[Idea] Hardware broadcast by active hub.

[Rapture]

Greetings,

I work in networking for a living and have always been interested in multiboxing WoW. I had an idea pop in my head after making a USB extender out of an ethernet cable (Cat5e). For those who don't know, an ethernet cable has 4 pairs of wires and USB has 2 pair. I was able to make a 50 foot dual USB extender. So I started thinking about hubs and usb broadcasting.

There are 3 types of network hubs:

1.) Passive hub. Repeats all signals (electrical only) through all ports. Does not regenerate signal.
2.) Active hub. Same as above, but has a power input to boost signal for longer distances.
3.) Intelligient (or Smart) Hub. Same as active, but can detect CRC errors (not what we need).

I was thinking of taking a 5 port (1 in, 4 out) active hub and making a bunch of USB->Ethernet cables to see if this would be an efficient way to broadcast a keyboard & mouse (ethernet can support 2 USB cables). Anyone ever try this?

Thanx for your time.

[Rapture]

I wanted to elaborate a little more on hubs for those who aren't as knowlegeable as I am. Here is a quick into to networking:

Networking is mostly based off of a simple model which consists of 7 layers (called the OSI Model). For what we are dealing with for WoW, we are only interested in 3 of the layers (Physical, Data Link, Network). The physical layer deals strictly with electrical signals (rise or fall in voltage). Data Link deals with frames and other things, but think of the MAC address of your NIC. Network layer deals mostly with IP addresses (IPv4 and IPv6) and some other things.

So, this is all fine and dandy, what what does this mean in terms of broadcasting input devices. Some software such as Keyclone use the Network Layer (layer 3) to send information over your network. From an IT perspective this is easy but is subject to traffic on your network (as in collisions and the behavior of the Ethernet Protocol). Devices that fall under the Network layer are Smart Switches , Routers, and Wireless Routers.

The Data Link layer (Layer 2) deals with, as I mentioned, MAC addresses. This type of networking is extremely fast and would be a nice candidate for what we need. Only problem is we aren't able to differentiate which "frame" (layer 2 version of a packet) is for our USB devices. It just sends a frame from MAC A to MAC B without concerning itself with the data other than making sure it isn't corrupted in transit. Speaking of, this is an important thing because USB signals would be interpreted as bad frames and thrown away instead of passed along. Devices that fall under this category are Layer 2 Switches, "Smart"/"Intelligent" Hubs, etc.

The Physical layer deals with only electrical signals. It doesn't care what the information is, how it is formed, or even where it is going. It just gets it in one port and sends it out another (or multiple) port. This is important for this application since if, for example, we plug in a mouse into a Network Hub and move the mouse, it will send a rapid pulse of +5V signals through the hub. The hub will just replicate this +5V pulse and send it out the other ports. This happens in less than 1 millisecond. There is no delay as it replicates the info, goes to each port successivlely, and sends the signal. It will instead replicate and send it to a chip which amplifies it and sends it out all other ports at the same time. Devices that work this way are Passive/Active hubs, NICs (for some things), and transcievers (i.e. ethernet to fiber adapter).

The only problem with hubs these days is they aren't made in quite that manner. They like to call them "Workgroup Switches" or "Smart Switches" or some marketing BS like that. Alot of vendors will take a Layer 2 device (switch) and turn off some features to sell it as a hub. It still checks for errors and tries to remember what goes where and from whom. We don't need that functionality. Marketing really makes it difficult to find what you need.

I will see if I can find a cheap/free active hub and see if it works. I'd appreciate any info if someone has attempted this in the past.

[Rapture]

Bah, missed your post while typing my novel.

One arguement to your statement would be Logitech wireless mice and just buying multiple recievers for $10/each. Does the mouse handshake with each reciever or just one? Ah, maybe each reciever does the handshaking to act as a mouse... Which do you think is the case?

[Rapture]

Actually, after thinking about how layer 5 works and the simplicity of mice. I believe this would work since I'm sure the mouse doesn't remember which peer it is connected to. I think Windows just associates a usb port (1-126) to that mouse. It would be interesting to attempt.

Edit: I meant IRQ instead of USB port.

[Rapture]

I would assume the RF mice are using FSSM (Frequency Spread Spectrum) which will attempt to sync with receiver A with is hopping from frequency to frequency (similar to cordless phones) to prevent interference from similar devices. That would explain the out-of-sync issue. Bluetooth uses something similar but also does a handshake off of a MAC address which syncs device A with only receiver A.

I read up on some RFCs and Windows USB whitepapers and I believe my idea is going to work given everything is using the same resolution. Hell, I wrote some software a couple years ago in AutoIt that will allow mouse movement within a smaller window on machine A to match the movements on a larger resolution on the same window on machine B. I would use this in conjunction with a software solution like keyclone. I want to avoid that though since my network is pretty active (3 file servers, 4 clients, firewall, Cisco router, and a couple other project machines).

A tip for those who do use Keyclone (or similar) on a home network, you can setup Quality of Service (QoS) for the port/protocol the software is using to ensure traffic hits all your machine before any other traffic flows. It's like a priority list saying this traffic can take the right of way over this other traffic. They do it with Vonage/Skype hardware to ensure clear calls.

Anyways, I'm gonna start a project log here soon but I need to draw up the wiring schematics in Visio first.

[Rapture]

Test 1: Failure
Test 2: Partial Success (Split 1 mouse into 2 machines)
Test 3: Pending

So, the short answer is the hub solution won't work. I guess the P2P connection does happen for each mouse to each PC. So after testing different ideas, I came up with a way to build this. I had a couple spare crappy wireless mice sitting around that I wasn't going to use. I ripped them apart and soldered the DATA+ and DATA- from a bare USB extender to both mice and plugged them into my machines. The PCs detected the mice and when I moved the primary (3rd) mouse, it worked! The best part was that machine A and machine B were at different resolutions and the mouse movement scaled correctly!

I need to tweak some things and then figure out how to do the same thing on a keyboard. I'm sure I can score 4-5 crappy keyboards for free. The hub idea was a failure, but I did manage to find a cheap way to multiplex mice. I'll submit my project log next week sometime.