|
Posted-By: auto-faq 3.3 (Perl 5.004)
Archive-name: privacy/anon-server/faq/use/part5
Changes: 1.10 2001/03/25 14:41:27
Posting-Frequency: monthly
A list of the recent changes to the FAQ list will appear
next week.
A how-to-find-the-FAQ article appears every Wednesday.
URL: http://www.eskimo.com/~turing/remailer/FAQ/
Subject: APAS Anonymous Remailer Use [FAQ 5/8]: Statistics
This is the fifth of eight parts of a list of frequently-asked
questions and their answers regarding anonymous remailer use. This
part is all about remailer statistics (stats). It has changed to
provide a more complete list of stats sources and to indicate how
current those sources are in opinion of the maintainer.
This FAQ is provided "as is" without any express or implied
warranties. While every effort has been taken to ensure the accuracy
of the information contained in these message digests, the maintainer
assumes no responsibility for errors or omissions, or for damages
resulting from the use of the information contained herein. This FAQ
is provided for information only; reference to a Web page does not
constitute endorsement of that page's content.
The following topics are in this FAQ:
1: [FAQ 5.1] What are stats pages?
2: [FAQ 5.2] How are stats Versions 1 and 2 different?
3: [FAQ 5.3] Where can I find stats pages?
4: [FAQ 5.4] Why are there dead remailers on the stats pages?
Date: 02 March 2001 12:00 Z
From: turing+apas-user-faq@eskimo.com (Computer Cryptology)
Subject: [FAQ 5.1] What are stats pages?
Summary: Stats pages have estimates of remailer reliability and latency.
Stats pages are a snapshot of the current status of the traditional
remailer network. They list the remailers the maintainer of that stats
page is "pinging", those remailers' capabilities and limitations, and
how well those remailers are replying to those pings. From this data,
remailer uptime and latency can be deduced.
Keep in mind that the results shown on different stats pages often
vary widely and that the stat page you are looking at shows the
results as seen from the remailer doing the pinging. Network
conditions between this remailer and the others can and will influence
the results.
For example, a stats source (pinger) temporarily without mail access
will list all pinged remailers as having low reliability (assuming
that Web access is present). In fact, it is the pinger itself whose
reliability has declined. Refreshing stats from such a pinger may
result in the remailer client complaining that there are not enough
remailers.
Other things that will influence the results are not having the
current key(s) for a remailer on the pinger's keyring. Often key
changes are made and announced, but a stats page maintainer might miss
this, or his remailer isn't automatically updating with new keys and
purging old keys. If possible (i.e., if the stats source provides
them), check the keyrings.
With new remailers coming on line all the time and others departing
the network after only a brief appearance, it can be very difficult to
maintain an accurate list of exactly which remailers are really
online. Also, when a remailer joins or leaves the network, the uptime
and latency stats for that remailer will not respond instantaneously.
Stats are an analysis of data taken over several days and presented as
a moving average.
Some stats pages are in question 5.3.
Elcaro posts his Remailer Reliability Statistics daily in APAS. This
guy is a stats maniac! His stats offer:
* 1 Hop Remailer Check for a day and the last week
* MultiHop(1-4) Remailer Check for a day
* Position Check for the MultiHop Check for a day each remailer
giving Success/Failure/Total/Percentage for every position
* Position Check for the MultiHop Check Summary for the last week
* Arc Analysis for Last 7 Days Success/Failure/Total/Percentage
Date: 02 March 2001 12:00 Z
From: turing+apas-user-faq@eskimo.com (Computer Cryptology)
Subject: [FAQ 5.2] How are stats Versions 1 and 2 different?
Summary: Different clients read different stats.
There isn't a really big difference between them. Version 2 is
somewhat more detailed and uses different symbols than Version 1 to
represent the measurements of Latency, History and Uptime. Here's a
typical Version 1 stats format
<http://mixmaster.shinn.net/stats/remailer-list.html> and here's a
typical Version 2 <http://www.neuropa.net/%7Egretchen/rlist2.html> .
These are in HTML. You may also download TXT versions of each. The
file names may vary (rlist, rlist.txt, remailer-list, etc...), but
here are the recommended file names:
Filename Remailer-Type Format Stats-Version
rlist Cypherpunk Plain Text 1
rlist.html Cypherpunk HTML 1
rlist2 Cypherpunk Plain Text 2
rlist2.html Cypherpunk HTML 2
mlist Mixmaster Plain Text 1
mlist.html Mixmaster HTML 1
mlist2 Mixmaster Plain Text 2
mlist2.html Mixmaster HTML 2
Further information on stats formats comes from RProcess'
Specification and RFC for Remailer Stats Version: 2.b
<http://www.skuz.net/potatoware/PSKB-032.html>. Most users need only
consider which stats version their client will read. Here is a
summary of the recommended version. Further comments are below.
Software Recommended Version Recommended Format
JBN1 1 HTML
JBN2 2 Plain Text
Mixmaster 1 HTML?
Quicksilver 1 Plain Text
Private Idaho 1 HTML
Reliable 2 HTML
Jack B. Nymble v.2 and Reliable v.1.0.5
JBN2 and Reliable 1.0.5 can read both types of stats. Pick a format
that you like and stick with it.
Quicksilver
Quicksilver reads only Version 1 stats in TXT format (e.g., mlist.txt
and rlist.txt).
Mixmaster
The operator of Cmeclax Remailer, <cmeclax@ixazon.dynip.com>, confirms
that, as far as he can tell, "...Mixmaster 2.9 doesn't understand
Version 2 [stats]. I tried it a while ago, and it had no reliability
data when I made chains."
