Web 2.0 tools can help in following a particular field for new publications. In addition to using Google Scholar to find new papers by monitoring which recent papers cite a certain seminal paper, RSS can help to monitor the field. As Daniel Lemire urged researchers to make their publications available through RSS in 2005, other services offer RSS feeds now. I try to give a brief overview on some that I use the most.

Preprints – arXiv

An archive for preprints of computer science papers (and other fields) which are not peer-reviewed is provided by arXiv, funded by Cornell University and the National Science Foundation. Currently, arXiv hosts more than half-million articles. Due to its popularity, it is worthwhile to follow submissions to categories of personal interest, e.g. Networking and Internet Architecture. More categories are provided at the home page of arXiv.

However, while it is painful to visit interesting categories frequently to follow new submissions made, it can be made faily easy by using RSS along with a feedreder. The whole procedure is described here. Example feed: Networking and Internet Architectures.

Journals – IEEE Transactions

Even the IEEE provides an RSS feed for papers in ther recent issues(e.g. Transactions on Multimedia or Transactions on Networking), which makes it very easy to monitor high-impact journals.

By keyword – CiteULike

A by keyword search can be monitored using CiteULike, which is a Web 2.0 service for reference sharing. An example can be found for QoE here (see the RSS feed here).

Depending on where one would purchase a Rolex watch or other, typically high priced, products of famous brands, the price can vary by orders of magnitude. The reason for this observation can be explained by the existence of cheap replicas, illegal copies of the product. Do such copies exist only in the materialized world, or can they be a matter in academia?

Yes, they can! Today, a colleague of mine discovered that his 2005 Sigmetrics paper [1] has been plagiarised. The copy of his paper has been presented in the 8th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2009), held from June 1-3, 2009 in Shanghai, China. The authors are with the “electronics and information department, Huazhong University of Science and Technology, Wuhan, China”.
Who are the authors? Liu Wei is an associate professor. Wenqing Cheng has some history of published papers in recent years and is a professor as well. They should have known the rules. Profiles of the remaining authors could not be found.

Besides the publication of the copied paper in the conference proceedings by IEEE, the copied version of the paper has been selected as an outstanding paper and published by Springer in an additional volume (see [2]). In case you don’t have an ACM or Springer subscription, that would allow you to download the original papers, I created excerpts of the first page of each paper (find there here: original, copy). The paper, including figures, has been copied word by word. The copy has been shortened, typesetted in Microsoft Word instead of LaTeX and adapted to the new page layout. Funny side note: The reference were changed, e.g. [7] has been added as new reference that did not appear in original version of the paper.

If you start looking for other papers by the authors, you might find [3]. To have some more fun, copy some phrases of the abstract and look them up using Google, which will lead to a result like this, showing parts of a book [4] published in 2007. The abstract has been copied almost word by word from two chapters of the book (first sentence from page 327 and the remainder from page 373). Only minor things have been changed: “In this chapter” reads now “In this paper”.

What a disgrace! Or, a very special way of saying that a paper or book is good …

(We have have taken initial action in those cases. Is there anyone willing to review other papers of those authors for possible cases of plagiarism?)

Update (2009-10-01): I checked another paper: The abstract of [5] has been taken from [6, Chapter 8.4, Page 221] (I can’t download the entire paper due to a lack of subscription). You can use Google Booksearch to verify this. I already mailed to the affected authors and editors.

[1] Original: Florin Ciucu, Almut Burchard and Jörg Liebeherr: “A network service curve approach for the stochastic analysis of networks” (2005)

[2] Copy: Deah J. Kadhim, Saba Q. Jobbar, Wei Liu, and Wenqing Cheng: “The Stochastic Network Calculus Methodology” (2009)

[3] Copy: Nawaf Hadhal Kamil, Deah J. Kadhim, Wei Liu, Wenqing Cheng, “Signal Processing Techniques for Robust Spectrum Sensing,” fcc, pp.120-123, 2009 ETP International Conference on Future Computer and Communication, 2009

