing would work."
Mgmt: "But we don't like having to pay someone to revise routes for us."
Me: "I understand that. But having propagation of bogus routes is more expensive."
Mgmt: "But the current system should work."
Me: "No, it can't."
Mgmt: "What do you mean? It used to work."
Me: "Yes, but you're using RIP, and RIP doesn't scale. You've grown to exactly the point where it starts failing. Put in routers that'll do OSPF, or use static routes. The latter is cheaper, and 100% effective."
Mgmt: "But writing routing tables is difficult."
Me: "No, it's not. Here are all the ones you need, free. I did them on the way here."
Mgmt: "We're not convinced RIP doesn't work. Surely there's a way to make it happen."
Me: "Good luck. I'm giving it to you as straight as I know it."
So, they paid an expensive consulting firm to study the problem. A week and a visitation by a squadron of network engineers later, management got a glossy folio report that said: "Put in static routes. RIP doesn't scale."
I guess I didn't charge enough.
Rick Moen
rick@linuxmafia.com