Technology

I have always been a fan of static web pages. When I started to think about a personal web page in 2005, I ended up using NanoBlogger as my content management system since it allowed me to write blog posts in a terminal (over ssh) using my favorite text editor, it did not require much on the server side, and it was a rather cool shell script hack. The world moved on since 2005 and content management systems became really big and also quite usable (I can edit content on them easily even without my most loved browser plugin It’s All text! that allows me to edit HTML textboxes in my favorite text editor).

The Universal Serial Bus (USB) has been a great success. Almost all peripheral devices such as printers, keyboards, cameras, audio devices, disk drives, wireless interfaces can be easily connected using a standard plug. And of course, USB memory sticks have become a standard way of data exchange, replacing CDs and DVDs. But exactly here is the problem. Some companies are rightfully afraid of leaking sensitive data and with the appearance of fast small USB memory sticks, the USB interface has been identified as a problem.

I just returned from Venice (Italy), where I attended DSOM 2009 and IPOM 2009. This was my first trip to Venice and the city is more than interesting. I enjoyed going straight to the airport by boat (not necessarily fast but pretty special) and I liked the water taxi I used every morning to reach the conference site, the Telecom Italia Future Centre.

The main conference room was truly exceptional, combining Venecian history with modern meeting room technology. The audience was sitting on the left and right side of the room facing each other but not the presenter, who was located at the head of the room. Being futuristic meant in this specific case that every participant had his own screen for watching the slides and a microphone for live interaction. Despite a few feedback issues, the microphones worked reasonably well for the audience - but less so for the presenter located at the head of the room since the speakers were all directed towards the audience. But the fact that it was difficult for the presenter to understand questions was just a minor challenge - the by far bigger challenge was to deliver a lively talk in front of an audience where nobody is facing the presenter and all people are starring into computer screens (either their own notebooks or the screens showing the slides or even both). Giving a lively presentation from an office via a web conferencing system is likely easier than this setup.

I am attending the Dagstuhl Seminar on the Management of the Future Internet and as some of you might know, I also love to create new terms when I go to this kind of events. (I somehow believe that for many hype terms, the words have been found before they were given an interpretation.) While sitting in the wine cellar, I heard my mouth suddenly saying “self-destructing networks” and so I started to think what this could possibly mean.

To understand the future of the Internet (or to better speculate about it), one should know something about the history of the Internet. There are two talks online which I highly recommend:

• A New Way to Look at Networking. Van Jacobson, August 2006
• The Internet History, Development and Forecast. Leonard Kleinrock, Infocom 2006

Especially young people who cannot imagine that there have been networks different than IEEE 802 and IP are encouraged to take a look at these videos (if you find them - it seems the recording of Leonard Kleinrock’s talk has disappeared from the Internet - which is kind of ironic if you know how he opened his talk - but good that I have a private copy).

Network operators know that mistakes can happen. This is why it is good to have humans involved in network management control loops so that they can identify and resolve errors. This simple principle, unfortunately, does not seem to apply to many online systems. My recent example comes from a well known scientific publisher. A paper submitted to a network management journal ended up being registered in the publisher’s database as a submission for the Journal of Polymers and the Environment (yes, this really exists). Since then, the system periodically reminds the author to approve the publication and to transfer the copyright. (Of course, meanwhile the paper appeared in the correct journal and the copyright has already been transferred but the database has no clue about this.) The reminder email says the following:

Yesterday, I attended one of the meanwhile popular “Future Internet” workshops. I ended up in a group tasked to picture how people will use the Internet in say 10-15 years from now and what services / technologies are driving the future Internet or can be expected to be invented. We came up with many interesting ideas; one of them was centered around the question how people will find out whether they are pregnant…

Today, I learned that I am holding together with four other nice guys a US patent called Methods and Apparatus for Managing Middleware Service in a Distributed System (US Patent 6,789,114, issued September 7th 2004). I was involved in this patent application during my stay at Bell Labs in 1997 and it was just today that I learned that this patent was finally posted in 2004. I should probably go and check my bank accounts now… And yes, please make heavy use of this patent!

This afternoon, the research and education network as well as the PlanetLab machines hosted at Jacobs University went offline. Our router reported a problem with a watchdog event. We quickly brought up identical replacement hardware - but the problem did persist. After some debugging, we finally figured out that one of our PlanetLab nodes started this afternoon to send packets that trigger an obscure bug in one of the Ethernet drivers of the OpenBSD based software router we use. We finally applied a one line patch and since then everything is fine again. (Of course, this all happened while we were facing a paper submission deadline. ;-)