[4] Original: Fitzek F H P., Katz M., “Cognitive Wireless Networks: Concepts, Methodologies and Visions Inspiring the Age of Enlightenment of Wireless Communications”, ISBN 978-1-4020-5978-0, 1st: Springer, pp. 714 , 2007
[5] Copy: Deah J. Kadhim, Wei Liu, Wenqing Cheng, “Ultra Wideband Cognitive Network Objective Issues,” fcc, pp.35-38, 2009 ETP International Conference on Future Computer and Communication, 2009
[6] Original: Hossain, E., and V. K. Bhargava (Eds.), Cognitive Wireless Communications Networks, Springer Publication, 2007

I arrived at the 21st International Teletraffic Congress held in Paris. Some numbers from the opening:

• 10 borderline papers accepted (out of 24)
• 48 out of 48 papers accepted from the accepted papers category
• acceptance rate 1/3
• TPC members:
  • 60 from Europe
  • 26 from France
  • 31 from USA/Canada
  • 21 from Asia/Pacific
  • 37 from Industy / 98 from academia
• 135 Attendees
  • 55 from Europe
  • 49 from France
  • 18 from USA/Canada
  • 10 from Asia/Pacific
  • 1 from South-America
  • 1 from Uruguay

Auditorium

I have spend the last weeks in the States and, among other things, attended ACM Sigmetrics 2009 held in Seattle to give a talk about my paper and visited Stanford University, located south of San Francisco.

Sigmetrics presented a technical program of high quality and was supplemented by some very good workshops.  I had the feeling that the technical program was a bit more practical this year, but I may be biased, especially as I know of some good theoretical submissions that have been rejected. All of the talks presented at the main conference have been recorded, and, those whose authors permit a publication will be available on the web soon (similar to Sigcomm last year). I personally like the availability of conference recordings. Unfortunately, this is very rare, still, so I hope this will be catching on!

Lunch

Regarding to some side-aspects of the conference like quality and quantity of lunch, one may think of feeling the economic down-turn directly Although the banquet was a very good event. Unfortunately, it’s common in some cultures to leave directly after having dinner and switch to another place, so it was harder to catch up and many possible discussions “abruptly” ended.

There is one point that keeps on puzzling me. I believe that it is a clear benefit for the conference to have a best presentation award, as this gives the speaker the incentive to prepare a good presentation that is easy to follow for the audience (see Rob’s talk at Sigcomm last year as an example. He created an own Python presentation framework for the sake of improving the presentation). I understand the point that selecting a best paper during the curse of the conference is not feasible for the audience, as that would practically mean everyone who is interested in voting would have to read every paper. So the best paper is selected by a small set of PC members, and, thus might be biased by their own believes. However, I don’t understand why it is not possible to let every attendee vote for the best presentation. One would listen to them anyways and the additional overhead for the voting is negligible.

Poster Session

In this years Sigmetrics conference, only travel grant holders were permitted to vote for the best presentation award, which excludes the majority of the attendees. Why is the major part of the audience excluded? Just to keep the organizational overhead in collecting the votes low? What does a best presentation award mean if only a minority is allowed to vote? Sceptically speaking, is it as meaningless as the best paper award? I believe this years best presentation award goes to the right person, but the concept should still be reconsidered.

Demo Session

One of my major criticism about Sigmetrics was the demo session. Sigmetrics considers itself as one of the flagship conferences, so this should not happen to this kind of conferences, I believe.  From what I heard from people presenting demos, the session was highly unorganized. When I entered the room, three “demonstrations” where shown, where only one demo was what I would consider being a demo, namely showing some program / hardware. All of the other demos where plain PowerPoint presentations, so in fact only talks. Funny enough, the best demo award goes to a power point talk. Moreover, Sigmetrics didn’t provide a popper setting for presenting the demos; some chairs were located around three tables where the presenters placed their laptops. There was no projector installed (the auditorium could have been used instead of the room where we used to have lunch), nor a possibility to show posters explaining the technical details of the presented program. As even local tier-3 conferences have much better demo sessions, I consider this session as a major flop of this years conference. Hope there will be a better one next year, that is more seriously organized.

