| blob | 987aecb2271350319b3a8006edf34aaafcfd9d1a |
1 <!--
2 - Copyright (C) 2004-2015 Internet Systems Consortium, Inc. ("ISC")
3 - Copyright (C) 2000-2003 Internet Software Consortium.
4 -
5 - Permission to use, copy, modify, and/or distribute this software for any
6 - purpose with or without fee is hereby granted, provided that the above
7 - copyright notice and this permission notice appear in all copies.
8 -
9 - THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH
10 - REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
11 - AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT,
12 - INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
13 - LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
14 - OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
15 - PERFORMANCE OF THIS SOFTWARE.
16 -->
17 <!-- $Id: Bv9ARM.ch04.html,v 1.5 2015/09/03 07:33:34 christos Exp $ -->
18 <html>
19 <head>
27 </head>
32 <tr>
37 </td>
38 </tr>
39 </table>
40 <hr>
41 </div>
47 <dl>
50 <dd><dl><dt><span class="sect2"><a href="Bv9ARM.ch04.html#journal">The journal file</a></span></dt></dl></dd>
51 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#incremental_zone_transfers">Incremental Zone Transfers (IXFR)</a></span></dt>
53 <dd><dl><dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2569938">Example split DNS setup</a></span></dt></dl></dd>
55 <dd><dl>
56 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2570439">Generate Shared Keys for Each Pair of Hosts</a></span></dt>
57 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2570581">Copying the Shared Secret to Both Machines</a></span></dt>
58 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2570592">Informing the Servers of the Key's Existence</a></span></dt>
59 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2570628">Instructing the Server to Use the Key</a></span></dt>
60 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2570685">TSIG Key Based Access Control</a></span></dt>
62 </dl></dd>
66 <dd><dl>
70 </dl></dd>
71 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#dnssec.dynamic.zones">DNSSEC, Dynamic Zones, and Automatic Signing</a></span></dt>
72 <dd><dl>
73 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2611126">Converting from insecure to secure</a></span></dt>
74 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2563650">Dynamic DNS update method</a></span></dt>
75 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2563686">Fully automatic zone signing</a></span></dt>
76 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2563933">Private-type records</a></span></dt>
78 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582689">Dynamic DNS update method</a></span></dt>
79 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582722">Automatic key rollovers</a></span></dt>
80 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582748">NSEC3PARAM rollovers via UPDATE</a></span></dt>
81 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582758">Converting from NSEC to NSEC3</a></span></dt>
82 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582768">Converting from NSEC3 to NSEC</a></span></dt>
83 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2582780">Converting from secure to insecure</a></span></dt>
86 </dl></dd>
87 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#rfc5011.support">Dynamic Trust Anchor Management</a></span></dt>
88 <dd><dl>
90 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2610730">Authoritative Server</a></span></dt>
91 </dl></dd>
92 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#pkcs11">PKCS#11 (Cryptoki) support</a></span></dt>
93 <dd><dl>
96 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2611390">OpenSSL-based PKCS#11</a></span></dt>
99 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2638892">Specifying the engine on the command line</a></span></dt>
100 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2639009">Running named with automatic zone re-signing</a></span></dt>
101 </dl></dd>
102 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#dlz-info">DLZ (Dynamically Loadable Zones)</a></span></dt>
103 <dd><dl>
106 </dl></dd>
107 <dt><span class="sect1"><a href="Bv9ARM.ch04.html#id2571523">IPv6 Support in <acronym class="acronym">BIND</acronym> 9</a></span></dt>
108 <dd><dl>
109 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2571789">Address Lookups Using AAAA Records</a></span></dt>
110 <dt><span class="sect2"><a href="Bv9ARM.ch04.html#id2571811">Address to Name Lookups Using Nibble Format</a></span></dt>
111 </dl></dd>
112 </dl>
113 </div>
117 <p>
119 servers to notify their slave servers of changes to a zone's data. In
121 slave will check to see that its version of the zone is the
122 current version and, if not, initiate a zone transfer.
123 </p>
124 <p>
127 <span><strong class="command">notify</strong></span> option in <a href="Bv9ARM.ch06.html#boolean_options" title="Boolean Options">the section called “Boolean Options”</a> and
129 <a href="Bv9ARM.ch06.html#zone_transfers" title="Zone Transfers">the section called “Zone Transfers”</a>. The <span><strong class="command">NOTIFY</strong></span>
130 protocol is specified in RFC 1996.
131 </p>
134 As a slave zone can also be a master to other slaves, <span><strong class="command">named</strong></span>,
137 cause <span><strong class="command">named</strong></span> to only send <span><strong class="command">NOTIFY</strong></span> for master
138 zones that it loads.
139 </div>
140 </div>
144 <p>
145 Dynamic Update is a method for adding, replacing or deleting
146 records in a master server by sending it a special form of DNS
147 messages. The format and meaning of these messages is specified
148 in RFC 2136.
149 </p>
150 <p>
151 Dynamic update is enabled by including an
152 <span><strong class="command">allow-update</strong></span> or an <span><strong class="command">update-policy</strong></span>
154 </p>
155 <p>
160 See <a href="Bv9ARM.ch06.html#dynamic_update_policies" title="Dynamic Update Policies">the section called “Dynamic Update Policies”</a> for more details.
161 </p>
162 <p>
163 Dynamic updates using Kerberos signed requests can be made
164 using the TKEY/GSS protocol by setting either the
168 Kerberos signed requests will be matched against the update
169 policies for the zone, using the Kerberos principal as the
170 signer for the request.
171 </p>
172 <p>
173 Updating of secure zones (zones using DNSSEC) follows RFC
174 3007: RRSIG, NSEC and NSEC3 records affected by updates are
175 automatically regenerated by the server using an online
176 zone key. Update authorization is based on transaction
177 signatures and an explicit server policy.
178 </p>
182 <p>
183 All changes made to a zone using dynamic update are stored
184 in the zone's journal file. This file is automatically created
185 by the server when the first dynamic update takes place.
186 The name of the journal file is formed by appending the extension
188 corresponding zone
189 file unless specifically overridden. The journal file is in a
190 binary format and should not be edited manually.
191 </p>
192 <p>
193 The server will also occasionally write ("dump")
194 the complete contents of the updated zone to its zone file.
195 This is not done immediately after
196 each dynamic update, because that would be too slow when a large
197 zone is updated frequently. Instead, the dump is delayed by
198 up to 15 minutes, allowing additional updates to take place.
199 During the dump process, transient files will be created
202 will be removed when the dump is complete, and can be safely
203 ignored.
204 </p>
205 <p>
206 When a server is restarted after a shutdown or crash, it will replay
207 the journal file to incorporate into the zone any updates that
208 took
209 place after the last zone dump.
210 </p>
211 <p>
212 Changes that result from incoming incremental zone transfers are
213 also
214 journalled in a similar way.
215 </p>
216 <p>
217 The zone files of dynamic zones cannot normally be edited by
218 hand because they are not guaranteed to contain the most recent
219 dynamic changes — those are only in the journal file.
220 The only way to ensure that the zone file of a dynamic zone
222 </p>
223 <p>
224 If you have to make changes to a dynamic zone
225 manually, the following procedure will work:
226 Disable dynamic updates to the zone using
227 <span><strong class="command">rndc freeze <em class="replaceable"><code>zone</code></em></strong></span>.
228 This will update the zone's master file with the changes
230 Edit the zone file. Run
231 <span><strong class="command">rndc thaw <em class="replaceable"><code>zone</code></em></strong></span>
232 to reload the changed zone and re-enable dynamic updates.
