UCI Housing Network and Cable TV Issues

It has been 3 1/2 years since the last update to the Perl script to log into CCA.
A kind soul recently fixed it to work with the current CCA installation at UCI Resnet.
Grab it here.
As always, know what you are doing and use at your own risk.

Way back in 2005, I wrote about UCI Resnet's completely brain-dead instructions to "set up" a router (actually, it would result in a complete screw up of the router.)

Obviously, they haven't learned from that episode. They are now requiring the same brain-dead setup again.
Like 6 years ago, this will be a complete failure, and will only result in lots and lots of problems, including rogue DHCP servers, each time such a misconfigured router is reset.
I see that in 6 years, UCI Resnet has not changed one bit. They are still bloody amateurs who should never be let near a network.
He who doesn't learn from history is damned to repeat it...

What makes this even more baffling is that in 2005, Ted Roberge of Resnet himself stated in a White Paper that this kind of router misconfiguration doesn't work...

Rogue DHCP servers. Although Resnet originally prohibited routers, this became an ?unenforceable? policy with the popularity of small home wireless routers. A combination of events caused many routers to be set up incorrectly, resulting in them becoming Rogue DHCP Servers. In an attempt to help residents set up their routers and still be able to register their computers, Resnet published guidelines on how to set them up as Access Points. This became a double edged sword. When configured correctly, they worked great. However, any reset, power failure, or disconnect resulted in the device defaulting back to its original configuration and then became a DHCP server. We removed all guidelines for configuring wireless routers and simply state that they must be set up in accordance with manufacturer instructions. We further went on to state that we would not support any routers, and removed all policies relating to the prohibition of routers on the network. We will not respond to any trouble call relating to home routers or wireless.

To sum it up: Never ever connect a LAN port to the wall plug. Always connect the WAN port to the wall plug. It is that simple.

Apparently, UCI Resnet doesn't understand their own network...
They seem to have some network performance issues, and instead of actually getting people who understand networks, they decided to just cap the bandwidth for everybody...
Needless to say, that didn't go well with a lot of people, and they got flooded with complaints.
Their standard answer was this boilerplate stuff:

Due to the connection instability experienced by various housing complexes, we have decided to focus first on providing a reliable and consistent connection speed on the internet, then secondly about providing higher speeds on a stable connection. We are currently working on the stability and reliability portion. As a result, the speed cap has been set at 2Mbp/s, and will likely remain in place for the foreseeable future.

But it seems the pressure from the complaints got too big. So, today they emailed this out:

Due to the numerous concerns for bandwidth restrictions, we have decided to go ahead and release the cap temporarily to see its effect on our network. If you notice any degradation in network stability, please let us know.

He speaks the oath as I write this.
What a great moment!

Recently, I have learned that CCA apparently can not detect certain configurations of the Windows Internet Connection Sharing system.
If the Internet Connection Sharing is working in NAT mode, only IP address translation is done, and multiple network card identification numbers, called MAC addresses, are visible to the outside world. That is detectable by CCA.
However, if the Internet Connection Sharing is working in Bridge mode, each shared machine gets its own IP address, and CCA can't detect that.
The CCA client agent, which is installed on the client Windows machine (no client agent for Linux...) can detect the presence of an Connection Sharing registry entry, but apparently, that registry entry exists even if connection sharing is switched off. That nicely demonstrates the futility of this whole idea of having a program run on the client computers, testing these things. Cisco always has to reverse-engineer all programs that they want to check for, and such reverse-engineering, aside from possibly being illegal under the DMCA, is highly ineffectual. All it does is provide for job security for Cisco engineers (granted, that counts for something in today's economy )