D-Link 2542B – Modem, 4-port router. Crashed on first night of download tests… 20% into the first file in queue. Punched random babies in face. Possibly the most messy UI and configuration interface I’ve ever seen (this includes the shit UIs that I code sometimes). On the bright side I saved 1 plug point vs my old setup of SDSL modem + 3Com router. EDIT: Forgot to mention that accessing the router web UI felt similar to accessing a poorly coded and very resource intensive website… running on a 386 SX 16MHz. Everything just felt like it was running on something with 30+ load averages.
Linksys AM300 – Modem. Nothing horrible… just impossible to configure. Nothing to do with Linksys though… apparently TIME engineers thought it would be funny to plug in an Aztech DSL605E somewhere on their switch configured to 192.168.1.1. Modem also refuses to allow configuration without DSL link plugged in (which leads to above problem). Stabbed random babies.
SMC Barricade (long model number that nobody bothers to remember) – Modem, 4-port router. A possible challenger to most horrible to configure UI… right after the previously mentioned D-Link. Couldn’t even get this bugger to activate the DSL connection. Dropped random babies from 16th floor lift shaft.
Aztech DSL605E+Buffalo WHR-HP-G54 (DD-WRT) – Standalone Modem, WiFi+4-port router. Most recent setup… and appears to be going to stay. Seems stable after about 12 hours of testing, DD-WRT only gave me problems initially with DHCP (apparently someone else used the router before and decided my box quadbiatch0 would be called subashini-PC). Buffalo AP/router refuses to clock at 250MHz, not an issue though, at least it didn’t hit 10.0 load and cause latency to 192.168.1.1 to reach almost 100ms when I was loading up the router admin page like the D-Link. Obviously stuff like that pisses me off because I hate having high WoW ping.
Moral of the story: Cheap thing no good, good thing no cheap.
Side note: Average of 260ms to WoW server. Average of 43ms latency to the TIME gateway. Anything above 10ms to the first hop outside the internal network is not acceptable IMHO… it’s plain ridiculous. On my pre-NetLynx connection, 43ms would have gotten me half-way to fucking Hong Kong. Verbal Molotov cocktails are incoming to a TIME engineer near you within the next few days if something doesn’t go right.