233 </p>
234 <p>
235 <span><strong class="command">rndc sync <em class="replaceable"><code>zone</code></em></strong></span>
236 will update the zone file with changes from the journal file
237 without stopping dynamic updates; this may be useful for viewing
239 file after updating the zone file, use
241 </p>
242 </div>
243 </div>
246 <a name="incremental_zone_transfers"></a>Incremental Zone Transfers (IXFR)</h2></div></div></div>
247 <p>
248 The incremental zone transfer (IXFR) protocol is a way for
249 slave servers to transfer only changed data, instead of having to
250 transfer the entire zone. The IXFR protocol is specified in RFC
252 </p>
253 <p>
255 supports IXFR for those zones
256 where the necessary change history information is available. These
257 include master zones maintained by dynamic update and slave zones
258 whose data was obtained by IXFR. For manually maintained master
259 zones, and for slave zones obtained by performing a full zone
260 transfer (AXFR), IXFR is supported only if the option
263 </p>
264 <p>
266 attempt to use IXFR unless
267 it is explicitly disabled. For more information about disabling
268 IXFR, see the description of the <span><strong class="command">request-ixfr</strong></span> clause
270 </p>
271 </div>
275 <p>
276 Setting up different views, or visibility, of the DNS space to
277 internal and external resolvers is usually referred to as a
279 reasons an organization would want to set up its DNS this way.
280 </p>
281 <p>
282 One common reason for setting up a DNS system this way is
284 Internet. There is some debate as to whether or not this is actually
285 useful.
286 Internal DNS information leaks out in many ways (via email headers,
287 for example) and most savvy "attackers" can find the information
288 they need using other means.
289 However, since listing addresses of internal servers that
290 external clients cannot possibly reach can result in
291 connection delays and other annoyances, an organization may
292 choose to use a Split DNS to present a consistent view of itself
293 to the outside world.
294 </p>
295 <p>
296 Another common reason for setting up a Split DNS system is
297 to allow internal networks that are behind filters or in RFC 1918
298 space (reserved IP space, as documented in RFC 1918) to resolve DNS
299 on the Internet. Split DNS can also be used to allow mail from outside
300 back in to the internal network.
301 </p>
305 <p>
308 has several corporate sites that have an internal network with
309 reserved
310 Internet Protocol (IP) space and an external demilitarized zone (DMZ),
311 or "outside" section of a network, that is available to the public.
312 </p>
313 <p>
315 to be able to resolve external hostnames and to exchange mail with
316 people on the outside. The company also wants its internal resolvers
317 to have access to certain internal-only zones that are not available
318 at all outside of the internal network.
319 </p>
320 <p>
321 In order to accomplish this, the company will set up two sets
322 of name servers. One set will be on the inside network (in the
323 reserved
324 IP space) and the other set will be on bastion hosts, which are
325 "proxy"
326 hosts that can talk to both sides of its network, in the DMZ.
327 </p>
328 <p>
329 The internal servers will be configured to forward all queries,
330 except queries for <code class="filename">site1.internal</code>, <code class="filename">site2.internal</code>, <code class="filename">site1.example.com</code>,
332 in the
333 DMZ. These internal servers will have complete sets of information
334 for <code class="filename">site1.example.com</code>, <code class="filename">site2.example.com</code>, <code class="filename">site1.internal</code>,
336 </p>
337 <p>
338 To protect the <code class="filename">site1.internal</code> and <code class="filename">site2.internal</code> domains,
339 the internal name servers must be configured to disallow all queries
340 to these domains from any external hosts, including the bastion
341 hosts.
342 </p>
343 <p>
344 The external servers, which are on the bastion hosts, will
345 be configured to serve the "public" version of the <code class="filename">site1</code> and <code class="filename">site2.example.com</code> zones.
346 This could include things such as the host records for public servers
347 (<code class="filename">www.example.com</code> and <code class="filename">ftp.example.com</code>),
348 and mail exchange (MX) records (<code class="filename">a.mx.example.com</code> and <code class="filename">b.mx.example.com</code>).
349 </p>
350 <p>
351 In addition, the public <code class="filename">site1</code> and <code class="filename">site2.example.com</code> zones
352 should have special MX records that contain wildcard (`*') records
353 pointing to the bastion hosts. This is needed because external mail
354 servers do not have any other way of looking up how to deliver mail
355 to those internal hosts. With the wildcard records, the mail will
356 be delivered to the bastion host, which can then forward it on to
357 internal hosts.
358 </p>
359 <p>
360 Here's an example of a wildcard MX record:
361 </p>
363 <p>
364 Now that they accept mail on behalf of anything in the internal
365 network, the bastion hosts will need to know how to deliver mail
366 to internal hosts. In order for this to work properly, the resolvers
367 on
368 the bastion hosts will need to be configured to point to the internal
369 name servers for DNS resolution.
370 </p>
371 <p>
372 Queries for internal hostnames will be answered by the internal
373 servers, and queries for external hostnames will be forwarded back
374 out to the DNS servers on the bastion hosts.
375 </p>
376 <p>
377 In order for all this to work properly, internal clients will
379 name servers for DNS queries. This could also be enforced via
380 selective
381 filtering on the network.
382 </p>
383 <p>
385 internal clients will now be able to:
386 </p>
388 <li>
390 and
392 </li>
393 <li>
396 </li>
399 </ul></div>
400 <p>
401 Hosts on the Internet will be able to:
402 </p>
404 <li>
406 and
408 </li>
409 <li>
412 </li>
413 </ul></div>
414 <p>
415 Here is an example configuration for the setup we just
416 described above. Note that this is only configuration information;
417 for information on how to configure your zone files, see <a href="Bv9ARM.ch03.html#sample_configuration" title="Sample Configurations">the section called “Sample Configurations”</a>.
418 </p>
419 <p>
420 Internal DNS server config:
421 </p>
428 options {
429 ...
430 ...
431 forward only;
432 // forward to external servers
433 forwarders {
435 };
436 // sample allow-transfer (no one)
437 allow-transfer { none; };
438 // restrict query access
439 allow-query { internals; externals; };
440 // restrict recursion
441 allow-recursion { internals; };
442 ...
443 ...
444 };
446 // sample master zone
447 zone "site1.example.com" {
448 type master;
449 file "m/site1.example.com";
450 // do normal iterative resolution (do not forward)
451 forwarders { };
452 allow-query { internals; externals; };
453 allow-transfer { internals; };
454 };
456 // sample slave zone
457 zone "site2.example.com" {
458 type slave;
459 file "s/site2.example.com";
460 masters { 172.16.72.3; };
461 forwarders { };
462 allow-query { internals; externals; };
463 allow-transfer { internals; };
464 };
466 zone "site1.internal" {
467 type master;
468 file "m/site1.internal";
469 forwarders { };
470 allow-query { internals; };
471 allow-transfer { internals; }
472 };
474 zone "site2.internal" {
475 type slave;
476 file "s/site2.internal";
477 masters { 172.16.72.3; };
478 forwarders { };
479 allow-query { internals };
480 allow-transfer { internals; }
481 };
482 </pre>
483 <p>
484 External (bastion host) DNS server config:
485 </p>
489 acl externals { bastion-ips-go-here; };
491 options {
492 ...
493 ...
494 // sample allow-transfer (no one)
495 allow-transfer { none; };
496 // default query access
497 allow-query { any; };
498 // restrict cache access
499 allow-query-cache { internals; externals; };
500 // restrict recursion
501 allow-recursion { internals; externals; };
502 ...
503 ...
504 };
506 // sample slave zone
507 zone "site1.example.com" {
508 type master;
509 file "m/site1.foo.com";
510 allow-transfer { internals; externals; };
511 };
513 zone "site2.example.com" {
514 type slave;
515 file "s/site2.foo.com";
516 masters { another_bastion_host_maybe; };
517 allow-transfer { internals; externals; }
518 };
519 </pre>
520 <p>
522 the bastion host(s):
523 </p>
525 search ...
526 nameserver 172.16.72.2
527 nameserver 172.16.72.3
528 nameserver 172.16.72.4
529 </pre>
530 </div>
531 </div>
535 <p>
536 This is a short guide to setting up Transaction SIGnatures
537 (TSIG) based transaction security in <acronym class="acronym">BIND</acronym>. It describes changes
538 to the configuration file as well as what changes are required for
539 different features, including the process of creating transaction
541 </p>
542 <p>
544 to server communication.
545 This includes zone transfer, notify, and recursive query messages.
546 Resolvers based on newer versions of <acronym class="acronym">BIND</acronym> 8 have limited support
547 for TSIG.
548 </p>
549 <p>
550 TSIG can also be useful for dynamic update. A primary
551 server for a dynamic zone should control access to the dynamic
552 update service, but IP-based access control is insufficient.
553 The cryptographic access control provided by TSIG
558 </p>
562 <p>
563 A shared secret is generated to be shared between <span class="emphasis"><em>host1</em></span> and <span class="emphasis"><em>host2</em></span>.
564 An arbitrary key name is chosen: "host1-host2.". The key name must
565 be the same on both hosts.
566 </p>
570 <p>
572 key as described above. Longer keys are better, but shorter keys
573 are easier to read. Note that the maximum key length is the digest
574 length, here 256 bits.
575 </p>
576 <p>
577 <strong class="userinput"><code>dnssec-keygen -a hmac-sha256 -b 128 -n HOST host1-host2.</code></strong>
578 </p>
579 <p>
581 Nothing directly uses this file, but the base-64 encoded string
583 can be extracted from the file and used as a shared secret:
584 </p>
586 <p>
588 be used as the shared secret.
589 </p>
590 </div>
594 <p>
595 The shared secret is simply a random sequence of bits, encoded
597 the length is a multiple of 4 and only valid characters are used),
598 so the shared secret can be manually generated.
599 </p>
600 <p>
601 Also, a known string can be run through <span><strong class="command">mmencode</strong></span> or
602 a similar program to generate base-64 encoded data.
603 </p>
604 </div>
605 </div>
609 <p>
610 This is beyond the scope of DNS. A secure transport mechanism
611 should be used. This could be secure FTP, ssh, telephone, etc.
612 </p>
613 </div>
617 <p>
618 Imagine <span class="emphasis"><em>host1</em></span> and <span class="emphasis"><em>host 2</em></span>
619 are
620 both servers. The following is added to each server's <code class="filename">named.conf</code> file:
621 </p>
623 key host1-host2. {
624 algorithm hmac-sha256;
625 secret "La/E5CjG9O+os1jq0a2jdA==";
626 };
627 </pre>
628 <p>
629 The secret is the one generated above. Since this is a secret, it
631 non-world readable, or the key directive be added to a non-world
633 </p>
634 <p>
635 At this point, the key is recognized. This means that if the
636 server receives a message signed by this key, it can verify the
637 signature. If the signature is successfully verified, the
638 response is signed by the same key.
639 </p>
640 </div>
644 <p>
645 Since keys are shared between two hosts only, the server must
646 be told when keys are to be used. The following is added to the <code class="filename">named.conf</code> file
647 for <span class="emphasis"><em>host1</em></span>, if the IP address of <span class="emphasis"><em>host2</em></span> is
648 10.1.2.3:
649 </p>
651 server 10.1.2.3 {
652 keys { host1-host2. ;};
653 };
654 </pre>
655 <p>
656 Multiple keys may be present, but only the first is used.
657 This directive does not contain any secrets, so it may be in a
658 world-readable
659 file.
660 </p>
661 <p>
663 to that address, the message will be signed with the specified key. <span class="emphasis"><em>host1</em></span> will
664 expect any responses to signed messages to be signed with the same
665 key.
666 </p>
667 <p>
669 configuration file (with <span class="emphasis"><em>host1</em></span>'s address) for <span class="emphasis"><em>host2</em></span> to
671 </p>
672 </div>
676 <p>
678 to be specified in ACL
679 definitions and
681 directives.
682 This has been extended to allow TSIG keys also. The above key would
684 </p>
685 <p>
687 </p>
689 allow-update { key host1-host2. ;};
690 </pre>
691 <p>
692 This allows dynamic updates to succeed only if the request
694 </p>
695 <p>
696 See <a href="Bv9ARM.ch06.html#dynamic_update_policies" title="Dynamic Update Policies">the section called “Dynamic Update Policies”</a> for a discussion of
698 </p>
699 </div>
703 <p>
704 The processing of TSIG signed messages can result in
705 several errors. If a signed message is sent to a non-TSIG aware
706 server, a FORMERR (format error) will be returned, since the server will not
707 understand the record. This is a result of misconfiguration,
708 since the server must be explicitly configured to send a TSIG
709 signed message to a specific server.
710 </p>
711 <p>
712 If a TSIG aware server receives a message signed by an
713 unknown key, the response will be unsigned with the TSIG
714 extended error code set to BADKEY. If a TSIG aware server
715 receives a message with a signature that does not validate, the
716 response will be unsigned with the TSIG extended error code set
717 to BADSIG. If a TSIG aware server receives a message with a time
718 outside of the allowed range, the response will be signed with
719 the TSIG extended error code set to BADTIME, and the time values
720 will be adjusted so that the response can be successfully
721 verified. In any of these cases, the message's rcode (response code) is set to
722 NOTAUTH (not authenticated).
723 </p>
724 </div>
725 </div>
730 is a mechanism for automatically generating a shared secret
731 between two hosts. There are several "modes" of
734 these modes, the Diffie-Hellman key exchange. Both hosts are
735 required to have a Diffie-Hellman KEY record (although this
736 record is not required to be present in a zone). The
738 signed either by TSIG or SIG(0). The result of
741 used to delete shared secrets that it had previously
742 generated.
743 </p>
744 <p>
746 client
748 query
749 (including any appropriate KEYs) to a TKEY-aware server. The
750 server response, if it indicates success, will contain a
752 After
753 this exchange, both participants have enough information to
754 determine the shared secret; the exact process depends on the
756 Diffie-Hellman
758 exchanged,
759 and the shared secret is derived by both participants.
760 </p>
761 </div>
765 <p>
768 SIG(0)
769 uses public/private keys to authenticate messages. Access control
770 is performed in the same manner as TSIG keys; privileges can be
771 granted or denied based on the key name.
772 </p>
773 <p>
774 When a SIG(0) signed message is received, it will only be
775 verified if the key is known and trusted by the server; the server
776 will not attempt to locate and/or validate the key.
777 </p>
778 <p>
779 SIG(0) signing of multiple-message TCP streams is not
780 supported.
781 </p>
782 <p>
785 </p>
786 </div>
790 <p>
791 Cryptographic authentication of DNS information is possible
794 This section describes the creation and use of DNSSEC signed zones.
795 </p>
796 <p>
797 In order to set up a DNSSEC secure zone, there are a series
799 9 ships
800 with several tools
801 that are used in this process, which are explained in more detail
803 full list of parameters. Note that the DNSSEC tools require the
804 keyset files to be in the working directory or the
806 that the tools shipped with BIND 9.2.x and earlier are not compatible
807 with the current ones.
808 </p>
809 <p>
810 There must also be communication with the administrators of
811 the parent and/or child zone to transmit keys. A zone's security
812 status must be indicated by the parent zone for a DNSSEC capable
813 resolver to trust its data. This is done through the presence
815 delegation
816 point.
817 </p>
818 <p>
819 For other servers to trust data in this zone, they must
820 either be statically configured with this zone's zone key or the
821 zone key of another zone above this one in the DNS tree.
822 </p>
826 <p>
828 generate keys.
829 </p>
830 <p>
831 A secure zone must contain one or more zone keys. The
832 zone keys will sign all other records in the zone, as well as
833 the zone keys of any secure delegated zones. Zone keys must
834 have the same name as the zone, a name type of
836 authentication.
837 It is recommended that zone keys use a cryptographic algorithm
838 designated as "mandatory to implement" by the IETF; currently
839 the only one is RSASHA1.
840 </p>
841 <p>
842 The following command will generate a 768-bit RSASHA1 key for
844 </p>
845 <p>
846 <strong class="userinput"><code>dnssec-keygen -a RSASHA1 -b 768 -n ZONE child.example.</code></strong>
847 </p>
848 <p>
849 Two output files will be produced:
852 (where
853 12345 is an example of a key tag). The key filenames contain
855 algorithm (3
857 this case).
859 file) is
860 used to generate signatures, and the public key (in the
862 verification.
863 </p>
864 <p>
865 To generate another key with the same properties (but with
866 a different key tag), repeat the above command.
867 </p>
868 <p>
870 to get a key pair from a crypto hardware and build the key
872 </p>
873 <p>
874 The public keys should be inserted into the zone file by
877 </p>
878 </div>
882 <p>
884 to sign a zone.
885 </p>
886 <p>
888 secure subzones should be present. The zone signer will
893 is not specified, then DS RRsets for the secure child
894 zones need to be added manually.
895 </p>
896 <p>
897 The following command signs the zone, assuming it is in a
899 default, all zone keys which have an available private key are
900 used to generate signatures.
901 </p>
902 <p>
903 <strong class="userinput"><code>dnssec-signzone -o child.example zone.child.example</code></strong>
904 </p>
905 <p>
906 One output file is produced:
908 file
910 as the
911 input file for the zone.
912 </p>
914 will also produce a keyset and dsset files and optionally a
915 dlvset file. These are used to provide the parent zone
918 secure entry point to the zone.
919 </p>
920 </div>
924 <p>
926 to DNS requests from DNSSEC aware clients,
928 (This is the default setting.)
929 </p>
930 <p>
935 <strong class="userinput"><code>yes</code></strong> or <strong class="userinput"><code>auto</code></strong>.
936 </p>
937 <p>
940 trust anchor for the DNS root zone will be used.
942 then at least one trust anchor must be configured
946 will not occur. The default setting is
948 </p>
949 <p>
951 for zones that are used to form the first link in the
952 cryptographic chain of trust. All keys listed in
954 are deemed to exist and only the listed keys will be used
955 to validated the DNSKEY RRset that they are from.
956 </p>
957 <p>
959 automatically kept up to date via RFC 5011 trust anchor
960 maintenance.
961 </p>
962 <p>
965 later in this document.
966 </p>
967 <p>
969 9 does not verify signatures on load, so zone keys for
970 authoritative zones do not need to be specified in the
971 configuration file.
972 </p>
973 <p>
974 After DNSSEC gets established, a typical DNSSEC configuration
975 will look something like the following. It has one or
976 more public keys for the root. This allows answers from
977 outside the organization to be validated. It will also
978 have several keys for parts of the namespace the organization
980 is immune to compromises in the DNSSEC components of the security
981 of parent zones.
982 </p>
984 managed-keys {
985 /* Root Key */
987 JxrGkxJWoZu6I7PzJu/E9gx4UC1zGAHlXKdE4zYIpRh
988 aBKnvcC2U9mZhkdUpd1Vso/HAdjNe8LmMlnzY3zy2Xy
989 4klWOADTPzSv9eamj8V18PHGjBLaVtYvk/ln5ZApjYg
990 hf+6fElrmLkdaz MQ2OCnACR817DF4BBa7UR/beDHyp
991 5iWTXWSi6XmoJLbG9Scqc7l70KDqlvXR3M/lUUVRbke
992 g1IPJSidmK3ZyCllh4XSKbje/45SKucHgnwU5jefMtq
993 66gKodQj+MiA21AfUVe7u99WzTLzY3qlxDhxYQQ20FQ
994 97S+LKUTpQcq27R7AT3/V5hRQxScINqwcz4jYqZD2fQ
995 dgxbcDTClU0CRBdiieyLMNzXG3";
996 };
998 trusted-keys {
999 /* Key for our organization's forward zone */
1001 5KbhTjrW1ZaARmPhEZZe3Y9ifgEuq7vZ/z
1002 GZUdEGNWy+JZzus0lUptwgjGwhUS1558Hb
1003 4JKUbbOTcM8pwXlj0EiX3oDFVmjHO444gL
1004 kBOUKUf/mC7HvfwYH/Be22GnClrinKJp1O
1005 g4ywzO9WglMk7jbfW33gUKvirTHr25GL7S
1006 TQUzBb5Usxt8lgnyTUHs1t3JwCY5hKZ6Cq
1007 FxmAVZP20igTixin/1LcrgX/KMEGd/biuv
1008 F4qJCyduieHukuY3H4XMAcR+xia2nIUPvm
1009 /oyWR8BW/hWdzOvnSCThlHf3xiYleDbt/o
1010 1OTQ09A0=";
1012 /* Key for our reverse zone. */
1014 xOdNax071L18QqZnQQQAVVr+i
1015 LhGTnNGp3HoWQLUIzKrJVZ3zg
1016 gy3WwNT6kZo6c0tszYqbtvchm
1017 gQC8CzKojM/W16i6MG/eafGU3
1018 siaOdS0yOI6BgPsw+YZdzlYMa
1019 IJGf4M4dyoKIhzdZyQ2bYQrjy
1020 Q4LB0lC7aOnsMyYKHHYeRvPxj
1021 IQXmdqgOJGq+vsevG06zW+1xg
1022 YJh9rCIfnm1GX/KMgxLPG2vXT
1023 D/RnLX+D3T3UL7HJYHJhAZD5L
1024 59VvjSPsZJHeDCUyWYrvPZesZ
1025 DIRvhDD52SKvbheeTJUm6Ehkz
1026 ytNN2SN96QRk8j/iI8ib";
1027 };
1029 options {
1030 ...
1031 dnssec-enable yes;
1032 dnssec-validation yes;
1033 };
1034 </pre>
1037 None of the keys listed in this example are valid. In particular,
1038 the root key is not valid.
1039 </div>
1040 <p>
1041 When DNSSEC validation is enabled and properly configured,
1042 the resolver will reject any answers from signed, secure zones
1043 which fail to validate, and will return SERVFAIL to the client.
1044 </p>
1045 <p>
1046 Responses may fail to validate for any of several reasons,
1047 including missing, expired, or invalid signatures, a key which
1048 does not match the DS RRset in the parent zone, or an insecure
1049 response from a zone which, according to its parent, should have
1050 been secure.
1051 </p>
1054 <p>
1055 When the validator receives a response from an unsigned zone
1056 that has a signed parent, it must confirm with the parent
1057 that the zone was intentionally left unsigned. It does
1058 this by verifying, via signed and validated NSEC/NSEC3 records,
1059 that the parent zone contains no DS records for the child.
1060 </p>
1061 <p>
1063 is insecure, then the response is accepted. However, if it
1064 cannot, then it must assume an insecure response to be a
1065 forgery; it rejects the response and logs an error.
1066 </p>
1067 <p>
1068 The logged error reads "insecurity proof failed" and
1069 "got insecure response; parent indicates it should be secure".
1071 This referred to the zone, not the response.)
1072 </p>
1073 </div>
1074 </div>
1075 </div>
1078 <a name="dnssec.dynamic.zones"></a>DNSSEC, Dynamic Zones, and Automatic Signing</h2></div></div></div>
1080 from insecure to signed and back again. A secure zone can use
1081 either NSEC or NSEC3 chains.</p>
1084 <p>Changing a zone from insecure to secure can be done in two
1085 ways: using a dynamic DNS update, or the
1087 <p>For either method, you need to configure
1090 parts of the keys that will be used to sign the zone. These files
1091 will have been generated by
1093 in the key-directory, as specified in
1096 zone example.net {
1097 type master;
1098 update-policy local;
1099 file "dynamic/example.net/example.net";
1100 key-directory "dynamic/example.net";
1101 };
1102 </pre>
1103 <p>If one KSK and one ZSK DNSKEY key have been generated, this
1104 configuration will cause all records in the zone to be signed
1105 with the ZSK, and the DNSKEY RRset to be signed with the KSK as
1106 well. An NSEC chain will be generated as part of the initial
1107 signing process.</p>
1112 % nsupdate
1114 > update add example.net DNSKEY 256 3 7 AwEAAZn17pUF0KpbPA2c7Gz76Vb18v0teKT3EyAGfBfL8eQ8al35zz3Y I1m/SAQBxIqMfLtIwqWPdgthsu36azGQAX8=
1115 > update add example.net DNSKEY 257 3 7 AwEAAd/7odU/64o2LGsifbLtQmtO8dFDtTAZXSX2+X3e/UNlq9IHq3Y0 XtC0Iuawl/qkaKVxXe2lo8Ct+dM6UehyCqk=
1116 > send
1117 </pre>
1118 <p>While the update request will complete almost immediately,
1119 the zone will not be completely signed until
1121 generate the NSEC and RRSIG records. The NSEC record at the apex
1122 will be added last, to signal that there is a complete NSEC
1123 chain.</p>
1124 <p>If you wish to sign using NSEC3 instead of NSEC, you should
1125 add an NSEC3PARAM record to the initial update request. If you
1126 wish the NSEC3 chain to have the OPTOUT bit set, set it in the
1127 flags field of the NSEC3PARAM record.</p>
1129 % nsupdate
1131 > update add example.net DNSKEY 256 3 7 AwEAAZn17pUF0KpbPA2c7Gz76Vb18v0teKT3EyAGfBfL8eQ8al35zz3Y I1m/SAQBxIqMfLtIwqWPdgthsu36azGQAX8=
1132 > update add example.net DNSKEY 257 3 7 AwEAAd/7odU/64o2LGsifbLtQmtO8dFDtTAZXSX2+X3e/UNlq9IHq3Y0 XtC0Iuawl/qkaKVxXe2lo8Ct+dM6UehyCqk=
1134 > send
1135 </pre>
1136 <p>Again, this update request will complete almost
1137 immediately; however, the record won't show up until
1139 relevant chain. A private type record will be created to record
1140 the state of the operation (see below for more details), and will
1141 be removed once the operation completes.</p>
1142 <p>While the initial signing and NSEC/NSEC3 chain generation
1143 is happening, other updates are possible as well.</p>
1146 <p>To enable automatic signing, add the
1152 <p>With
1155 matching the zone, insert them into the zone, and use them to
1156 sign the zone. It will do so only when it receives an
1158 <p>
1161 functionality, but will also automatically adjust the zone's
1162 DNSKEY records on schedule according to the keys' timing metadata.
1163 (See <a href="man.dnssec-keygen.html" title="dnssec-keygen"><span class="refentrytitle"><span class="application">dnssec-keygen</span></span>(8)</a> and
1164 <a href="man.dnssec-settime.html" title="dnssec-settime"><span class="refentrytitle"><span class="application">dnssec-settime</span></span>(8)</a> for more information.)
1165 </p>
1166 <p>
1168 for keys matching the zone, and if the keys' metadata indicates
1169 that any change should be made the zone, such as adding, removing,
1170 or revoking a key, then that action will be carried out. By default,
1171 the key directory is checked for changes every 60 minutes; this period
1173 to a maximum of 24 hours. The <span><strong class="command">rndc loadkeys</strong></span> forces
1175 </p>
1176 <p>
1177 If keys are present in the key directory the first time the zone
1178 is loaded, the zone will be signed immediately, without waiting for an
1179 <span><strong class="command">rndc sign</strong></span> or <span><strong class="command">rndc loadkeys</strong></span>
1180 command. (Those commands can still be used when there are unscheduled
1181 key changes, however.)
1182 </p>
1183 <p>
1184 When new keys are added to a zone, the TTL is set to match that
1185 of any existing DNSKEY RRset. If there is no existing DNSKEY RRset,
1186 then the TTL will be set to the TTL specified when the key was
1188 any, or to the SOA TTL.
1189 </p>
1190 <p>
1191 If you wish the zone to be signed using NSEC3 instead of NSEC,
1192 submit an NSEC3PARAM record via dynamic update prior to the
1193 scheduled publication and activation of the keys. If you wish the
1194 NSEC3 chain to have the OPTOUT bit set, set it in the flags field
1195 of the NSEC3PARAM record. The NSEC3PARAM record will not appear in
1196 the zone immediately, but it will be stored for later reference. When
1197 the zone is signed and the NSEC3 chain is completed, the NSEC3PARAM
1198 record will appear in the zone.
1199 </p>
1200 <p>Using the
1202 configured to allow dynamic updates, by adding an
1205 configuration. If this has not been done, the configuration will
1206 fail.</p>
1209 <p>The state of the signing process is signaled by
1210 private-type records (with a default type value of 65534). When
1211 signing is complete, these records will have a nonzero value for
1212 the final octet (for those records which have a nonzero initial
1213 octet).</p>
1214 <p>The private type record format: If the first octet is
1215 non-zero then the record indicates that the zone needs to be
1216 signed with the key matching the record, or that all signatures
1217 that match the record should be removed.</p>
1218 <p>
1219 </p>
1221 <br>
1226 </p></div>
1227 <p>
1228 </p>
1230 dynamic update. Attempts to remove other private type records
1231 will be silently ignored.</p>
1232 <p>If the first octet is zero (this is a reserved algorithm
1233 number that should never appear in a DNSKEY record) then the
1234 record indicates changes to the NSEC3 chains are in progress. The
1235 rest of the record contains an NSEC3PARAM record. The flag field
1236 tells what operation to perform based on the flag bits.</p>
1237 <p>
1238 </p>
1240 <br>
1245 </p></div>
1246 <p>
1247 </p>
1250 <p>As with insecure-to-secure conversions, rolling DNSSEC
1251 keys can be done in two ways: using a dynamic DNS update, or the
1255 <p> To perform key rollovers via dynamic update, you need to add
1258 DNSKEY RRs via dynamic update.
1260 with the new keys. When the signing is complete the private type
1261 records will be updated so that the last octet is non
1262 zero.</p>
1263 <p>If this is for a KSK you need to inform the parent and any
1264 trust anchor repositories of the new KSK.</p>
1265 <p>You should then wait for the maximum TTL in the zone before
1266 removing the old DNSKEY. If it is a KSK that is being updated,
1267 you also need to wait for the DS RRset in the parent to be
1268 updated and its TTL to expire. This ensures that all clients will
1269 be able to verify at least one signature when you remove the old
1270 DNSKEY.</p>
1271 <p>The old DNSKEY can be removed via UPDATE. Take care to
1272 specify the correct key.
1274 by the old key after the update completes.</p>
1277 <p>When a new key reaches its activation date (as set by
1278 <span><strong class="command">dnssec-keygen</strong></span> or <span><strong class="command">dnssec-settime</strong></span>),
1280 <code class="constant">maintain</code>, <span><strong class="command">named</strong></span> will
1281 automatically carry out the key rollover. If the key's algorithm
1282 has not previously been used to sign the zone, then the zone will
1283 be fully signed as quickly as possible. However, if the new key
1284 is replacing an existing key of the same algorithm, then the
1285 zone will be re-signed incrementally, with signatures from the
1286 old key being replaced with signatures from the new key as their
1287 signature validity periods expire. By default, this rollover
1288 completes in 30 days, after which it will be safe to remove the
1289 old key from the DNSKEY RRset.</p>
1292 <p>Add the new NSEC3PARAM record via dynamic update. When the
1293 new NSEC3 chain has been generated, the NSEC3PARAM flag field
1294 will be zero. At this point you can remove the old NSEC3PARAM
1295 record. The old chain will be removed after the update request
1296 completes.</p>
1299 <p>To do this, you just need to add an NSEC3PARAM record. When
1300 the conversion is complete, the NSEC chain will have been removed
1301 and the NSEC3PARAM record will have a zero flag field. The NSEC3
1302 chain will be generated before the NSEC chain is
1303 destroyed.</p>
1307 remove all NSEC3PARAM records with a zero flag
1308 field. The NSEC chain will be generated before the NSEC3 chain is
1309 removed.</p>
1312 <p>To convert a signed zone to unsigned using dynamic DNS,
1313 delete all the DNSKEY records from the zone apex using
1315 and associated NSEC3PARAM records will be removed automatically.
1316 This will take place after the update request completes.</p>
1317 <p> This requires the
1322 zone statement is used, it should be removed or changed to
1324 </p>
1327 <p>In any secure zone which supports dynamic updates, named
1328 will periodically re-sign RRsets which have not been re-signed as
1329 a result of some update action. The signature lifetimes will be
1330 adjusted so as to spread the re-sign load over time rather than
1331 all at once.</p>
1334 <p>
1336 where all the NSEC3 records in the zone have the same OPTOUT
1337 state.
1339 records in the chain have mixed OPTOUT state.
1341 state of an individual NSEC3 record, the entire chain needs to be
1342 changed if the OPTOUT state of an individual NSEC3 needs to be
1343 changed.</p>
1344 </div>
1349 anchor management. Using this feature allows
1351 DNSSEC keys without any need for the operator to make changes to
1352 configuration files.</p>
1357 maintain a trust anchor, configure the trust anchor using a
1359 this can be found in
1361 and Usage">the section called “<span><strong class="command">managed-keys</strong></span> Statement Definition
1363 </div>
1368 maintenance, generate two (or more) key signing keys (KSKs) for
1369 the zone. Sign the zone with one of them; this is the "active"
1370 KSK. All KSK's which do not sign the zone are "stand-by"
1371 keys.</p>
1372 <p>Any validating resolver which is configured to use the
1373 active KSK as an RFC 5011-managed trust anchor will take note
1374 of the stand-by KSKs in the zone's DNSKEY RRset, and store them
1375 for future reference. The resolver will recheck the zone
1376 periodically, and after 30 days, if the new key is still there,
1377 then the key will be accepted by the resolver as a valid trust
1378 anchor for the zone. Any time after this 30-day acceptance
1379 timer has completed, the active KSK can be revoked, and the
1381 <p>The easiest way to place a stand-by key in a zone is to
1382 use the "smart signing" features of
1385 date in the past, but an activation date which is unset or in
1386 the future, "
1388 record in the zone, but will not sign with it:</p>
1390 $ <strong class="userinput"><code>dnssec-keygen -K keys -f KSK -P now -A now+2y example.net</code></strong>
1392 </pre>
1393 <p>To revoke a key, the new command
1395 REVOKED bit to the key flags and re-generates the
1398 <p>After revoking the active key, the zone must be signed
1399 with both the revoked KSK and the new active KSK. (Smart
1400 signing takes care of this automatically.)</p>
1401 <p>Once a key has been revoked and used to sign the DNSKEY
1402 RRset in which it appears, that key will never again be
1403 accepted as a valid trust anchor by the resolver. However,
1404 validation can proceed using the new active key (which had been
1405 accepted by the resolver when it was a stand-by key).</p>
1407 scenarios.</p>
1408 <p>When a key has been revoked, its key ID changes,
1413 revoked, then the two key ID's will collide, causing several
1414 problems. To prevent this,
1416 another key is present which may collide. This checking will
1417 only occur if the new keys are written to the same directory
1418 which holds all other keys in use for that zone.</p>
1420 Exercise caution if using key revocation on keys that were
1421 generated by previous releases, or if using keys stored in
1422 multiple directories or on multiple machines.</p>
1424 address this problem in a different way, by storing revoked
1425 keys with their original unrevoked key ID's.</p>
1426 </div>
1427 </div>
1431 <p>
1433 platform-independent API for the control of hardware security
1434 modules (HSMs) and other cryptographic support devices.
1435 </p>
1436 <p>
1437 BIND 9 is known to work with three HSMs: The AEP Keyper, which has
1438 been tested with Debian Linux, Solaris x86 and Windows Server 2003;
1439 the Thales nShield, tested with Debian Linux; and the Sun SCA 6000
1440 cryptographic acceleration board, tested with Solaris x86. In
1441 addition, BIND can be used with all current versions of SoftHSM,
1442 a software-based HSM simulator library produced by the OpenDNSSEC
1443 project.
1444 </p>
1445 <p>
1447 library which provides a low-level PKCS#11 interface to drive the HSM
1448 hardware. The PKCS#11 provider library comes from the HSM vendor, and
1449 it is specific to the HSM to be controlled.
1450 </p>
1451 <p>
1454 mechanism, BIND uses a modified version of OpenSSL, which loads
1455 the provider library and operates the HSM indirectly; any
1456 cryptographic operations not supported by the HSM can be carried
1457 out by OpenSSL instead. The second mechanism enables BIND to bypass
1458 OpenSSL completely; BIND loads the provider library itself, and uses
1459 the PKCS#11 API to drive the HSM directly.
1460 </p>
1464 <p>
1465 See the documentation provided by your HSM vendor for
1466 information about installing, initializing, testing and
1467 troubleshooting the HSM.
1468 </p>
1469 </div>
1473 <p>
1474 Native PKCS#11 mode will only work with an HSM capable of carrying
1476 need. The HSM's provider library must have a complete implementation
1477 of the PKCS#11 API, so that all these functions are accessible. As of
1478 this writing, only the Thales nShield HSM and SoftHSMv2 can be used
1479 in this fashion. For other HSMs, including the AEP Keyper, Sun SCA
1481 (Note: Eventually, when more HSMs become capable of supporting
1483 be deprecated.)
1484 </p>
1485 <p>
1486 To build BIND with native PKCS#11, configure as follows:
1487 </p>
1492 </pre>
1493 <p>
1495 and the <span><strong class="command">dnssec-*</strong></span> and <span><strong class="command">pkcs11-*</strong></span>
1496 tools, to use the PKCS#11 provider library specified in
1498 (The provider library path can be overridden using the
1500 <span><strong class="command">dnssec-*</strong></span> tools, or the <code class="option">-m</code> in
1502 </p>
1506 <p>
1507 SoftHSMv2, the latest development version of SoftHSM, is available
1508 from
1510 https://github.com/opendnssec/SoftHSMv2
1511 </a>.
1512 It is a software library developed by the OpenDNSSEC project
1514 http://www.opendnssec.org
1515 </a>)
1516 which provides a PKCS#11 interface to a virtual HSM, implemented in
1517 the form of a SQLite3 database on the local filesystem. It provides
1518 less security than a true HSM, but it allows you to experiment with
1519 native PKCS#11 when an HSM is not available. SoftHSMv2 can be
1520 configured to use either OpenSSL or the Botan library to perform
1521 cryptographic functions, but when using it for native PKCS#11 in
1522 BIND, OpenSSL is required.
1523 </p>
1524 <p>
1525 By default, the SoftHSMv2 configuration file is
1528 This location can be overridden by the SOFTHSM2_CONF environment
1529 variable. The SoftHSMv2 cryptographic store must be installed and
1530 initialized before using it with BIND.
1531 </p>
1534 $ <strong class="userinput"><code> configure --with-crypto-backend=openssl --prefix=/opt/pkcs11/usr --enable-gost </code></strong>
1537 $ <strong class="userinput"><code> /opt/pkcs11/usr/bin/softhsm-util --init-token 0 --slot 0 --label softhsmv2 </code></strong>
1538 </pre>
1539 </div>
1540 </div>
1544 <p>
1545 OpenSSL-based PKCS#11 mode uses a modified version of the
1546 OpenSSL library; stock OpenSSL does not fully support PKCS#11.
1547 ISC provides a patch to OpenSSL to correct this. This patch is
1548 based on work originally done by the OpenSolaris project; it has been
1549 modified by ISC to provide new features such as PIN management and
1550 key-by-reference.
1551 </p>
1552 <p>
1554 the patched OpenSSL, one of which must be chosen at
1555 configuration time. The correct choice depends on the HSM
1556 hardware:
1557 </p>
1559 <li><p>
1560 Use 'crypto-accelerator' with HSMs that have hardware
1561 cryptographic acceleration features, such as the SCA 6000
1562 board. This causes OpenSSL to run all supported
1563 cryptographic operations in the HSM.
1564 </p></li>
1565 <li><p>
1566 Use 'sign-only' with HSMs that are designed to
1567 function primarily as secure key storage devices, but lack
1568 hardware acceleration. These devices are highly secure, but
1569 are not necessarily any faster at cryptography than the
1570 system CPU — often, they are slower. It is therefore
1571 most efficient to use them only for those cryptographic
1572 functions that require access to the secured private key,
1573 such as zone signing, and to use the system CPU for all
1574 other computationally-intensive operations. The AEP Keyper
1575 is an example of such a device.
1576 </p></li>
1577 </ul></div>
1578 <p>
1579 The modified OpenSSL code is included in the BIND 9 release,
1580 in the form of a context diff against the latest versions of
1582 separate diffs for each version. In the examples to follow,
1583 we use OpenSSL 0.9.8, but the same methods work with OpenSSL
1585 </p>
1588 The latest OpenSSL versions as of this writing (January 2015)
1590 ISC will provide updated patches as new versions of OpenSSL
1591 are released. The version number in the following examples
1592 is expected to change.
1593 </div>
1594 <p>
1596 necessary to build OpenSSL with the patch in place, and configure
1597 it with the path to your HSM's PKCS#11 provider library.
