ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/gvpe/doc/gvpe.protocol.7
Revision: 1.16
Committed: Wed Mar 30 04:02:50 2016 UTC (8 years, 8 months ago) by root
Branch: MAIN
CVS Tags: rel-3_0, HEAD
Changes since 1.15: +2 -2 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 .\" Automatically generated by Pod::Man 2.28 (Pod::Simple 3.30)
2 .\"
3 .\" Standard preamble:
4 .\" ========================================================================
5 .de Sp \" Vertical space (when we can't use .PP)
6 .if t .sp .5v
7 .if n .sp
8 ..
9 .de Vb \" Begin verbatim text
10 .ft CW
11 .nf
12 .ne \\$1
13 ..
14 .de Ve \" End verbatim text
15 .ft R
16 .fi
17 ..
18 .\" Set up some character translations and predefined strings. \*(-- will
19 .\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
20 .\" double quote, and \*(R" will give a right double quote. \*(C+ will
21 .\" give a nicer C++. Capital omega is used to do unbreakable dashes and
22 .\" therefore won't be available. \*(C` and \*(C' expand to `' in nroff,
23 .\" nothing in troff, for use with C<>.
24 .tr \(*W-
25 .ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
26 .ie n \{\
27 . ds -- \(*W-
28 . ds PI pi
29 . if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
30 . if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
31 . ds L" ""
32 . ds R" ""
33 . ds C`
34 . ds C'
35 'br\}
36 .el\{\
37 . ds -- \|\(em\|
38 . ds PI \(*p
39 . ds L" ``
40 . ds R" ''
41 . ds C`
42 . ds C'
43 'br\}
44 .\"
45 .\" Escape single quotes in literal strings from groff's Unicode transform.
46 .ie \n(.g .ds Aq \(aq
47 .el .ds Aq '
48 .\"
49 .\" If the F register is turned on, we'll generate index entries on stderr for
50 .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
51 .\" entries marked with X<> in POD. Of course, you'll have to process the
52 .\" output yourself in some meaningful fashion.
53 .\"
54 .\" Avoid warning from groff about undefined register 'F'.
55 .de IX
56 ..
57 .nr rF 0
58 .if \n(.g .if rF .nr rF 1
59 .if (\n(rF:(\n(.g==0)) \{
60 . if \nF \{
61 . de IX
62 . tm Index:\\$1\t\\n%\t"\\$2"
63 ..
64 . if !\nF==2 \{
65 . nr % 0
66 . nr F 2
67 . \}
68 . \}
69 .\}
70 .rr rF
71 .\"
72 .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
73 .\" Fear. Run. Save yourself. No user-serviceable parts.
74 . \" fudge factors for nroff and troff
75 .if n \{\
76 . ds #H 0
77 . ds #V .8m
78 . ds #F .3m
79 . ds #[ \f1
80 . ds #] \fP
81 .\}
82 .if t \{\
83 . ds #H ((1u-(\\\\n(.fu%2u))*.13m)
84 . ds #V .6m
85 . ds #F 0
86 . ds #[ \&
87 . ds #] \&
88 .\}
89 . \" simple accents for nroff and troff
90 .if n \{\
91 . ds ' \&
92 . ds ` \&
93 . ds ^ \&
94 . ds , \&
95 . ds ~ ~
96 . ds /
97 .\}
98 .if t \{\
99 . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
100 . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
101 . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
102 . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
103 . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
104 . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
105 .\}
106 . \" troff and (daisy-wheel) nroff accents
107 .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
108 .ds 8 \h'\*(#H'\(*b\h'-\*(#H'
109 .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
110 .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
111 .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
112 .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
113 .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
114 .ds ae a\h'-(\w'a'u*4/10)'e
115 .ds Ae A\h'-(\w'A'u*4/10)'E
116 . \" corrections for vroff
117 .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
118 .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
119 . \" for low resolution devices (crt and lpr)
120 .if \n(.H>23 .if \n(.V>19 \
121 \{\
122 . ds : e
123 . ds 8 ss
124 . ds o a
125 . ds d- d\h'-1'\(ga
126 . ds D- D\h'-1'\(hy
127 . ds th \o'bp'
128 . ds Th \o'LP'
129 . ds ae ae
130 . ds Ae AE
131 .\}
132 .rm #[ #] #H #V #F C
133 .\" ========================================================================
134 .\"
135 .IX Title "GVPE.PROTOCOL 7"
136 .TH GVPE.PROTOCOL 7 "2015-10-31" "2.25" "GNU Virtual Private Ethernet"
137 .\" For nroff, turn off justification. Always turn off hyphenation; it makes
138 .\" way too many mistakes in technical documents.
139 .if n .ad l
140 .nh
141 .SH "The GNU-VPE Protocols"
142 .IX Header "The GNU-VPE Protocols"
143 .SH "Overview"
144 .IX Header "Overview"
145 \&\s-1GVPE\s0 can make use of a number of protocols. One of them is the \s-1GNU VPE\s0
146 protocol which is used to authenticate tunnels and send encrypted data
147 packets. This protocol is described in more detail the second part of this
148 document.
149 .PP
150 The first part of this document describes the transport protocols which
151 are used by \s-1GVPE\s0 to send its data packets over the network.
152 .SH "PART 1: Transport protocols"
153 .IX Header "PART 1: Transport protocols"
154 \&\s-1GVPE\s0 offers a wide range of transport protocols that can be used to
155 interchange data between nodes. Protocols differ in their overhead, speed,
156 reliability, and robustness.
157 .PP
158 The following sections describe each transport protocol in more
159 detail. They are sorted by overhead/efficiency, the most efficient
160 transport is listed first:
161 .SS "\s-1RAW IP\s0"
162 .IX Subsection "RAW IP"
163 This protocol is the best choice, performance-wise, as the minimum
164 overhead per packet is only 38 bytes.
165 .PP
166 It works by sending the \s-1VPN\s0 payload using raw \s-1IP\s0 frames (using the
167 protocol set by \f(CW\*(C`ip\-proto\*(C'\fR).
168 .PP
169 Using raw \s-1IP\s0 frames has the drawback that many firewalls block \*(L"unknown\*(R"
170 protocols, so this transport only works if you have full \s-1IP\s0 connectivity
171 between nodes.
172 .SS "\s-1ICMP\s0"
173 .IX Subsection "ICMP"
174 This protocol offers very low overhead (minimum 42 bytes), and can
175 sometimes tunnel through firewalls when other protocols can not.
176 .PP
177 It works by prepending an \s-1ICMP\s0 header with type \f(CW\*(C`icmp\-type\*(C'\fR and a code
178 of \f(CW255\fR. The default \f(CW\*(C`icmp\-type\*(C'\fR is \f(CW\*(C`echo\-reply\*(C'\fR, so the resulting
179 packets look like echo replies, which looks rather strange to network
180 administrators.
181 .PP
182 This transport should only be used if other transports (i.e. raw \s-1IP\s0) are
183 not available or undesirable (due to their overhead).
184 .SS "\s-1UDP\s0"
185 .IX Subsection "UDP"
186 This is a good general choice for the transport protocol as \s-1UDP\s0 packets
187 tunnel well through most firewalls and routers, and the overhead per
188 packet is moderate (minimum 58 bytes).
189 .PP
190 It should be used if \s-1RAW IP\s0 is not available.
191 .SS "\s-1TCP\s0"
192 .IX Subsection "TCP"
193 This protocol is a very bad choice, as it not only has high overhead (more
194 than 60 bytes), but the transport also retries on its own, which leads
195 to congestion when the link has moderate packet loss (as both the \s-1TCP\s0
196 transport and the tunneled traffic will retry, increasing congestion more
197 and more). It also has high latency and is quite inefficient.
198 .PP
199 It's only useful when tunneling through firewalls that block better
200 protocols. If a node doesn't have direct internet access but a \s-1HTTP\s0 proxy
201 that supports the \s-1CONNECT\s0 method it can be used to tunnel through a web
202 proxy. For this to work, the \f(CW\*(C`tcp\-port\*(C'\fR should be \f(CW443\fR (\f(CW\*(C`https\*(C'\fR), as
203 most proxies do not allow connections to other ports.
204 .PP
205 It is an abuse of the usage a proxy was designed for, so make sure you are
206 allowed to use it for \s-1GVPE.\s0
207 .PP
208 This protocol also has server and client sides. If the \f(CW\*(C`tcp\-port\*(C'\fR is
209 set to zero, other nodes cannot connect to this node directly. If the
210 \&\f(CW\*(C`tcp\-port\*(C'\fR is non-zero, the node can act both as a client as well as a
211 server.
212 .SS "\s-1DNS\s0"
213 .IX Subsection "DNS"
214 \&\fB\s-1WARNING:\s0\fR Parsing and generating \s-1DNS\s0 packets is rather tricky. The code
215 almost certainly contains buffer overflows and other, likely exploitable,
216 bugs. You have been warned.
217 .PP
218 This is the worst choice of transport protocol with respect to overhead
219 (overhead can be 2\-3 times higher than the transferred data), and latency
220 (which can be many seconds). Some \s-1DNS\s0 servers might not be prepared to
221 handle the traffic and drop or corrupt packets. The client also has to
222 constantly poll the server for data, so the client will constantly create
223 traffic even if it doesn't need to transport packets.
224 .PP
225 In addition, the same problems as the \s-1TCP\s0 transport also plague this
226 protocol.
227 .PP
228 Its only use is to tunnel through firewalls that do not allow direct
229 internet access. Similar to using a \s-1HTTP\s0 proxy (as the \s-1TCP\s0 transport
230 does), it uses a local \s-1DNS\s0 server/forwarder (given by the \f(CW\*(C`dns\-forw\-host\*(C'\fR
231 configuration value) as a proxy to send and receive data as a client,
232 and an \f(CW\*(C`NS\*(C'\fR record pointing to the \s-1GVPE\s0 server (as given by the
233 \&\f(CW\*(C`dns\-hostname\*(C'\fR directive).
234 .PP
235 The only good side of this protocol is that it can tunnel through most
236 firewalls mostly undetected, iff the local \s-1DNS\s0 server/forwarder is sane
237 (which is true for most routers, wireless \s-1LAN\s0 gateways and nameservers).
238 .PP
239 Fine-tuning needs to be done by editing \f(CW\*(C`src/vpn_dns.C\*(C'\fR directly.
240 .SH "PART 2: The GNU VPE protocol"
241 .IX Header "PART 2: The GNU VPE protocol"
242 This section, unfortunately, is not yet finished, although the protocol
243 is stable (until bugs in the cryptography are found, which will likely
244 completely change the following description). Nevertheless, it should give
245 you some overview over the protocol.
246 .SS "Anatomy of a \s-1VPN\s0 packet"
247 .IX Subsection "Anatomy of a VPN packet"
248 The exact layout and field lengths of a \s-1VPN\s0 packet is determined at
249 compile time and doesn't change. The same structure is used for all
250 transport protocols, be it \s-1RAWIP\s0 or \s-1TCP.\s0
251 .PP
252 .Vb 3
253 \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+
254 \& | HMAC | TYPE | SRCDST | DATA |
255 \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+
256 .Ve
257 .PP
258 The \s-1HMAC\s0 field is present in all packets, even if not used (e.g. in auth
259 request packets), in which case it is set to all zeroes. The \s-1MAC\s0 itself is
260 calculated over the \s-1TYPE, SRCDST\s0 and \s-1DATA\s0 fields in all cases.
261 .PP
262 The \s-1TYPE\s0 field is a single byte and determines the purpose of the packet
263 (e.g. \s-1RESET, COMPRESSED/UNCOMPRESSED DATA, PING, AUTH REQUEST/RESPONSE,
264 CONNECT REQUEST/INFO\s0 etc.).
265 .PP
266 \&\s-1SRCDST\s0 is a three byte field which contains the source and destination
267 node IDs (12 bits each).
268 .PP
269 The \s-1DATA\s0 portion differs between each packet type, naturally, and is the
270 only part that can be encrypted. Data packets contain more fields, as
271 shown:
272 .PP
273 .Vb 3
274 \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-+
275 \& | HMAC | TYPE | SRCDST | SEQNO | DATA |
276 \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-+
277 .Ve
278 .PP
279 \&\s-1SEQNO\s0 is a 32\-bit sequence number. It is negotiated at every connection
280 initialization and starts at some random 31 bit value. \s-1GVPE\s0 currently uses
281 a sliding window of 512 packets/sequence numbers to detect reordering,
282 duplication and replay attacks.
283 .PP
284 The encryption is done on \s-1SEQNO+DATA\s0 in \s-1CTR\s0 mode with \s-1IV\s0 generated from
285 the seqno (for \s-1AES:\s0 seqno || seqno || seqno || (u32)0), which ensures
286 uniqueness for a given key.
287 .SS "The authentication/key exchange protocol"
288 .IX Subsection "The authentication/key exchange protocol"
289 Before nodes can exchange packets, they need to establish authenticity of
290 the other side and a key. Every node has a private \s-1RSA\s0 key and the public
291 \&\s-1RSA\s0 keys of all other nodes.
292 .PP
293 When a node wants to establish a connection to another node, it sends an
294 RSA-OEAP-encrypted challenge and an \s-1ECDH \s0(curve25519) key. The other node
295 replies with its own \s-1ECDH\s0 key and a \s-1HKDF\s0 of the challenge and both \s-1ECDH\s0
296 keys to prove its identity.
297 .PP
298 The remote node enganges in exactly the same protocol. When both nodes
299 have exchanged their challenge and verified the response, they calculate a
300 cipher key and a \s-1HMAC\s0 key and start exchanging data packets.
301 .PP
302 In detail, the challenge consist of:
303 .PP
304 .Vb 1
305 \& RSA\-OAEP (SEQNO MAC CIPHER SALT EXTRA\-AUTH) ECDH1
306 .Ve
307 .PP
308 That is, it encrypts (with the public key of the remote node) an initial
309 sequence number for data packets, key material for the \s-1HMAC\s0 key, key
310 material for the cipher key, a salt used by the \s-1HKDF \s0(as shown later) and
311 some extra random bytes that are unused except for authentication. It also
312 sends the public key of a curve25519 exchange.
313 .PP
314 The remote node decrypts the \s-1RSA\s0 data, generates its own \s-1ECDH\s0 key (\s-1ECDH2\s0),
315 and replies with:
316 .PP
317 .Vb 1
318 \& HKDF\-Expand (HKDF\-Extract (ECDH2, RSA), ECDH1, AUTH_DIGEST_SIZE) ECDH2
319 .Ve
320 .PP
321 That is, it extracts from the decrypted \s-1RSA\s0 challenge, using its \s-1ECDH\s0
322 key as salt, and then expands using the requesting node's \s-1ECDH1\s0 key. The
323 resulting hash is returned as a proof that the node could decrypt the \s-1RSA\s0
324 challenge data, together with the \s-1ECDH\s0 key.
325 .PP
326 After both nodes have done this to each other, they calculate the shared
327 \&\s-1ECDH\s0 secret, cipher and \s-1HMAC\s0 keys for the session (each node generates two
328 cipher and \s-1HMAC\s0 keys, one for sending and one for receiving).
329 .PP
330 The \s-1HMAC\s0 key for sending is generated as follow:
331 .PP
332 .Vb 1
333 \& HMAC_KEY = HKDF\-Expand (HKDF\-Extract (REMOTE_SALT, MAC ECDH_SECRET), info, HMAC_MD_SIZE)
334 .Ve
335 .PP
336 It extracts from \s-1MAC\s0 and \s-1ECDH_SECRET\s0 using the \fIremote\fR \s-1SALT,\s0 then
337 expands using a static info string.
338 .PP
339 The cipher key is generated in the same way, except using the \s-1CIPHER\s0 part
340 of the original challenge.
341 .PP
342 The result of this process is to authenticate each node to the other
343 node, while exchanging keys using both \s-1RSA\s0 and \s-1ECDH,\s0 the latter providing
344 perfect forward secrecy.
345 .PP
346 The protocol has been overdesigned where this was possible without
347 increasing implementation complexity, in an attempt to protect against
348 implementation or protocol failures. For example, if the \s-1ECDH\s0 challenge
349 was found to be flawed, perfect forward secrecy would be lost, but the
350 data would likely still be protected. Likewise, standard algorithms and
351 implementations are used where possible.
352 .SS "Retrying"
353 .IX Subsection "Retrying"
354 When there is no response to an auth request, the node will send auth
355 requests in bursts with an exponential back-off. After some time it will
356 resort to \s-1PING\s0 packets, which are very small (8 bytes + protocol header)
357 and lightweight (no \s-1RSA\s0 operations required). A node that receives ping
358 requests from an unconnected peer will respond by trying to create a
359 connection.
360 .PP
361 In addition to the exponential back-off, there is a global rate-limit on
362 a per-IP base. It allows long bursts but will limit total packet rate to
363 something like one control packet every ten seconds, to avoid accidental
364 floods due to protocol problems (like a \s-1RSA\s0 key file mismatch between two
365 nodes).
