Assembling Etherkillers


	BOFHcam Menu
	Camera I
	Camera II
	Copyleft

Current
Designs

Assembling
Etherkillers

Black Operations in the
Corporate IT Theatre

BOFH
in a Nutshell

Distributing Clue
to Users

LART
Pocket Reference

Practical UNIX
Terrorism

Snooping Email
for Fun and Profit

Tracing Spammers

Windows NT's
Infernal Filesystem

Windows NT
User Obliteration

Why You Can't Find
Your UNIX System Administrator

Writing Word
Macro Viruses

Special
Interest

Ask BOFH

This month's Journal

Other Writing

Assembling Etherkillers

Full Description

Etherkillers. You know what they are. You may have even toyed with building one to see how easy it is. You may, like myself, have five of them hanging up behind you as a warning to any other systems administrators who might visit. It might not say "hey look at me, I'm kewl" but it might say "I do so little work I had time to put these together, which means I have time to slip out of work, come round your place and plug them into out-of-the-way ports you'd forgotten all about".

Etherkillers. The world over they have a common construction. Take one mains power cable, remove the appliance-specific connector from one end with your tool of choice. Bare the wires.

In this modern age, BNC and AUI are fading into obscurity, but we will cover BNC here for completeness. The "Baby 'N' Connector" was used to terminate coaxial "thinnet" (10base-2) ethernet cabling. A bayonet-type connector plugged into the wall socket and a suitably equipped network card and could manage well over the widely used 10Mb/sec. 10base-5 used a much thicker co-axial cable and could possibly give higher voltage potentials.

Given the nature of coaxial cable, attaching a power cable to a BNC connector, or indeed a native electricity plug to a length of coaxial (probably much likely to be commented upon than a length of power cable connected to something other than the power socket on the back of a PC) is simplicity itself. You make a choice, live to data, or shield. Countries with a ground wire should make a decision as to whether to include that in some way. I'd recommend not. It's better that way.

When you get onto twisted pair wiring, things become a little more complicated. For preference you should seek out a 3-phase power supply. If you're in the possession of some big iron (Sun Enterprise 5000 or similar), your machine room should be able to supply you with this. One suggested wiring scheme for 3-phase as suggested by Arvid is indicated below (including working comments):

     Live phase 1 <--> RJ45 pin 3  
     Live phase 2 <--> RJ45 pin 6 
     Live phase 3 <--> RJ45 pin 2     (is this suitable?)
     Neutral      <--> RJ45 pin 1     (or should we kill phones too?)
     GND <--> RJ45 pin 8

Working with AC 2-phase current leaves you with the option to only connect the four wires used for transmission and receiving. TX+, TX-, RX+ and RX- all need to be connected if you're going to do the most damage. And you are aiming to do that. No half measures here. There is the consideration that an unfused plug might melt the small wires of twisted pair. It is recommended that you use solid core, rather than stranded.

Given that not everyone subscribes to the A standard for wiring twisted pair and RJ-45 connectors the best advice that can be given here is to splice the live and neutral wires into two and connect TX+ and RX+ to live and TX- and RX- to neutral. As Arvid states, if you run your phones over the same structured wiring then connecting a "cable saver" and fully wiring your etherkiller can take out phones as well as computers. With luck you can make the phones ring before they melt.

Finally, fibre etherkillers. Little research has been done into this new area. Some progress and academic discussion by [Tanuki and Jasper] sheds some light on current thinking.

Tanuki writes:

>My current problem is the design of an Etherkiller that works
>on fibre-based Gigabit. Current prototype involves a short
>length of fibre with ST-connector one end, and a laser-pointer 
>the other: the question is, do i need to arrange for the laser-
>source to be pulse-modulated in a particularly-evil mark:space
>ratio in order to deliver sufficient intensity to permanently
>zap the remote receiver photodiode?

I doubt that a laserdiode will be able to permanently zap the receiver at all. 
Unless you go and get yourself one of them fancy .5W IR laserdiodes (single 
diodes are available up to 2-4 Watts of optical output IIRC). Added advantage: 
These wavelengths are definitely _not_ eyesafe. Be careful about reflections off 
of, say, your window though...

You could always go for a pulsed Diode-pumped Nd:YAG laser, though your lusers 
may be suspicious when they see a box at least the size of a shoebox on your 
cable.

Jasper

This T-shirt has a rich heritage grounded in a background of the ultimate parting shot from a departing systems administrator. To be used with restraint unless your time really is up, an etherkiller can become a treasured possession or a talisman for jaded burnt out operators. Wear a T-shirt that signifies if not your intentions, then your desires and mindset.

Return to Assembling Etherkillers