Given more than a few choices it is nearly paralyzing to make a good decision.  This becomes more challenging in infotech because there are almost always a dozen ways to accomplish any technical outcome.  Applied to the design of a small network the engineer is, from minute one, at odds with themselves.  I try, even when examining very large networks, to maintain the KISS principle whenever possible and focus on modularizing components whenever possible.  Start at the “top” and work your way to the enduser and then work your way back to the gateway(s) checking for redundancies and loops.  Although it is never done without consideration I like to use nontraditional components for routing – maybe a particular location has a need that could be fulfilled by a BSD or Linux based routing device – but a modular approach allows you to replace a proprietary device with something perhaps less conventional and, depending on your experience level, you can collect and analyze performance differences and apply that at some later date.

I’ve been analyzing a smaller network for a few days, accumulating data, and trying to simplify unnecessarily complex segments.  The inter-relatedness of all things, however, slows the process and I find myself repairing things as I go.  Given a clean slate I wonder what decisions I’d make…

I’ve almost completed the HTPC – I completed the assembly, installed my OS, patched everything, updated all my drivers, updated all software after running Secunia’s vulnerability scanner, and started messing with the network.  ‘Mo’s wireless access points are starting to pay their way – after installing dd-wrt on each of three and setting up 2 as repeaters for my wireless segment – my wireless network extends about three times it’s former distance.  Of course this also raises questions about how I’m authenticating and whether I want to finally incorporate RADIUS or TACACS on the network.  I don’t think I do.  Soon I’ll be recording television shows while I watch other television shows – the manchild geek’s dream.

Fall is a good time for projects – it gets colder, damp, and the leaves change – and if I don’t have a project I succumb to seasonal nostalgia.  Stefani wants to see the Etsy offices so I feel like I’m going to be owing Chris a favor soon.

Also I’m well aware I owe some project updates – I build a couple routers and wrote up a comparison of JunOS, Untangle, Smoothwall Express, and a basic BSD router running openBGP.

Advertisement