Statistics on KubeCon 2018 North America
I just attended KubeCon 2018 North America which was hosted in Seattle. It was a great experience full of interesting sessions and interesting people. It was the highest attended KubeCon to date as shown in the following chart that was presented during the keynote by Liz Rice of Aqua security.
The number of attendees was nearly double that of the Kubecon 2017 North America conference that took place in Austin Texas last year. It was almost double the Europe and Shanghai KubeCon this year. This is a good sign as shows the growing interest with Kubernetes especially in North America.
The dates of the KubeCon event was officially from the 10–13 December (however there were additional registration only events on the 9 and 14 December i.e. co-located events). During the event there was a large number of sessions to go to as well as a large number of speakers.
Out of curiosity I decided to figure out some statistics concerning the event i.e. distribution of the number of speakers, types of events, number of jobs on the jobs board etc. Thus this blog is a presentation of my findings.
Speakers
One of the first data points that I was interested in was the distribution of speakers. The data for this analysis was scraped from the speakers page(s) on the KubeCon schedule website. The data from the site suggested there were 453 speakers from 167 unique companies. Next I was curious about the distribution of speakers, i.e. which company had the highest number of speakers. The following pie chart shows the distribution of speakers across the companies:
It can be seen that 102 companies only had 1 speaker at the event. However, one company had 81 speakers i.e. 17.8 % of all the speakers at the conference.
Companies with the most speakers
In order to drill down further, I decided to focus on the companies that had more than 6 speakers, as it would not have been easy to visualize a large number of speakers. The data from the filtered result is as follows:
As could have been guessed, Google had the most number of speakers at the conference coming in at 81 speakers. This makes sense as Kubernetes is based off the Google Borg project, hence Google would probably have the most familiarity with the technology and therefore more developers would be involved with Kubernetes. My hypothesis was confirmed by the contributors chart by companies hosted on Stackalytics (The graphic has been provided below for easy reference).
Red Hat followed with 40 speakers which is less than half of Google and follows the same trend as shown in Stackalytics. This is possibly due to its involvement in OpenShift which is a platform as a service based on Kubernetes.
It seems like most of the big cloud providers had sent more than 15 speakers to the conference i.e. Google, Microsoft and IBM. However, I was surprised that Amazon didn’t have a high number of speakers at the conference considering how competitive this space is becoming. I have 2 guesses about this. The first is after looking at Stackalytics, it can be seen that Amazon is not a big contributor of Kubernetes so possibly doesn’t have as much expertise in the project as the other big companies (again it’s just a guess), my second guess is perhaps the close time proximity to Amazon ReInvent (which was from 25 — Nov. 30) and Kubecon (from 10–13 Dec) which limits the availability of speakers.
Speakers per talk
I then investigated the number of speakers per talk to see how the speakers were distributed.
Majority i.e. 94.8 % of the sessions had 1 or 2 speakers, and only a small subset had multiple speakers. On further analysis these were found to be mostly panel sessions and tutorials.
Conference Events
There were about 459 speaker type sessions i.e. sessions excluding events such as registration, breaks, social activities etc. The breakdown for the number of sessions was as follows:
There seems to be a somewhat even distribution across the sessions, with the exception of the maintainers track (21.1 % of the sessions). This is because it didn’t cover only one type of technology but included intros and deep dives into the different aspects of Kubernetes. Most of the time there was more than one event occurring in parallel, which made choosing an event to go to potentially challenging.
Event time distribution
The next key metric I was interested in was the time distribution across the sessions. This is shown in the following pie chart.
As can be seen above, majority of the sessions i.e. 93 %, took less than 90 minutes. The minority that took over 90 minutes i.e. 7 % of the sessions represented summits and sessions which occurred outside the official KubeCon conference time i.e co-located events and required additional registration. The highest number of sessions (i.e. 290), took 35 minutes where the attendees could learn the specifics behind a technology.
Event time aggregate
The aggregate time for all the sessions in the conference was approximately 408.85 hours. I found that the total time taken by sessions less than 90 minutes was 214.35 hours, and greater than 90 minutes was 194.5 hours. This was quite surprising, however one needs to realize that the minority of the sessions took quite a long time to complete i.e. the longest event was OpenShift Commons Gathering by Red Hat, which was scheduled for about 12 hours (I doubt this was actually 12 hours of talk time).
Diversity Scholarship Recipients
As I was a recipient of the diversity scholarship, I was interested in how many attendees also received it. The data for this was not easy to come by, but I reached out to the linux foundation and they were happy enough to provide it. For the conference, CNCF, in partnership with member companies Aspen Mesh, MongoDB, Twistlock, Two Sigma, and VMware offered 147 scholarships which translated to a $ 220 500 in travel stipends at $ 1 500 per attendee, and $ 80 850 in complementary conference tickets assumed to be at $ 550 per ticket, resulting in a total of $ 301 350 for all the recipients. This brings the total scholarship recipients to about 485 for the past 2 years.
Other information that was provided by linux foundation was:
- 11 percent of all attendees were women
- 14 percent of all speakers were women
- 40 percent of all keynote speakers were women
Jobs Board
One of the things that caught my eye was the jobs whiteboard near the sponsors showcase i.e:
There seemed to be a large number of jobs posted which is an indication of the shortage of developers in the industry. However, I was curious as to how many jobs was actually posted on the board. Therefore, I added them to a spreadsheet and counted them all, as a result I found there were approximately 142 jobs posted. It’s only an approximation as some of the writing was not legible. Of the 142 jobs, I found that 38 of them had some sort of email address where an applicant could post a resume to. I also tried to map the names to a website, out of all of them, I found that only 4 of them did not have a website, perhaps due to the fact that the names weren’t legible so was hard to track them online.
As a bonus, I posted the jobs listing on my github.io page i.e. (https://gitumarkk.github.io/kubecon2018jobs/) for anyone who is interested. However, I redacted the emails to protect the privacy of the people who posted the jobs.
Conclusion
All in all it was a great conference and hope to meet you at the next KubeCon.
If you like this post, don’t forget to like and/or recommend it. You can find me on Twitter as @MarkGituma.
