[tor-dev] Why the seeming anticorrelation between obfs3 and vanilla bridges in metrics graphs?

David Fifield david at bamsoftware.com
Fri Nov 7 06:44:58 UTC 2014


On Wed, Nov 05, 2014 at 12:17:50PM +0100, Karsten Loesing wrote:
> On 01/11/14 01:44, David Fifield wrote:
> > This might be the key to the mystery. It must be that PacificSunset,
> > despite being an obfs3 bridge, doesn't have ExtORPort enabled, so all
> > its obfs3 connections are being counted as <OR> connections.
> 
> Hi David,
> 
> sorry for not replying earlier and for not replying in more detail.
> 
> I just wanted to say that your analysis looks plausible to me.  Fixing
> PacificSunset's ExtORPort might indeed fix the metrics graphs.

Here's one more graph on the subject. The shading in the background
indicates when PacificSunset was running (blue) and not running (white).
It looks like the periods of highest correlation come while it is
running.

I got the uptime information from
https://onionoo.torproject.org/uptime?search=pacificsunset.

David Fifield
-------------- next part --------------
A non-text attachment was scrubbed...
Name: obfs3-or-pacificsunset.png
Type: image/png
Size: 41851 bytes
Desc: not available
URL: <http://lists.torproject.org/pipermail/tor-dev/attachments/20141106/8cbc796f/attachment-0001.png>
-------------- next part --------------
library(ggplot2)

uptime = data.frame(
	date=seq(from=as.Date("2013-11-04"), to=as.Date("2014-11-05"), by="2 days"),
	frac=c(999,957,0,20,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,520,0,228,874,353,999,999,999,457,0,0,0,0,0,853,999,999,999,749,749,0,457,999,999,999,999,999,999,999,NA,999,999,999,478,0,0,0,185,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,478,0,0,NA,NA,NA,NA,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,166,208,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,104,395,999,999,999,790,541,0,0,686,999,999,999,999,999,999,999,999,999,999,999,166,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,20,0,0,0,0,66)/999
)

c <- read.csv("clients.csv", colClasses=c(date="Date"))

p <- ggplot(c[c$date >= as.Date("2014-04-01") & c$date < as.Date("2014-11-01") & c$node=="bridge" & (c$transport=="obfs3" | c$transport=="<OR>"), ])
p <- p + geom_rect(data=uptime, aes(xmin=uptime$date, xmax=uptime$date+2, ymin=-Inf, ymax=Inf), fill="lightblue", alpha=uptime$frac)
p <- p + geom_line(aes(date, clients, color=transport))
p <- p + xlim(c(as.Date("2014-04-01"), as.Date("2014-11-01")))
p <- p + theme_bw()
ggsave("obfs3-or-pacificsunset.png", p, width=4.8, height=3.2, dpi=120)


More information about the tor-dev mailing list