syncthing/man/syncthing-device-ids.7

267 lines
10 KiB
Groff
Raw Normal View History

2015-05-30 11:05:37 +00:00
.\" Man page generated from reStructuredText.
.
.TH "SYNCTHING-DEVICE-IDS" "7" "Sep 02, 2018" "v0.14" "Syncthing"
2015-05-30 11:05:37 +00:00
.SH NAME
syncthing-device-ids \- Understanding Device IDs
.
.nr rst2man-indent-level 0
.
.de1 rstReportMargin
\\$1 \\n[an-margin]
level \\n[rst2man-indent-level]
level margin: \\n[rst2man-indent\\n[rst2man-indent-level]]
-
\\n[rst2man-indent0]
\\n[rst2man-indent1]
\\n[rst2man-indent2]
..
.de1 INDENT
.\" .rstReportMargin pre:
. RS \\$1
. nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin]
. nr rst2man-indent-level +1
.\" .rstReportMargin post:
..
.de UNINDENT
. RE
.\" indent \\n[an-margin]
.\" old: \\n[rst2man-indent\\n[rst2man-indent-level]]
.nr rst2man-indent-level -1
.\" new: \\n[rst2man-indent\\n[rst2man-indent-level]]
.in \\n[rst2man-indent\\n[rst2man-indent-level]]u
..
.sp
Every device is identified by a device ID. The device ID is used for address
2017-11-29 06:45:17 +00:00
resolution, authentication and authorization. The term “device ID” could
interchangeably have been “key ID” since the device ID is a direct property of
2015-05-30 11:05:37 +00:00
the public key in use.
.SH KEYS
.sp
To understand device IDs we need to look at the underlying mechanisms. At first
2015-09-27 20:31:19 +00:00
startup, Syncthing will create a public/private keypair.
2015-05-30 11:05:37 +00:00
.sp
2018-02-28 06:45:16 +00:00
Currently this is a 384 bit ECDSA key (3072 bit RSA prior to v0.12.5,
which is what is used as an example in this article). The keys are saved in
the form of the private key (\fBkey.pem\fP) and a self signed certificate
(\fBcert.pem\fP). The self signing part doesnt actually add any security or
functionality as far as Syncthing is concerned but it enables the use of the
keys in a standard TLS exchange.
2015-05-30 11:05:37 +00:00
.sp
The typical certificate will look something like this, inspected with
\fBopenssl x509\fP:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
Certificate:
Data:
Version: 3 (0x2)
Serial Number: 0 (0x0)
Signature Algorithm: sha1WithRSAEncryption
Issuer: CN=syncthing
Validity
Not Before: Mar 30 21:10:52 2014 GMT
Not After : Dec 31 23:59:59 2049 GMT
Subject: CN=syncthing
Subject Public Key Info:
Public Key Algorithm: rsaEncryption
RSA Public Key: (3072 bit)
Modulus (3072 bit):
00:da:83:8a:c0:95:af:0a:42:af:43:74:65:29:f2:
30:e3:b9:12:d2:6b:70:93:da:0b:7b:8a:1e:e5:79:
...
99:09:4c:a9:7b:ba:4a:6a:8b:3b:e6:e7:c7:2c:00:
90:aa:bc:ad:94:e7:80:95:d2:1b
Exponent: 65537 (0x10001)
X509v3 extensions:
X509v3 Key Usage: critical
Digital Signature, Key Encipherment
X509v3 Extended Key Usage:
TLS Web Server Authentication, TLS Web Client Authentication
X509v3 Basic Constraints: critical
CA:FALSE
Signature Algorithm: sha1WithRSAEncryption
68:72:43:8b:83:61:09:68:f0:ef:f0:43:b7:30:a6:73:1e:a8:
d9:24:6c:2d:b4:bc:c9:e8:3e:0b:1e:3c:cc:7a:b2:c8:f1:1d:
...
88:7e:e2:61:aa:4c:02:e3:64:b0:da:70:3a:cd:1c:3d:86:db:
df:54:b9:4e:be:1b
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We can see here that the certificate is little more than a container for the
public key; the serial number is zero and the Issuer and Subject are both
2017-11-29 06:45:17 +00:00
“syncthing” where a qualified name might otherwise be expected.
2015-05-30 11:05:37 +00:00
.sp
An advanced user could replace the \fBkey.pem\fP and \fBcert.pem\fP files with a
keypair generated directly by the \fBopenssl\fP utility or other mechanism.
.SH DEVICE IDS
.sp
To form a device ID the SHA\-256 hash of the certificate data in DER form is
calculated. This means the hash covers all information under the
\fBCertificate:\fP section above.
.sp
2015-09-27 20:31:19 +00:00
The hashing results in a 256 bit hash which we encode using base32. Base32
encodes five bits per character so we need 256 / 5 = 51.2 characters to encode
2015-05-30 11:05:37 +00:00
the device ID. This becomes 52 characters in practice, but 52 characters of
2015-09-27 20:31:19 +00:00
base32 would decode to 260 bits which is not a whole number of bytes. The
2015-05-30 11:05:37 +00:00
base32 encoding adds padding to 280 bits (the next multiple of both 5 and 8
2015-06-07 10:57:26 +00:00
bits) so the resulting ID looks something like:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
MFZWI3DBONSGYYLTMRWGC43ENRQXGZDMMFZWI3DBONSGYYLTMRWA====
.ft P
.fi
.UNINDENT
.UNINDENT
2015-05-30 11:05:37 +00:00
.sp
2015-09-27 20:31:19 +00:00
The padding (\fB====\fP) is stripped away, the device ID split into four
2015-05-30 11:05:37 +00:00
groups, and \fI\%check
2016-12-26 13:23:55 +00:00
digits\fP <\fBhttps://forum.syncthing.net/t/v0-9-0-new-node-id-format/478\fP>
2015-05-30 11:05:37 +00:00
are added for each group. For presentation purposes the device ID is
grouped with dashes, resulting in the final value:
2015-06-07 10:57:26 +00:00
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
2015-06-14 11:52:00 +00:00
MFZWI3D\-BONSGYC\-YLTMRWG\-C43ENR5\-QXGZDMM\-FZWI3DP\-BONSGYY\-LTMRWAD
2015-06-07 10:57:26 +00:00
.ft P
.fi
.UNINDENT
.UNINDENT
2015-05-30 11:05:37 +00:00
.SS Connection Establishment
.sp
2017-11-29 06:45:17 +00:00
Now we know what device IDs are, heres how they are used in Syncthing. When
2015-09-27 20:31:19 +00:00
you add a device ID to the configuration, Syncthing will attempt to
2015-05-30 11:05:37 +00:00
connect to that device. The first thing we need to do is figure out the IP and
2015-09-27 20:31:19 +00:00
port to connect to. There are three possibilities here:
2015-05-30 11:05:37 +00:00
.INDENT 0.0
.IP \(bu 2
The IP and port can be set statically in the configuration. The IP
2015-09-27 20:31:19 +00:00
can equally well be a host name, so if you have a static IP or a
2015-05-30 11:05:37 +00:00
dynamic DNS setup this might be a good option.
.IP \(bu 2
Using local discovery, if enabled. Every Syncthing instance on a LAN
periodically broadcasts information about itself (device ID, address,
2017-11-29 06:45:17 +00:00
port number). If weve seen one of these broadcasts for a given
device ID thats where we try to connect.
2015-05-30 11:05:37 +00:00
.IP \(bu 2
Using global discovery, if enabled. Every Syncthing instance
announces itself to the global discovery service (device ID and
external port number \- the internal address is not announced to the
2017-11-29 06:45:17 +00:00
global server). If we dont have a static address and havent seen
2015-05-30 11:05:37 +00:00
any local announcements the global discovery server will be queried
for an address.
.UNINDENT
.sp
2015-09-27 20:31:19 +00:00
Once we have an address and port a TCP connection is established and a TLS
2015-05-30 11:05:37 +00:00
handshake performed. As part of the handshake both devices present their
certificates. Once the handshake has completed and the peer certificate is
2015-09-27 20:31:19 +00:00
known, the following steps are performed:
2015-05-30 11:05:37 +00:00
.INDENT 0.0
.IP 1. 3
2015-09-27 20:31:19 +00:00
Calculate the remote device ID by processing the received certificate as above.
2015-05-30 11:05:37 +00:00
.IP 2. 3
Weed out a few possible misconfigurations \- i.e. if the device ID is
that of the local device or of a device we already have an active
connection to. Drop the connection in these cases.
.IP 3. 3
Verify the remote device ID against the configuration. If it is not a
device ID we are expecting to talk to, drop the connection.
.IP 4. 3
Verify the certificate \fBCommonName\fP against the configuration. By
default, we expect it to be \fBsyncthing\fP, but when using custom
certificates this can be changed.
.IP 5. 3
If everything checks out so far, accept the connection.
.UNINDENT
.SH AN ASIDE ABOUT COLLISIONS
.sp
The SHA\-256 hash is cryptographically collision resistant. This means
that there is no way that we know of to create two different messages
with the same hash.
.sp
2017-11-29 06:45:17 +00:00
You can argue that of course there are collisions \- theres an infinite
2015-09-27 20:31:19 +00:00
amount of inputs and a finite amount of outputs \- so by definition there
2015-05-30 11:05:37 +00:00
are infinitely many messages that result in the same hash.
