The Second International Workshop on NS-2

2008/10/25

On Thursday I attended the second international workshop on ns-2 which took place at the hotel Amarilia. Since the main conference lounge was reserved for VALUETOOLS, the workshop presentations were given at a smaller room. Some quick observations:

  • Why does every small conference room in Greek Hotels have paintings of small fishing boats?
  • People! Please rehearse your presentation in the language you are going to give it! You must never run out of time!
  • Please find a way to block internet access where the actual sessions take place. How would you like it to be giving a presentation, only to see that your audience just bangs away on their keyboards instead of paying (or trying to pay) attention to what you are saying?

I particularly liked the following presentations:

  • Simulation of wireless multi-* networks in ns-2 by Laurent Paquereau and Bjarne Helvik.

    This was judged as the best paper of the workshop. If you live and die by ns-2 code, you must read it.

  • Implementation of an IPv6 stack for NS-3 by Julien Montavont, Sebastien Vincent and Nicolas Montavont.

    Very interesting work! It is going to be merged in version 3.3 of ns-3. You can see more at http://ns3v6.enstb.fr.

  • XAV: A Tracing Framework for Exploring Large Network Simulation Outputs by Ryad Ben-El-Kezadri, Guy Pujolle and Farouk Kamoun.

    Oh how I wish I had known of XAV last year when I was madly scripting awk and perl over gigabytes of ns-2 traces! I particularly liked the idea of using a data cube in order to do the trace analysis, instead of the normal ns-2 trace files. And although I am not a big fan of columnar databases, I find it rather interesting to use MonetDB as the database backend (for more columnar database propaganda go to the Database Column).

Tutorial on NS-3

This is the real reason I attended the workshop. The tutorial was given by Dr. George Riley. He is an excellent presenter and it is a pity that this tutorial was not taped. The presentation slides cover all the main points of his talk, so I want to focus on some peripheral stuff that I found interesting:

  • In the words of George Riley himself: “When I reached 45, both me and my wife watched each other in the eyes and decided that we needed something else professionally. So we both enrolled in PhD programs, got our PhDs and joined the academia”. Silence full of admiration covered the room.
  • ns-3 is not the same thing as ns-2. There is no Tcl, nor are any Tcl bindings planned (most of the audience seemed to like this fact). In ns-3, the simulations are programmed in C++, although one of the lead developers has contributed bindings in Python. I think this is an opportunity for the SmallTalk crowd to continue in the spirit of SmalTalk-80: The Language* and provide hooks for ns-3 (Part III of the book describes simulation classes and ns-3 is a discrete event simulator).
  • There is no compatibility between ns-2 and ns-3. NS-3 is more of a spiritual successor of ns-2.
  • There is no nam for ns-3.
  • The simulator’s core is very robust and generic. At one time they even thought of releasing it separately.
  • ns-3 development is going to be funded for two more years by the NSF. However one of the full time developers is fully sponsored by INRIA.
  • Different simulation platforms give different results even for the same scenarios. One has to be careful and read On the Accuracy of MANET Simulators (.ppt slides here).
  • When reading comments in ns-2 code be careful. There are lots of comments that are really older than the code beneath them, sometimes to the point of being irrelevant. Yes people when you patch / change / improve code, you have to patch / change / improve comments too. At the very same moment, not later because you will not. And lots of people will bang their heads on the wall because they will believe the comment first. Hey! Even no comments is better than irrelevant comments.

From what I saw and heard, if your simulation involves protocols already implemented in ns-2, go ahead and use ns-2. If on the other hand you are going to implement some new functionality or protocol that you are willing to study, ns-3 is a much more promising platform.

As usual, the most interesting stuff in conferences and workshops happens during the breaks. When you bump into old friends, exchange news and ideas, learning about interesting tools like scapy, VNUML, PRIME and books like Analytical Network and System Administration: Managing Human-Computer Systems.

All in all, attending this workshop was money well spent.


[*] – You can download a copy of a previous edition of the book from the ACM classics.

About these ads

Υποβολή απάντησης

Εισάγετε τα παρακάτω στοιχεία ή επιλέξτε ένα εικονίδιο για να συνδεθείτε:

WordPress.com Logo

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό WordPress.com. Log Out / Αλλαγή )

Twitter picture

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Twitter. Log Out / Αλλαγή )

Facebook photo

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Facebook. Log Out / Αλλαγή )

Google+ photo

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Google+. Log Out / Αλλαγή )

Σύνδεση με %s

Follow

Get every new post delivered to your Inbox.

Join 1,671 other followers

%d bloggers like this: