ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/gvpe/doc/gvpe.protocol.7
Revision: 1.12
Committed: Sat Jul 13 04:10:29 2013 UTC (10 years, 10 months ago) by root
Branch: MAIN
CVS Tags: rel-2_25
Changes since 1.11: +371 -0 lines
Log Message:
*** empty log message ***

File Contents

# User Rev Content
1 root 1.12 .\" Automatically generated by Pod::Man 2.25 (Pod::Simple 3.20)
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     'br\}
42     .\"
43     .\" Escape single quotes in literal strings from groff's Unicode transform.
44     .ie \n(.g .ds Aq \(aq
45     .el .ds Aq '
46     .\"
47     .\" If the F register is turned on, we'll generate index entries on stderr for
48     .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
49     .\" entries marked with X<> in POD. Of course, you'll have to process the
50     .\" output yourself in some meaningful fashion.
51     .ie \nF \{\
52     . de IX
53     . tm Index:\\$1\t\\n%\t"\\$2"
54     ..
55     . nr % 0
56     . rr F
57     .\}
58     .el \{\
59     . de IX
60     ..
61     .\}
62     .\"
63     .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
64     .\" Fear. Run. Save yourself. No user-serviceable parts.
65     . \" fudge factors for nroff and troff
66     .if n \{\
67     . ds #H 0
68     . ds #V .8m
69     . ds #F .3m
70     . ds #[ \f1
71     . ds #] \fP
72     .\}
73     .if t \{\
74     . ds #H ((1u-(\\\\n(.fu%2u))*.13m)
75     . ds #V .6m
76     . ds #F 0
77     . ds #[ \&
78     . ds #] \&
79     .\}
80     . \" simple accents for nroff and troff
81     .if n \{\
82     . ds ' \&
83     . ds ` \&
84     . ds ^ \&
85     . ds , \&
86     . ds ~ ~
87     . ds /
88     .\}
89     .if t \{\
90     . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
91     . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
92     . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
93     . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
94     . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
95     . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
96     .\}
97     . \" troff and (daisy-wheel) nroff accents
98     .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
99     .ds 8 \h'\*(#H'\(*b\h'-\*(#H'
100     .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
101     .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
102     .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
103     .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
104     .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
105     .ds ae a\h'-(\w'a'u*4/10)'e
106     .ds Ae A\h'-(\w'A'u*4/10)'E
107     . \" corrections for vroff
108     .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
109     .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
110     . \" for low resolution devices (crt and lpr)
111     .if \n(.H>23 .if \n(.V>19 \
112     \{\
113     . ds : e
114     . ds 8 ss
115     . ds o a
116     . ds d- d\h'-1'\(ga
117     . ds D- D\h'-1'\(hy
118     . ds th \o'bp'
119     . ds Th \o'LP'
120     . ds ae ae
121     . ds Ae AE
122     .\}
123     .rm #[ #] #H #V #F C
124     .\" ========================================================================
125     .\"
126     .IX Title "GVPE.PROTOCOL 7"
127     .TH GVPE.PROTOCOL 7 "2013-07-12" "2.24" "GNU Virtual Private Ethernet"
128     .\" For nroff, turn off justification. Always turn off hyphenation; it makes
129     .\" way too many mistakes in technical documents.
130     .if n .ad l
131     .nh
132     .SH "The GNU-VPE Protocols"
133     .IX Header "The GNU-VPE Protocols"
134     .SH "Overview"
135     .IX Header "Overview"
136     \&\s-1GVPE\s0 can make use of a number of protocols. One of them is the \s-1GNU\s0 \s-1VPE\s0
137     protocol which is used to authenticate tunnels and send encrypted data
138     packets. This protocol is described in more detail the second part of this
139     document.
140     .PP
141     The first part of this document describes the transport protocols which
142     are used by \s-1GVPE\s0 to send it's data packets over the network.
143     .SH "PART 1: Transport protocols"
144     .IX Header "PART 1: Transport protocols"
145     \&\s-1GVPE\s0 offers a wide range of transport protocols that can be used to
146     interchange data between nodes. Protocols differ in their overhead, speed,
147     reliability, and robustness.
148     .PP
149     The following sections describe each transport protocol in more
150     detail. They are sorted by overhead/efficiency, the most efficient
151     transport is listed first:
152     .SS "\s-1RAW\s0 \s-1IP\s0"
153     .IX Subsection "RAW IP"
154     This protocol is the best choice, performance-wise, as the minimum
155     overhead per packet is only 38 bytes.
156     .PP
157     It works by sending the \s-1VPN\s0 payload using raw \s-1IP\s0 frames (using the
158     protocol set by \f(CW\*(C`ip\-proto\*(C'\fR).
159     .PP
160     Using raw \s-1IP\s0 frames has the drawback that many firewalls block \*(L"unknown\*(R"
161     protocols, so this transport only works if you have full \s-1IP\s0 connectivity
162     between nodes.
163     .SS "\s-1ICMP\s0"
164     .IX Subsection "ICMP"
165     This protocol offers very low overhead (minimum 42 bytes), and can
166     sometimes tunnel through firewalls when other protocols can not.
167     .PP
168     It works by prepending an \s-1ICMP\s0 header with type \f(CW\*(C`icmp\-type\*(C'\fR and a code
169     of \f(CW255\fR. The default \f(CW\*(C`icmp\-type\*(C'\fR is \f(CW\*(C`echo\-reply\*(C'\fR, so the resulting
170     packets look like echo replies, which looks rather strange to network
171     administrators.
172     .PP
173     This transport should only be used if other transports (i.e. raw \s-1IP\s0) are
174     not available or undesirable (due to their overhead).
175     .SS "\s-1UDP\s0"
176     .IX Subsection "UDP"
177     This is a good general choice for the transport protocol as \s-1UDP\s0 packets
178     tunnel well through most firewalls and routers, and the overhead per
179     packet is moderate (minimum 58 bytes).
180     .PP
181     It should be used if \s-1RAW\s0 \s-1IP\s0 is not available.
182     .SS "\s-1TCP\s0"
183     .IX Subsection "TCP"
184     This protocol is a very bad choice, as it not only has high overhead (more
185     than 60 bytes), but the transport also retries on it's own, which leads
186     to congestion when the link has moderate packet loss (as both the \s-1TCP\s0
187     transport and the tunneled traffic will retry, increasing congestion more
188     and more). It also has high latency and is quite inefficient.
189     .PP
190     It's only useful when tunneling through firewalls that block better
191     protocols. If a node doesn't have direct internet access but a \s-1HTTP\s0 proxy
192     that supports the \s-1CONNECT\s0 method it can be used to tunnel through a web
193     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
194     most proxies do not allow connections to other ports.
195     .PP
196     It is an abuse of the usage a proxy was designed for, so make sure you are
197     allowed to use it for \s-1GVPE\s0.
198     .PP
199     This protocol also has server and client sides. If the \f(CW\*(C`tcp\-port\*(C'\fR is
200     set to zero, other nodes cannot connect to this node directly. If the
201     \&\f(CW\*(C`tcp\-port\*(C'\fR is non-zero, the node can act both as a client as well as a
202     server.
203     .SS "\s-1DNS\s0"
204     .IX Subsection "DNS"
205     \&\fB\s-1WARNING:\s0\fR Parsing and generating \s-1DNS\s0 packets is rather tricky. The code
206     almost certainly contains buffer overflows and other, likely exploitable,
207     bugs. You have been warned.
208     .PP
209     This is the worst choice of transport protocol with respect to overhead
210     (overhead can be 2\-3 times higher than the transferred data), and latency
211     (which can be many seconds). Some \s-1DNS\s0 servers might not be prepared to
212     handle the traffic and drop or corrupt packets. The client also has to
213     constantly poll the server for data, so the client will constantly create
214     traffic even if it doesn't need to transport packets.
215     .PP
216     In addition, the same problems as the \s-1TCP\s0 transport also plague this
217     protocol.
