This is the second major update of my overview of the top System Security conferences.

CURRENT VERSION: 3.0.2
(major version changes when I add new stats or new venues, minor if stats are modified, build number keeps track of typos and spellings mistakes)

WHAT’s NEW
  • Updated information, now including all papers published in 2017

  • Tier-two conferences: I received many requests to extend the work to other security conferences beyond the four tier-one venues. However, it is not easy to find long-lasting tier2 conferences in system security.. For the moment, I started importing ACSAC and RAID (2000 to 2017).

  • A new authors table that simply reports papers count.
    The old authors page with all stats for the Top4 only is still available here

  • The affiliations table (in all its forms) and the institution pages now also include Tier2 information.

WARNING

This post contains only an update of the plots and few lines of comment to point out things that have changed from last year. If you did not read the previous Circus posts, please go and check the last version first to have more information of what I am doing and a discussion of the current trends.

WARNING

Unless clearly marked, I re-plotted all graphs and stats for the Top4 only.
If you are curious about Tier2 conferences, you can find some comments about the differences at the bottom of this page.

Community Size and Number of Submissions


Overall, the entire dataset now includes 108 venues, for a total of 4128 papers written by 6754 unique authors from 972 affiliations located in 55 different countries.

Let’s first look at the submission graphs (Top4 only, you can find the Tier2 ones at the bottom):

./submission_tot.png ./acceptance_rate.png

Nothing new here.
2017 confirmed the exponential increase in the number of submitted papers and the effort of the top4 to maintain a constant acceptance rate despite it.

./new_auth.png ./active.png ./new_affl.png

Also the community graphs are in line with the last trends. In particular, the number of new authors is increasing very fast.

./authors_x_paper.png ./single_affl.png

The average number of authors per paper continues to slowly increase.
This year we also have a new record in the number of authors for a single paper. The previous record belonged to "Manufacturing Compromise: The Emergence of Exploit-as-a-Service" with 18 authors from six different institutions. Now let’s welcome the new giant: "Understanding the Mirai Botnet" with 19 authors from seven different affiliations!

At the other end of the spectrum, there were only two single-author papers in 2017.

A bit of Geography

./us_vs_world.png ./country_bars.png

China is coming up fast, passing Canada and Switzerland and taking now fifth place.
Italy is now out of the top ten, replaced by the Netherlands.

Program Committees


./committees.png ./tpc_overlap.png

Good news from these graphs. NDSS managed to go back to a more healthy paper-per-tpc-member ratio and both Usenix and CCS are back from their spikes last year in the overlap graph. Good job everyone!

Also, the program chairs managed to fill some gaps and now only one researcher in the top-100 never served in any committee.

Authors


Rank Name Papers Active Max Chair TPC Slams Venues Top4 Co-Authors Avg

1

Christopher Kruegel

69

02-17

4

1

20

3

41

All

111

5.29

2

Dawn Song

66

00-17

4

0

9

3

45

All

119

4.23

3

Giovanni Vigna

60

02-17

4

2

20

2

38

All

99

5.42

4

Wenke Lee

58

01-17

3

2

31

6

42

All

103

4.88

5

XiaoFeng Wang

54

03-17

5

0

16

3

33

All

108

5.65

6

Michael Backes

48

03-17

5

1

17

2

30

All

83

4.06

7

Vern Paxson

46

00-17

3

2

17

2

34

All

113

5.63

8

Michael K. Reiter

44

01-16

4

1

26

0

30

All

63

3.59

9

Adrian Perrig

40

00-16

4

2

19

1

29

All

66

3.98

9

Dan Boneh

40

01-17

2

2

16

0

30

All

98

4.28

The complete list (with sortable columns) which includes everyone with at least two publications is available here.

Legend:

  • active : the first and last year in which the author published a paper in the circus

  • max : max number of papers in the same venue

  • chairs : number of Top-4 venues for which the person was PC Chair

  • slams : number of times in which the author had a paper in all the Top-4 conferences in the same year.

  • tpc : number of technical program committee memberships

  • venues : total number of venues in which the author had a paper

  • top4 : in how many of the Top-4 the authors had a paper (1 to 4)

  • co-authors : total number of co-authors in Top-4 publications

  • avg : average number of authors per publication

Same names in the top ten, but with differences in the order.
Sadly, I still count only 5% of women in the top-100.
Instead, the European presence increased from 13% to 18%.

./max_papers_per_person.png

Nothing new year, except more people with five papers in the same conference.

Affiliations


Rank Name Papers Coverage Size Country

1

Carnegie Mellon University

185

56.5%

180

USA

2

University of California - Berkeley

171

62.0%

136

USA

3

Microsoft Research - US

152

47.2%

118

USA

4

Georgia Institute of Technology

126

44.4%

116

USA

5

Stanford University

106

51.9%

115

USA

6

University of Maryland - College Park

94

36.1%

78

USA

7

University of California - Santa Barbara

91

45.4%

81

USA

8

University of California - San Diego

89

43.5%

92

USA

9

University of Illinois - Urbana Champaign

87

43.5%

91

USA

10

University of Michigan - Ann Arbor

86

39.8%

88

USA

(the complete list with sortable columns is available here)

Legend:

  • coverage : percentage of venues in which the institution had at least one paper

  • size : number of Top-4 authors affiliated to the institution

Same names (still all from the US) in the top10, with the only difference of Urbana Champaign loosing few positions and University of Maryland gaining two.
On this side of the Atlantic, Bochum gained the first spot in the European ranking.

./institutions_cdf.png

Groups and Collaborations


The entire graph with all authors contains now over 5.1K nodes and almost 18K edges. Over 86% (was less than 85% last year) of the authors are part of a giant connected cluster! In other words, you can move from co-author to co-author and reach almost everyone in the field. There are only eight researchers who published more than three papers and are not connected to this giant component.

If we want the main groups to emerge from the graph, we can emphasize the main clusters by limiting the analysis to authors with at least five papers in common (and a total of 10 each in the top4). This results in 115 nodes (91 in 2016), grouped in 26 clusters (22 in 2016). From the figure below you can clearly see one macro-cluster (Xiafeng + Reiter, connected to Dawn Song through Perrig and McCune). However, the largest clique remains Savage + Paxson + McCoy + Levchenko + Voelker. (big picture here)

./main_groups.png

If we zoom in even further we can finally see the most prolific collaborations in our field (minimum 10 papers together in the Top4).
Here is what it looks like:

./connected_graph_10.png

Second-Tier Venues

./tier2_submission_tot.png ./tier2_acceptance_rate.png

These graphs are interesting.
First, we can see that also Tier2 conferences are growing, but not as much as the Top4. The curve shows a liner increase rather than the exponential one we saw before. And this is strange… if the top conferences receive more and more papers, and therefore the number of rejected ones also increases exponentially, where do all these rejected papers go? Just resubmissions?

The plot on the right side (RAID in blue and ACSAC in red) show again how the PC chairs over the years were able to stabilize the respective acceptance rates, which converged to somewhere around 20%. Not bad.

./tier2_us_vs_world.png

The US-vs-World distribution is also different for Tier2 conferences, with a similar trend but a less pronounced domination of US-based institutions.

Finally, looking at individual authors, if we count the total number of papers (Top4 + Tier2) the main change to the top ranking is Thorsten Holz jumping several positions ahead. Interestingly, there are few researchers in the top twenty position of the Top4 who never published at RAID nor ACSAC. The same happens the other way around, with some authors who published 8 or 9 papers at Tier2 conferences but never in the Top4.
Check the new authors table for more info.