366 .PP
367 The intervals between retries are limited by the \f(CW\*(C`max\-retry\*(C'\fR
368 configuration value. A node with \f(CW\*(C`connect\*(C'\fR = \f(CW\*(C`always\*(C'\fR will always retry,
369 a node with \f(CW\*(C`connect\*(C'\fR = \f(CW\*(C`ondemand\*(C'\fR will only try (and re-try) to connect
370 as long as there are packets in the queue, usually this limits the retry
371 period to \f(CW\*(C`max\-ttl\*(C'\fR seconds.
372 .PP
373 Sending packets over the \s-1VPN\s0 will reset the retry intervals as well, which
374 means as long as somebody is trying to send packets to a given node, \s-1GVPE\s0
375 will try to connect every few seconds.
376 .SS "Routing and Protocol translation"
377 .IX Subsection "Routing and Protocol translation"
378 The \s-1GVPE\s0 routing algorithm is easy: there isn't much routing to speak
379 of: When routing packets to another node, \s-1GVPE\s0 tries the following
380 options, in order:
381 .IP "If the two nodes should be able to reach each other directly (common protocol, port known), then \s-1GVPE\s0 will send the packet directly to the other node." 4
382 .IX Item "If the two nodes should be able to reach each other directly (common protocol, port known), then GVPE will send the packet directly to the other node."
383 .PD 0
384 .ie n .IP "If this isn't possible (e.g. because the node doesn't have a \*(C`hostname\*(C' or known port), but the nodes speak a common protocol and a router is available, then \s-1GVPE\s0 will ask a router to ""mediate"" between both nodes (see below)." 4
385 .el .IP "If this isn't possible (e.g. because the node doesn't have a \f(CW\*(C`hostname\*(C'\fR or known port), but the nodes speak a common protocol and a router is available, then \s-1GVPE\s0 will ask a router to ``mediate'' between both nodes (see below)." 4
386 .IX Item "If this isn't possible (e.g. because the node doesn't have a hostname or known port), but the nodes speak a common protocol and a router is available, then GVPE will ask a router to mediate between both nodes (see below)."
387 .ie n .IP "If a direct connection isn't possible (no common protocols) or forbidden (\*(C`deny\-direct\*(C') and there are any routers, then \s-1GVPE\s0 will try to send packets to the router with the highest priority that is connected already \fIand\fR is able (as specified by the config file) to connect directly to the target node." 4
388 .el .IP "If a direct connection isn't possible (no common protocols) or forbidden (\f(CW\*(C`deny\-direct\*(C'\fR) and there are any routers, then \s-1GVPE\s0 will try to send packets to the router with the highest priority that is connected already \fIand\fR is able (as specified by the config file) to connect directly to the target node." 4
389 .IX Item "If a direct connection isn't possible (no common protocols) or forbidden (deny-direct) and there are any routers, then GVPE will try to send packets to the router with the highest priority that is connected already and is able (as specified by the config file) to connect directly to the target node."
390 .IP "If no such router exists, then \s-1GVPE\s0 will simply send the packet to the node with the highest priority available." 4
391 .IX Item "If no such router exists, then GVPE will simply send the packet to the node with the highest priority available."
392 .IP "Failing all that, the packet will be dropped." 4
393 .IX Item "Failing all that, the packet will be dropped."
394 .PD
395 .PP
396 A host can usually declare itself unreachable directly by setting its
397 port number(s) to zero. It can declare other hosts as unreachable by using
398 a config-file that disables all protocols for these other hosts. Another
399 option is to disable all protocols on that host in the other config files.
400 .PP
401 If two hosts cannot connect to each other because their \s-1IP\s0 address(es)
402 are not known (such as dial-up hosts), one side will send a \fImediated\fR
403 connection request to a router (routers must be configured to act as
404 routers!), which will send both the originating and the destination host
405 a connection info request with protocol information and \s-1IP\s0 address of the
406 other host (if known). Both hosts will then try to establish a direct
407 connection to the other peer, which is usually possible even when both
408 hosts are behind a \s-1NAT\s0 gateway.
409 .PP
410 Routing via other nodes works because the \s-1SRCDST\s0 field is not encrypted,
411 so the router can just forward the packet to the destination host. Since
412 each host uses its own private key, the router will not be able to
413 decrypt or encrypt packets, it will just act as a simple router and
414 protocol translator.