1598 </p>
1603 $ <strong class="userinput"><code>wget <a href="" target="_top">http://www.openssl.org/source/openssl-0.9.8zc.tar.gz</a></code></strong>
1604 </pre>
1608 </pre>
1613 </pre>
1616 Note that the patch file may not be compatible with the
1617 "patch" utility on all operating systems. You may need to
1618 install GNU patch.
1619 </div>
1620 <p>
1621 When building OpenSSL, place it in a non-standard
1622 location so that it does not interfere with OpenSSL libraries
1623 elsewhere on the system. In the following examples, we choose
1624 to install into "/opt/pkcs11/usr". We will use this location
1625 when we configure BIND 9.
1626 </p>
1627 <p>
1628 Later, when building BIND 9, the location of the custom-built
1629 OpenSSL library will need to be specified via configure.
1630 </p>
1631 </div>
1635 <p>
1636 The AEP Keyper is a highly secure key storage device,
1637 but does not provide hardware cryptographic acceleration. It
1638 can carry out cryptographic operations, but it is probably
1639 slower than your system's CPU. Therefore, we choose the
1640 'sign-only' flavor when building OpenSSL.
1641 </p>
1642 <p>
1643 The Keyper-specific PKCS#11 provider library is
1644 delivered with the Keyper software. In this example, we place
1645 it /opt/pkcs11/usr/lib:
1646 </p>
1648 $ <strong class="userinput"><code>cp pkcs11.GCC4.0.2.so.4.05 /opt/pkcs11/usr/lib/libpkcs11.so</code></strong>
1649 </pre>
1650 <p>
1651 This library is only available for Linux as a 32-bit
1652 binary. If we are compiling on a 64-bit Linux system, it is
1653 necessary to force a 32-bit build, by specifying -m32 in the
1654 build options.
1655 </p>
1656 <p>
1657 Finally, the Keyper library requires threads, so we
1658 must specify -pthread.
1659 </p>
1663 --pk11-libname=/opt/pkcs11/usr/lib/libpkcs11.so \
1664 --pk11-flavor=sign-only \
1665 --prefix=/opt/pkcs11/usr</code></strong>
1666 </pre>
1667 <p>
1669 and "<span><strong class="command">make test</strong></span>". If "<span><strong class="command">make
1671 add the -pthread above.
1672 </p>
1673 </div>
1677 <p>
1679 library, libpkcs11. It is a true crypto accelerator, up to 4
1680 times faster than any CPU, so the flavor shall be
1681 'crypto-accelerator'.
1682 </p>
1683 <p>
1684 In this example, we are building on Solaris x86 on an
1685 AMD64 system.
1686 </p>
1690 --pk11-libname=/usr/lib/64/libpkcs11.so \
1691 --pk11-flavor=crypto-accelerator \
1692 --prefix=/opt/pkcs11/usr</code></strong>
1693 </pre>
1694 <p>
1696 </p>
1697 <p>
1698 After configuring, run
1701 </p>
1702 </div>
1706 <p>
1707 SoftHSM (version 1) is a software library developed by the
1708 OpenDNSSEC project
1710 http://www.opendnssec.org
1711 </a>)
1712 which provides a
1713 PKCS#11 interface to a virtual HSM, implemented in the form of
1714 a SQLite3 database on the local filesystem. SoftHSM uses
1715 the Botan library to perform cryptographic functions. Though
1716 less secure than a true HSM, it can allow you to experiment
1717 with PKCS#11 when an HSM is not available.
1718 </p>
1719 <p>
1720 The SoftHSM cryptographic store must be installed and
1721 initialized before using it with OpenSSL, and the SOFTHSM_CONF
1722 environment variable must always point to the SoftHSM configuration
1723 file:
1724 </p>
1730 $ <strong class="userinput"><code> export SOFTHSM_CONF=/opt/pkcs11/softhsm.conf </code></strong>
1731 $ <strong class="userinput"><code> echo "0:/opt/pkcs11/softhsm.db" > $SOFTHSM_CONF </code></strong>
1732 $ <strong class="userinput"><code> /opt/pkcs11/usr/bin/softhsm --init-token 0 --slot 0 --label softhsm </code></strong>
1733 </pre>
1734 <p>
1735 SoftHSM can perform all cryptographic operations, but
1736 since it only uses your system CPU, there is no advantage to using
1737 it for anything but signing. Therefore, we choose the 'sign-only'
1738 flavor when building OpenSSL.
1739 </p>
1743 --pk11-libname=/opt/pkcs11/usr/lib/libsofthsm.so \
1744 --pk11-flavor=sign-only \
1745 --prefix=/opt/pkcs11/usr</code></strong>
1746 </pre>
1747 <p>
1750 </p>
1751 </div>
1752 <p>
1753 Once you have built OpenSSL, run
1755 that PKCS#11 support was compiled in correctly. The output
1756 should be one of the following lines, depending on the flavor
1757 selected:
1758 </p>
1760 (pkcs11) PKCS #11 engine support (sign only)
1761 </pre>
1764 (pkcs11) PKCS #11 engine support (crypto accelerator)
1765 </pre>
1766 <p>
1767 Next, run
1769 attempt to initialize the PKCS#11 engine. If it is able to
1770 do so successfully, it will report
1772 </p>
1773 <p>
1774 If the output is correct, run
1777 </p>
1780 <a name="id2638385"></a>Configuring BIND 9 for Linux with the AEP Keyper</h4></div></div></div>
1781 <p>
1782 To link with the PKCS#11 provider, threads must be
1783 enabled in the BIND 9 build.
1784 </p>
1785 <p>
1786 The PKCS#11 library for the AEP Keyper is currently
1787 only available as a 32-bit binary. If we are building on a
1789 the CC options on the "configure" command line.
1790 </p>
1794 --with-openssl=/opt/pkcs11/usr \
1795 --with-pkcs11=/opt/pkcs11/usr/lib/libpkcs11.so</code></strong>
1796 </pre>
1797 </div>
1800 <a name="id2638417"></a>Configuring BIND 9 for Solaris with the SCA 6000</h4></div></div></div>
1801 <p>
1802 To link with the PKCS#11 provider, threads must be
1803 enabled in the BIND 9 build.
1804 </p>
1808 --with-openssl=/opt/pkcs11/usr \
1810 </pre>
1812 <p>
1813 If configure complains about OpenSSL not working, you
1815 incorrectly specified the path to OpenSSL (it should be the
1816 same as the --prefix argument to the OpenSSL
1817 Configure).
1818 </p>
1819 </div>
1826 --with-openssl=/opt/pkcs11/usr \
1827 --with-pkcs11=/opt/pkcs11/usr/lib/libsofthsm.so</code></strong>
1828 </pre>
1829 </div>
1830 <p>
1831 After configuring, run
1835 </p>
1836 <p>
1838 have forgotten to set the SOFTHSM_CONF environment variable.)
1839 </p>
1840 </div>
1844 <p>
1845 BIND 9 includes a minimal set of tools to operate the
1846 HSM, including
1848 within the HSM,
1850 available,
1853 </p>
1854 <p>
1855 In UNIX/Linux builds, these tools are built only if BIND
1856 9 is configured with the --with-pkcs11 option. (Note: If
1857 --with-pkcs11 is set to "yes", rather than to the path of the
1858 PKCS#11 provider, then the tools will be built but the
1859 provider will be left undefined. Use the -m option or the
1860 PKCS11_PROVIDER environment variable to specify the path to the
1861 provider.)
1862 </p>
1863 </div>
1867 <p>
1868 For OpenSSL-based PKCS#11, we must first set up the runtime
1869 environment so the OpenSSL and PKCS#11 libraries can be loaded:
1870 </p>
1872 $ <strong class="userinput"><code>export LD_LIBRARY_PATH=/opt/pkcs11/usr/lib:${LD_LIBRARY_PATH}</code></strong>
1873 </pre>
1874 <p>
1877 rather than from the default location. This step is not necessary
1878 when using native PKCS#11.
