I promised I would write about the new network topology I realised by using my brand new wireless router. I installed DD-WRT on my new linksys WRT54G2. Now I will show how to configure it, so you don’t need an extra switch in a common telenet configuration.Â First I will describe the old situation and explain why it is done that way. Next I will describe how I managed toÂ get the same result, using just one device and a bit of configuration magic.
So what is the old situation?
This is how telenet installs it. They provide an extra hub, to split the connection between the digibox / digicorder and the internet connection. It is necessary for the digibox/digicorder to work, that it can get an external (telenet-internal) ip address from a dedicated range (10.169.*.*). If the hub or switch would be removed, and the digibox/digicorder plugged into the router, it would get a private ip (192.168.1.*), from the routers lan-dhcp range. This is because the usual configuration of cheap router, does not allow to the split its switch-interface between internal and external ranges. It has one fixed WAN (wide area network) port, and the other ports are for the private network.
The new situation
This is my current setup. It’s almost the same, but now without a separate device for giving the digibox/digicorder its dedicated external ip address. It is now done using some configuration magic.
I used one of the fixed internal network ports to form a WAN-switch. This can be done by configuring a VLAN where this interface, and the WAN port are combined, which leaves me with only 3 internal fixed network ports. This is enough for me. One for my fixed computer, one for my XBOX360, and I even have on left unused. My other computers are connected to my private network using Wifi.
This is how you configure it using DD-WRT: It is as easy as unchecking, and then checking another checkbox, and saving the settings. (rebooting the router if it doesn’t work the first time)
Click the image to get the full size. Â The result is the same as if you would have the extra physical switch in between, only now, it is implemented as a “virtual” switch inside the router.
You only need one device, where you used to have two. This saves a UTP cable, a wall-plug, the space to put the switch and lots of energy (= money). I have not calculated how much energy my old setup consumed, and now that the router has died, it is also no longer possible, but I guess that in this case one device consumes less energy than two. If someone has a kWh meter, I will gladly measure the energy consumption. We can than compare and calculate how long it takes before the new router has paid itself back.
What would be the impact if all telenet customers (that have internet + digital tv) changed their configuration? How much energy would be saved? I have always wondered how much energy the telenet customer side setup costs in terms of energy. There are so many boxes in my closet, all plugged in, are they really all necessary?
Last week, my old wireless router died on me. Actually the power supply stopped working. It was a topcomrouter, I bought at aldi, for just under 80â‚¬ a little more than three years ago. So the warranty had only just expired for a couple of months. It had been acting strang lately, losing dns-caches and disconnecting wired clients, for no obvious reason. I was still very happy with it. It was cheap, fast (108Mbps wireless), secure and reliable. It came with a wifi-usb key too, which I rarely used. If someone has a spare 5.0V Â DC 2.0A power supply, just let me know 🙂 I used to experiment a lot with that little router. Once I made a parabolic signal director for it, boosting its range to up to 300m.
So last week, I searched for a solution. I wanted a cheap, but comprehensive solution. Something I could use to play with a bit, without spending too much money. I started reading some reviews, and found that the Linksys WRT54GL is a favourite among open source firmware enthousiasts. I found that it looked ugly and was more expensive than the WRT54G2. I found that the WRT54G2 too is supported by DD-WRT, a fully featured open source firmware. So I ordered it from routershop.nl, and two days later, yesterday, it arrived by mail.
I played with it a bit, but found that out of the box, the possibilities for configuration are rather limited. So I flahed it with the new firmware. This page details the process, which is actually rather painless. There were only two (known) problems I encountered. The first is that the tftp program for sending the new firmware to the device didn’t work under windows Vista. This was not a problem for me, I just booted into XP and everything worked as described. The second problem was that after setting a new password, I couldn’t save any of the settings I wanted to change. I did the 30/30/30 reset, which means pressing the reset button for 90 seconds, while killing and restoring the power to the device after half a minute. After that I was able to configure the device to my liking.
I was very impressed with the possibilities the DD-WRT firmware opens for such a cheap device.
In a next post, I will describe the network topology I realized with it. I used to need an extra network-switch to give an external IP to my telenet digibox/digicorder.
On 11, 12 and 13 May 2 colleagues and I went to london for the Progressive .NET exchange organized by Skills Matter. It turned out to be a very interesting conference, with many renowned speakers from the open source .net scene. The conference had 2 concurrent tracks to choose from.
On the first day I chose to go to Gojko Adzics sessions on specification by example and fitnesse. In this session Gojko presented the idea to use examples to drive requirements specification. It was a very intense workshop that used the game of Blackjack as an example to make the point. In small groups we were asked to write a full specification for the rules of Blackjack using just examples. It turned out to be very useful to use examples, as it presses on the importance of “the what” and “the why” over “the how”. In the afternoon session we implemented Fitnesse to automate the acceptance testing for the application. I had never seen Fitnesse and it was funny to see some concepts converge. The easiest way to describe Fitnesse is to say that it is a wiki with green and red bits. You write wiki pages to describe the acceptance criteria for a project, and you specify examples to back up the description. You use tables to layout the examples, which can than be automatically verified by Fitnesse. If someone is interested in what this is, and how it works, I might be able to give a short demo during one of the next KSSs.
Tuesday morning I chose the Robert Pickering session. It was on the programming language F#, a topic on which he has written a book and has extensive knowledge, but the session was more on syntax, and he’s not a very good speaker. A little bit of a disappointment. The afternoon I went to Ayendes Advanced NHibernate workshop. It was very intense. He tackled 25 topics in 4 hours, answering questions from the audience. From caching over security toward meta data. Later that night we went to the alt.net UK beers, which was a very interesting experience. It was a dynamic discussion based on topics suggested and elected by the audience, hosted in the cellar of a pub, with lots of free beer. People shouting, giving opinions and laughing. Very fun indeed.
Wednesday – the final day – I attended the sessions of David Laribee. Towards a new Architect, Lean, Kanban, Team Values, Diverging and converging brainstorming specification sessions. These sessions were very practical. In small groups we discussed team values – defined what matters most in a development team, where the priorities for all stakeholders essentially lay. We worked together to define a new product – a medical device to manage patient data and facilitate doctor interaction and his daily work flow. We did this using a “design storm”. First everyone created something individually, then we came together in small groups of five and compared the results. Made a group synthesis and presented the results. Then the entire group discussed the results and moved towards an agreed upon result. In the afternoon, we continued to build a product, but this time a site to play a fantasy soccer strategy game, based on the real soccer players and results from matches. It was a lot of fun, and the exercise tried to show how agile, lean and kanban can be useful in software planning and development.
Other interesting subjects that came up during this sessions were the pomodorotechnique, Towards new Architecture, ten usability heuristics, …
The only downside of the conference was the dodgy Internet connection. Otherwise it would have been even more enjoyable and interesting. Overall it was very instructive and lots of fun!