218     .PP
219     Its only use is to tunnel through firewalls that do not allow direct
220     internet access. Similar to using a \s-1HTTP\s0 proxy (as the \s-1TCP\s0 transport
221     does), it uses a local \s-1DNS\s0 server/forwarder (given by the \f(CW\*(C`dns\-forw\-host\*(C'\fR
222     configuration value) as a proxy to send and receive data as a client,
223     and an \f(CW\*(C`NS\*(C'\fR record pointing to the \s-1GVPE\s0 server (as given by the
224     \&\f(CW\*(C`dns\-hostname\*(C'\fR directive).
225     .PP
226     The only good side of this protocol is that it can tunnel through most
227     firewalls mostly undetected, iff the local \s-1DNS\s0 server/forwarder is sane
228     (which is true for most routers, wireless \s-1LAN\s0 gateways and nameservers).
229     .PP
230     Fine-tuning needs to be done by editing \f(CW\*(C`src/vpn_dns.C\*(C'\fR directly.
231     .SH "PART 2: The GNU VPE protocol"
232     .IX Header "PART 2: The GNU VPE protocol"
233     This section, unfortunately, is not yet finished, although the protocol
234     is stable (until bugs in the cryptography are found, which will likely
235     completely change the following description). Nevertheless, it should give
236     you some overview over the protocol.
237     .SS "Anatomy of a \s-1VPN\s0 packet"
238     .IX Subsection "Anatomy of a VPN packet"
239     The exact layout and field lengths of a \s-1VPN\s0 packet is determined at
240     compile time and doesn't change. The same structure is used for all
241     transport protocols, be it \s-1RAWIP\s0 or \s-1TCP\s0.
242     .PP
243     .Vb 3
244     \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+
245     \& | HMAC | TYPE | SRCDST | DATA |
246     \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+
247     .Ve
248     .PP
249     The \s-1HMAC\s0 field is present in all packets, even if not used (e.g. in auth
250     request packets), in which case it is set to all zeroes. The checksum
251     itself is calculated over the \s-1TYPE\s0, \s-1SRCDST\s0 and \s-1DATA\s0 fields in all cases.
252     .PP
253     The \s-1TYPE\s0 field is a single byte and determines the purpose of the packet
254     (e.g. \s-1RESET\s0, \s-1COMPRESSED/UNCOMPRESSED\s0 \s-1DATA\s0, \s-1PING\s0, \s-1AUTH\s0 \s-1REQUEST/RESPONSE\s0,
255     \&\s-1CONNECT\s0 \s-1REQUEST/INFO\s0 etc.).
256     .PP
257     \&\s-1SRCDST\s0 is a three byte field which contains the source and destination
258     node IDs (12 bits each).
259     .PP
260     The \s-1DATA\s0 portion differs between each packet type, naturally, and is the
261     only part that can be encrypted. Data packets contain more fields, as
262     shown:
263     .PP
264     .Vb 3
265     \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-+
266     \& | HMAC | TYPE | SRCDST | RAND | SEQNO | DATA |
267     \& +\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-\-+\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-+
268     .Ve
269     .PP
270     \&\s-1RAND\s0 is a sequence of fully random bytes, used to increase the entropy of
271     the data for encryption purposes.
272     .PP
273     \&\s-1SEQNO\s0 is a 32\-bit sequence number. It is negotiated at every connection
274     initialization and starts at some random 31 bit value. \s-1VPE\s0 currently uses
275     a sliding window of 512 packets/sequence numbers to detect reordering,
276     duplication and replay attacks.
277     .PP
278     The encryption is done on \s-1RAND+SEQNO+DATA\s0 in \s-1CBC\s0 mode with zero \s-1IV\s0 (or,
279     equivalently, the \s-1IV\s0 is \s-1RAND+SEQNO\s0, encrypted with the block cipher,
280     unless \s-1RAND\s0 size is decreased or increased over the default value).
281     .SS "The authentication protocol"
282     .IX Subsection "The authentication protocol"
283     Before nodes can exchange packets, they need to establish authenticity of
284     the other side and a key. Every node has a private \s-1RSA\s0 key and the public
285     \&\s-1RSA\s0 keys of all other nodes.
286     .PP
287     A host establishes a simplex connection by sending the other node an \s-1RSA\s0
288     encrypted challenge containing a random challenge (consisting of the
289     encryption and authentication keys to use when sending packets, more
290     random data and \s-1PKCS1_OAEP\s0 padding) and a random 16 byte \*(L"challenge-id\*(R"
291     (used to detect duplicate auth packets). The destination node will respond
292     by replying with an (unencrypted) hash of the decrypted challenge, which
293     will authenticate that node. The destination node will also set the
294     outgoing encryption parameters as given in the packet.
295     .PP
296     When the source node receives a correct auth reply (by verifying the
297     hash and the id, which will expire after 120 seconds), it will start to
298     accept data packets from the destination node.
299     .PP
300     This means that a node can only initiate a simplex connection, telling the
301     other side the key it has to use when it sends packets. The challenge
302     reply is only used to set the current \s-1IP\s0 address of the other side and
303     protocol parameters.
304     .PP
305     This protocol is completely symmetric, so to be able to send packets the
306     destination node must send a challenge in the exact same way as already
307     described (so, in essence, two simplex connections are created per node
308     pair).
309     .SS "Retrying"
310     .IX Subsection "Retrying"
311     When there is no response to an auth request, the node will send auth
312     requests in bursts with an exponential back-off. After some time it will
313     resort to \s-1PING\s0 packets, which are very small (8 bytes + protocol header)
314     and lightweight (no \s-1RSA\s0 operations required). A node that receives ping
315     requests from an unconnected peer will respond by trying to create a
316     connection.
317     .PP
318     In addition to the exponential back-off, there is a global rate-limit on
319     a per-IP base. It allows long bursts but will limit total packet rate to
320     something like one control packet every ten seconds, to avoid accidental
321     floods due to protocol problems (like a \s-1RSA\s0 key file mismatch between two
322     nodes).
323     .PP
324     The intervals between retries are limited by the \f(CW\*(C`max\-retry\*(C'\fR
325     configuration value. A node with \f(CW\*(C`connect\*(C'\fR = \f(CW\*(C`always\*(C'\fR will always retry,
326     a node with \f(CW\*(C`connect\*(C'\fR = \f(CW\*(C`ondemand\*(C'\fR will only try (and re-try) to connect
327     as long as there are packets in the queue, usually this limits the retry
328     period to \f(CW\*(C`max\-ttl\*(C'\fR seconds.
329     .PP
330     Sending packets over the \s-1VPN\s0 will reset the retry intervals as well, which
331     means as long as somebody is trying to send packets to a given node, \s-1GVPE\s0
332     will try to connect every few seconds.
333     .SS "Routing and Protocol translation"
334     .IX Subsection "Routing and Protocol translation"
335     The \s-1GVPE\s0 routing algorithm is easy: there isn't much routing to speak
336     of: When routing packets to another node, \s-1GVPE\s0 tries the following
337     options, in order:
338     .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
339     .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."
340     .PD 0
341     .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
342     .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
343     .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)."
344     .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
345     .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
346     .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."
347     .IP "If no such router exists, then \s-1GVPE\s0 will simply send the packet to the node with the highest priority available." 4
348     .IX Item "If no such router exists, then GVPE will simply send the packet to the node with the highest priority available."
349     .IP "Failing all that, the packet will be dropped." 4
350     .IX Item "Failing all that, the packet will be dropped."
351     .PD
352     .PP
353     A host can usually declare itself unreachable directly by setting it's
354     port number(s) to zero. It can declare other hosts as unreachable by using
355     a config-file that disables all protocols for these other hosts. Another
356     option is to disable all protocols on that host in the other config files.
357     .PP
358     If two hosts cannot connect to each other because their \s-1IP\s0 address(es)
359     are not known (such as dial-up hosts), one side will send a \fImediated\fR
360     connection request to a router (routers must be configured to act as
361     routers!), which will send both the originating and the destination host
362     a connection info request with protocol information and \s-1IP\s0 address of the
363     other host (if known). Both hosts will then try to establish a direct
364     connection to the other peer, which is usually possible even when both
365     hosts are behind a \s-1NAT\s0 gateway.
366     .PP
367     Routing via other nodes works because the \s-1SRCDST\s0 field is not encrypted,
368     so the router can just forward the packet to the destination host. Since
369     each host uses it's own private key, the router will not be able to
370     decrypt or encrypt packets, it will just act as a simple router and
371     protocol translator.