Jack B. Nymble v.1
Frog-Admin says that JBN1 only reads version 1 stats. In that
operator's experience, JBN1 works better with HTML files
than with plain text files.
Private Idaho
As for Private Idaho, there are lots of different variations of this
older program. But it's safe to assume that most of them require
Version 1 stats in HTML format (e.g., remailer.htm) and won't work
with Version 2.
Other Clients
Be aware that some of the older client software like Potato, Decrypt
and Mixmaster 2.0.4 cannot read Version 2 Stats.
Date: 02 March 2001 12:00 Z
From: turing+apas-user-faq@eskimo.com (Computer Cryptology)
Subject: [FAQ 5.3] Where can I find stats pages?
Summary: Various URLs may point to stats pages.
First, consider where you might find stats pages that are no longer
useful. RProcess packaged JBN2 and Reliable when different stats
pages were active. The included stats lists are no longer current.
Next, there are several indexes to stats sources (pingers). Starting
with the newest, below are the indexes available on the date shown
above.
lefarris "Les autres remailers"
<http://www.citeweb.net/arris/stats/index.htm> OR
<http://pages.globetrotter.net/arris/stats/index.htm>
weasel "Anonymous Remailer Stats, Meta-Stats and other Information"
<http://anon.noreply.org/stats/index.html>
frog "All Pingers' Index"
<http://www.privacyresources.org/frogadmin/Pingers.html> OR
<http://www.chez.com/frogadmin/Pingers.html> OR
<http://members.nbci.com/frogadmin/Pingers.html>
As of the date of this question, the following pingers (in
alphabetical order) are accurate and current or up-to-date--according
to frog (see above) and turing, at least:
austria <http://www.tahina.priv.at/~cm/stats/>
efga <http://anon.efga.org/Remailers/>
farout <http://www.nuther-planet.net/farout/stats/>
frog <http://www.privacyresources.org/frogadmin/Main.html>
helferlein
<http://www.helferlein.net/mixmaster/>
senshi <http://private.addcom.de/SenshiRemailer/>
shinn <//www.mit.edu:8001/finger?rlist@mixmaster.shinn.net">http://www.mit.edu:8001/finger?rlist@mixmaster.shinn.net>
<//www.mit.edu:8001/finger?mlist@mixmaster.shinn.net">http://www.mit.edu:8001/finger?mlist@mixmaster.shinn.net>
subzer0 <http://www.press.nu/leiurus/subzer0/>
turing <http://www.eskimo.com/~turing/remailer/stats/>
The following stats sources produce lists significantly different from
the stats sources above. The cmeclax page, for example, says "Note: I
am behind a modem, so my latency figures include my own latency as
well as those of the pinged remailers. Use these stats only to decide
what remailer to put after or before me." Consider the comments in
question 5.4 before using the others.
cmeclax <http://lexx.shinn.net/cmeclax/>
lefarris
<http://pages.globetrotter.net/arris/stats/>
publius <http://www.publius.net/>
In addition to the pages above, active remailers with stats pages
include the following that are CURRENTLY OUT OF DATE, but may someday
return because the remailer still operates:
bruble2 <http://www.angelfire.com/pe/rijto/remailer/>
gretchen
<http://www.neuropa.net/~gretchen/>
xganon <http://anon.xg.nu/list/>
Check the date of this FAQ and of the stats pages! The three above
are NOT UP-TO-DATE as of the date of this writing.
Date: 02 March 2001 12:00 Z
From: turing+apas-user-faq@eskimo.com (Computer Cryptology)
Subject: [FAQ 5.4] Why are there dead remailers on the stats pages?
Summary: Stats pages may go out of date when remailer network changes.
This happens because the keeper of that stats page is either not aware
that the dead remailers have officially left the network, or is behind
on maintaining his stats pages. Often stats pages are maintained by
remailer operators and their other remailer duties will take priority
over updating their stats pages.
Unfortunately, a poorly maintained stats site will often lead to
problems for remailer users when they plug those URLs into programs
like Jack B. Nymble, Private Idaho or QuickSilver and receive
configuration error messages. The problem will repeat itself every
time your stats are updated and the site with the stale keys is
polled.
One attempt to provide information on the problem described above is
found in Computer Cryptology's Comparison tables:
<http://www.eskimo.com/~turing/remailer/stats/db/rlist.html>
<http://www.eskimo.com/~turing/remailer/stats/db/mlist.html>
Further information is available at Frog's MetaStats Page
<http://www.privacyresources.org/frogadmin/MetaStats/index.html>.
Two comparisons between stats sources can help in selecting pingers to
include in a remailer client's list. First, the "Last update" time
indicates if the values are current. Second, a comparison of the
remailers each stats source list can indicate if the problem described
above is occurring. For example, imagine a particular stats source
lists many remailers at 0.00% reliability ("uptime"), is the only
pinger to list several remailers, and doesn't list a dozen remailers
that the majority of other pingers include. That stats source is
probably out of date, even if the "Last update" time is current.
For further consideration, the cells in Computer Cryptology's
Comparison shade to indicate suspected error values or outliers, i.e.,
values far from the average of other stats sources. The details are
on those pages.
Similarly, Frog's "All Pingers' Index Page"
<http://www.privacyresources.org/frogadmin/Pingers.html> rates pingers
as up to date versus poorly or not maintained. These ratings express
the opinion of the operator.
------------------------------
End of faq.5 Digest
*******************
|
Comment about this article, ask questions, or add new information about this topic: