CONEXT

It is a great pleasure to welcome everyone to the 10th ACM International Conference on emerging Networking EXperiments and Technologies (ACM CoNEXT'14). At the close of a decade as one of the major forums for presentations and discussions of novel computing technologies that will shape the future of Internet-working, the conference travels to another continent.

The single track program including what we found the most promising submissions on computer and communication networks research at large is a reflection of commitment of all PC members to the three values of CoNEXT: recognizing farsighted excellence, promoting a broad definition of networking research, and assuring authors of submitted papers a fair and thorough review process. We hope it will bring new technical interactions and substantial novelty to the field of networking for the years to come.

• The program, 27 long papers and 10 short ones, was selected from the 187 papers we received (including 133 long and 54 short).

• The geographical diversity of our research community, with submission from authors in 25 countries, lead to 18 countries being represented by at least one author in the program. Often illrepresented as US-centric, our community is more balanced than one may have thought: although 21 papers have at least one author affiliated in North America, 19 and 5, contain one author affiliated in Europe or Pacific/Asia, respectively.

• Looked over topics, contributions in the program reflect our set of submissions: 12 papers on wireless, 10 on theory, 7 in novel application and other fields, and an expected high number 16 on papers on management, including on data-centers and SDN. Acceptance rate for all topics above are around 20%. The only exception we noticed is in systems and security, less represented in submissions and with a lower acceptance rate, ending with only 4 papers in our program.

As can be judged by these numbers, the competition has been worldwide and the selection drastic. We would like to report information on how the reviewing process was organized.

• Apart from two papers with clear content or formatting problems, each paper was formally reviewed by 3-5 reviewers, producing a full review and a binary 'accept/reject' suggestion. These reviews and suggestions have been used by the chairs and other PC members to start a discussion on the merits of the submission. The online discussion was vigorous (995 comments were posted by PC members, an average of 5 per paper) and members updated their review to reflect additional information gathered and discussed. A plenary PC meeting attended by the vast majority of PC members at SIGCOMM made final decisions on all controversial papers w.r.t. their merits and limitations.

• Multiple factors are important to select works to be presented, and we summarized them in two criteria: Impact and Technical Quality, graded on a scale from 1 to 4 with 4 being the strongest, allowing reviewers to specify what part of this paper they appreciated to justify their decisions.

• While the decision to include or not, each paper was done individually, and all papers have been considered independently of raw scores, we thought it might be important to share statistics on the 187 submissions we received

• 19 received only accept scores (of which 100% were accepted)

• 18 received a strict majority of accept (72% accepted)

• 16 received a tie (30% accepted)

• 134 ended with a strict majority of reject (0% accepted) Distributions of impact scores (1, 2, 3, 4) was (9%, 57%, 33%, 1%); technical quality (1, 2, 3, 4) was (11%, 45%, 38%, 6%).

• A minority of papers received interaction from PC members, a feature we encouraged whenever a short clarification could help reviewers assessing a particular fact about the paper. Note that this was not a rebuttal. In the end, 8 papers received such requests which were all answered promptly and used in the decision; 3 of those papers were accepted into the program. We think this was a useful approach and not one that causes excessive burden to authors and reviewers.

Following last year's CoNEXT, short papers were considered under the same term as long papers (i.e., mature work with novelty and impact), only with a more open mind towards criticism on the scope of evaluation or broadness of the topics. The acceptance ratio for long and short papers was nearly identical, and we will continue to support authors to use a variety of presentation formats while sticking to the same excellent standard of research.