.sp
2017-11-29 06:45:17 +00:00
Im going to quote \fI\%stack
2017-02-05 12:54:49 +00:00
overflow\fP <\fBhttps://stackoverflow.com/questions/4014090/is-it-safe-to-ignore-the-possibility-of-sha-collisions-in-practice\fP>
2015-05-30 11:05:37 +00:00
here:
.INDENT 0.0
.INDENT 3.5
The usual answer goes thus: what is the probability that a rogue
asteroid crashes on Earth within the next second, obliterating
civilization\-as\-we\- know\-it, and killing off a few billion people ?
It can be argued that any unlucky event with a probability lower
than that is not actually very important.
.sp
2017-11-29 06:45:17 +00:00
If we have a “perfect” hash function with output size n, and we have
2015-05-30 11:05:37 +00:00
p messages to hash (individual message length is not important),
then probability of collision is about p2/2n+1 (this is an
2017-11-29 06:45:17 +00:00
approximation which is valid for “small” p, i.e. substantially
2015-05-30 11:05:37 +00:00
smaller than 2n/2). For instance, with SHA\-256 (n=256) and one
billion messages (p=10^9) then the probability is about 4.3*10^\-60.
.sp
A mass\-murderer space rock happens about once every 30 million years
on average. This leads to a probability of such an event occurring
2017-11-29 06:45:17 +00:00
in the next second to about 10^\-15. Thats 45 orders of magnitude
2015-05-30 11:05:37 +00:00
more probable than the SHA\-256 collision. Briefly stated, if you
find SHA\-256 collisions scary then your priorities are wrong.
.UNINDENT
.UNINDENT
.sp
2017-11-29 06:45:17 +00:00
Its also worth noting that the property of SHA\-256 that we are using is not
2015-09-27 20:31:19 +00:00
simply collision resistance but resistance to a preimage attack, i.e. even if
2017-11-29 06:45:17 +00:00
you can find two messages that result in a hash collision that doesnt help you
2015-05-30 11:05:37 +00:00
attack Syncthing (or TLS in general). You need to create a message that hashes
2017-11-29 06:45:17 +00:00
to exactly the hash that my certificate already has or you wont get in.
2015-05-30 11:05:37 +00:00
.sp
2017-11-29 06:45:17 +00:00
Note also that its not good enough to find a random blob of bits that happen to
2015-09-27 20:31:19 +00:00
have the same hash as my certificate. You need to create a valid DER\-encoded,
2015-05-30 11:05:37 +00:00
signed certificate that has the same hash as mine. The difficulty of this is
staggeringly far beyond the already staggering difficulty of finding a SHA\-256
collision.
.SH PROBLEMS AND VULNERABILITIES
.sp
As far as I know, these are the issues or potential issues with the
above mechanism.
.SS Discovery Spoofing
.sp
2018-02-07 06:45:17 +00:00
Currently, the local discovery mechanism isnt protected by crypto. This
means that any device can in theory announce itself for any device ID and
potentially receive connections for that device from the local network.
2015-05-30 11:05:37 +00:00
.SS Long Device IDs are Painful
.sp
2017-11-29 06:45:17 +00:00
Its a mouthful to read over the phone, annoying to type into an SMS or even
2015-05-30 11:05:37 +00:00
into a computer. And it needs to be done twice, once for each side.
.sp
2017-11-29 06:45:17 +00:00
This isnt a vulnerability as such, but a user experience problem. There are
2015-05-30 11:05:37 +00:00
various possible solutions:
.INDENT 0.0
.IP \(bu 2
2017-11-29 06:45:17 +00:00
Use shorter device IDs with verification based on the full ID (“You
2015-05-30 11:05:37 +00:00
entered MFZWI3; I found and connected to a device with the ID
MFZWI3\-DBONSG\-YYLTMR\-WGC43E\-NRQXGZ\-DMMFZW\-I3DBON\-SGYYLT\-MRWA, please
2017-11-29 06:45:17 +00:00
confirm that this is correct”).
2015-05-30 11:05:37 +00:00
.IP \(bu 2
Use shorter device IDs with an out of band authentication, a la
Bluetooth pairing. You enter a one time PIN into Syncthing and give
that PIN plus a short device ID to another user. On initial connect,
both sides verify that the other knows the correct PIN before
accepting the connection.
.UNINDENT
.SH AUTHOR
The Syncthing Authors
.SH COPYRIGHT
2018-02-07 06:45:17 +00:00
2014-2018, The Syncthing Authors
2015-05-30 11:05:37 +00:00
.\" Generated by docutils manpage writer.
.