Besides these two points of cristism, Sigmetrics 2009 was a very good conference and I enjoyed attending.

Our paper submitted to the 21st International Teletraffic Congress (ITC 21) has been accepted for publication. Abstract:

The migration of voice communication from the Public Switched Telephone Network to the Internet pushes the need to adequately size network resources such as buffers and capacity. This paper addresses the problem of how these resources should be scaled in the number of voice flows N in order to guarantee predefined packet loss probabilities and end-to-end delays. By deriving non-asymptotic buffer overflow probabilities at both edge and interior network nodes, the paper demonstrates that O(1) buffers are sufficient to ensure probabilistic packet loss constraints at all utilizations. Also, by deriving end-to-end delay bounds, the paper shows that the required per-flow capacities decrease as O(1/N) when probabilistic end-to-end delay guarantees are sought. Numerical examples illustrate that statistical multiplexing dominates the effect of scheduling in multi-nodes scenarios with high capacities.

My journal paper entitled Stochastic Packet Loss Model to Evaluate QoE Impairments that appeared in issue 1 / 2009 of the PIK journal is now online.

The venue is finaly over and I’m back in Berlin since a couple of days. Some final thoughts:

• I may be biased, but, it appeared to me that the focus of this years KiVS moved from presenting “blueprints” and considerations on layer 21 down to the network layer again by also taking into account more theoretical work, which is a good thing.
• For the first time, they introduced a software award consisting of two categories; the awarded the best software written by a student and the best software written by a group. Matt Welch recently pointed out that application papers and software in general get much less credit then they should and are less regarded than “hard” research. From this perspective, I like the idea of having a software award in order to make these outcomes of research more visible within the community.
• I like the idea of putting proceedings on an USB memory stick drive. Unfortunately, KiVS only did this for workshop proceedings, which have been published in an open access journal in adition.
• The KiVS badge included a ticked allowing unrestricted usage of the public transportation in the time when the venue was held. While attaching USB stick drives becomes more common, I’m not aware of any other venue having something similar. (Although I also learned than, up to a certain distance, walking might be faster than using trams and other means of public transportation)

At the KiVS 2009 conference, I received the Master Thesis Award from the communication in distributed systems (KuVS) group for my thesis entitled Statistical Error Model to Impair an H.264 Decoder. See a copy of the award here and retrieve the slides of my talk here. Please find below some pictures of the award session, taken by KiVS organisers.

Me giving the talk.

Professor Lars Wolf, head of the award comitee, ceremoniously presented the KuVS award to me.

The first conference day is over and was quite interesting. While the KiVS is considered to deal with more practical issues of communication in distributed systems, more theoretical work is typically presented at MMB. From this perspective it was interesting to see that the KiVS now shifted a bit and presented some more theoretical work and greatly benefits from people like professor Jens Schmitt. Jens gave an excellent tutorial on Network Calculus and held two talks. The first session, dealing with issues of wireless networks, included two excellent papers (S. ElRakabawy et al.: Practical Rate-based Congestion Control for Wireless Mesh Networks and Frei et al.: Paving the Way Towards Reactive Planar Spanner Construction in Wireless Networks). The latter combined graph theoretical considerations with practical issues in a wonderful way.

The second session contained a panel discussion entitled “Standardization and Research – How do these two fit together?“. The interesting part of this panel where the fact that people involved in standardisation bodies (3GPP, ETSI and IETF) were present and briefly introduced the basic workflows in each body of how a standard is formed. However. the discussion was quite converse; panelists involved in standardisation highlighted that researchers do benefit from standardisation by providing interesting and practically relevant problems while researchers claimed that they get recognition within the community only from refreed publications and not from standards. Moreover, they highlighted the fact that there is mainly a lack of funding for standardisation related activities, which may make them quite uninteresting.

At the end of the day, there was a software demo session where some interesting projects were presented, e.g. a video multicast framework or a secured data link layer for home networks.