1879 </p>
1880 <p>
1881 Some HSMs require other environment variables to be set.
1882 For example, when operating an AEP Keyper, it is necessary to
1883 specify the location of the "machine" file, which stores
1884 information about the Keyper for use by the provider
1885 library. If the machine file is in
1887 use:
1888 </p>
1890 $ <strong class="userinput"><code>export KEYPER_LIBRARY_PATH=/opt/Keyper/PKCS11Provider</code></strong>
1891 </pre>
1892 <p>
1893 Such environment variables must be set whenever running
1894 any tool that uses the HSM, including
1902 </p>
1903 <p>
1904 We can now create and use keys in the HSM. In this case,
1905 we will create a 2048 bit key and give it the label
1906 "sample-ksk":
1907 </p>
1910 </pre>
1914 Enter PIN:
1917 </pre>
1918 <p>
1919 Before using this key to sign a zone, we must create a
1921 does this. In this case, we will be using the HSM key
1923 </p>
1925 $ <strong class="userinput"><code>dnssec-keyfromlabel -l sample-ksk -f KSK example.net</code></strong>
1926 </pre>
1927 <p>
1928 The resulting K*.key and K*.private files can now be used
1929 to sign the zone. Unlike normal K* files, which contain both
1930 public and private key data, these files will contain only the
1931 public key data, plus an identifier for the private key which
1932 remains stored within the HSM. Signing with the private key takes
1933 place inside the HSM.
1934 </p>
1935 <p>
1936 If you wish to generate a second key in the HSM for use
1937 as a zone-signing key, follow the same procedure above, using a
1938 different keylabel, a smaller key size, and omitting "-f KSK"
1939 from the dnssec-keyfromlabel arguments:
1940 </p>
1941 <p>
1942 (Note: When using OpenSSL-based PKCS#11 the label is an arbitrary
1943 string which identifies the key. With native PKCS#11, the label is
1944 a PKCS#11 URI string which may include other details about the key
1945 and the HSM, including its PIN. See
1946 <a href="man.dnssec-keyfromlabel.html" title="dnssec-keyfromlabel"><span class="refentrytitle"><span class="application">dnssec-keyfromlabel</span></span>(8)</a> for details.)
1947 </p>
1950 $ <strong class="userinput"><code>dnssec-keyfromlabel -l sample-zsk example.net</code></strong>
1951 </pre>
1952 <p>
1953 Alternatively, you may prefer to generate a conventional
1954 on-disk key, using dnssec-keygen:
1955 </p>
1958 </pre>
1959 <p>
1960 This provides less security than an HSM key, but since
1961 HSMs can be slow or cumbersome to use for security reasons, it
1962 may be more efficient to reserve HSM keys for use in the less
1963 frequent key-signing operation. The zone-signing key can be
1964 rolled more frequently, if you wish, to compensate for a
1965 reduction in key security. (Note: When using native PKCS#11,
1966 there is no speed advantage to using on-disk keys, as cryptographic
1967 operations will be done by the HSM regardless.)
1968 </p>
1969 <p>
1970 Now you can sign the zone. (Note: If not using the -S
1973 files to the zone master file before signing it.)
1974 </p>
1977 Enter PIN:
1978 Verifying the zone using the following algorithms:
1979 NSEC3RSASHA1.
1980 Zone signing complete:
1982 example.net.signed
1983 </pre>
1984 </div>
1988 <p>
1993 the --with-pkcs11 option, this option defaults to "pkcs11".
1994 Specifying the engine will generally not be necessary unless
1995 for some reason you wish to use a different OpenSSL
1996 engine.
1997 </p>
1998 <p>
2000 for troubleshooting purposes, or because the HSM is unavailable
2001 — set the engine to the empty string. For example:
2002 </p>
2005 </pre>
2006 <p>
2007 This causes
2009 without the --with-pkcs11 option.
2010 </p>
2011 <p>
2013 different meaning: it specifies the path to the PKCS#11 provider
2014 library. This may be useful when testing a new provider library.
2015 </p>
2016 </div>
2020 <p>
2022 using HSM keys, and/or to to sign new records inserted via nsupdate,
2023 then named must have access to the HSM PIN. In OpenSSL-based PKCS#11,
2024 this is accomplished by placing the PIN into the openssl.cnf file
2025 (in the above examples,
2027 </p>
2028 <p>
2029 The location of the openssl.cnf file can be overridden by
2030 setting the OPENSSL_CONF environment variable before running
2031 named.
2032 </p>
2035 openssl_conf = openssl_def
2036 [ openssl_def ]
2037 engines = engine_section
2038 [ engine_section ]
2039 pkcs11 = pkcs11_section
2040 [ pkcs11_section ]
2042 </pre>
2043 <p>
2044 This will also allow the dnssec-* tools to access the HSM
2045 without PIN entry. (The pkcs11-* tools access the HSM directly,
2046 not via OpenSSL, so a PIN will still be required to use
2047 them.)
2048 </p>
2049 <p>
2050 In native PKCS#11 mode, the PIN can be provided in a file specified
2051 as an attribute of the key's label. For example, if a key had the label
2052 <strong class="userinput"><code>pkcs11:object=local-zsk;pin-source=/etc/hsmpin</code></strong>,
2053 then the PIN would be read from the file
2055 </p>
2058 <p>
2059 Placing the HSM's PIN in a text file in this manner may reduce the
2060 security advantage of using an HSM. Be sure this is what you want to
2061 do before configuring the system in this way.
2062 </p>
2063 </div>
2064 </div>
2065 </div>
2069 <p>
2070 DLZ (Dynamically Loadable Zones) is an extension to BIND 9 that allows
2071 zone data to be retrieved directly from an external database. There is
2072 no required format or schema. DLZ drivers exist for several different
2073 database backends including PostgreSQL, MySQL, and LDAP and can be
2074 written for any other.
2075 </p>
2076 <p>
2077 Historically, DLZ drivers had to be statically linked with the named
2078 binary and were turned on via a configure option at compile time (for
2080 Currently, the drivers provided in the BIND 9 tarball in
2082 way.
2083 </p>
2084 <p>
2085 In BIND 9.8 and higher, it is possible to link some DLZ modules
2086 dynamically at runtime, via the DLZ "dlopen" driver, which acts as a
2087 generic wrapper around a shared object implementing the DLZ API. The
2088 "dlopen" driver is linked into named by default, so configure options
2089 are no longer necessary when using these dynamically linkable drivers,
2090 but are still needed for the older drivers in
2092 </p>
2093 <p>
2094 When the DLZ module provides data to named, it does so in text format.
2095 The response is converted to DNS wire format by named. This
2096 conversion, and the lack of any internal caching, places significant
2097 limits on the query performance of DLZ modules. Consequently, DLZ is
2098 not recommended for use on high-volume servers. However, it can be
2099 used in a hidden master configuration, with slaves retrieving zone
2100 updates via AXFR. (Note, however, that DLZ has no built-in support for
2101 DNS notify; slaves are not automatically informed of changes to the
2102 zones in the database.)
2103 </p>
2107 <p>
2110 </p>
2112 dlz example {
2114 search yes;
2115 };
2116 </pre>
2117 <p>
2118 This specifies a DLZ module to search when answering queries; the
2120 loaded at runtime by the dlopen DLZ driver. Multiple
2124 they contain an answer for the query name; the best available
2125 answer will be returned to the client.
2126 </p>
2127 <p>
2130 </p>
2131 <p>
2134 match when a query is received. Instead, zones in this DLZ must be
2135 separately specified in a zone statement. This allows you to
2136 configure a zone normally using standard zone option semantics,
2137 but specify a different database back-end for storage of the
2138 zone's data. For example, to implement NXDOMAIN redirection using
2139 a DLZ module for back-end storage of redirection rules:
2140 </p>
2142 dlz other {
2144 search no;
2145 };
2147 zone "." {
2148 type redirect;
2149 dlz other;
2150 };
2151 </pre>
2152 </div>
2156 <p>
2157 For guidance in implementation of DLZ modules, the directory
2159 dynamically-linkable DLZ module--i.e., one which can be
2160 loaded at runtime by the "dlopen" DLZ driver.
2161 The example sets up a single zone, whose name is passed
2163 statement:
2164 </p>
2166 dlz other {
2167 database "dlopen driver.so example.nil";
2168 };
2169 </pre>
2170 <p>
2171 In the above example, the module is configured to create a zone
2172 "example.nil", which can answer queries and AXFR requests, and
2173 accept DDNS updates. At runtime, prior to any updates, the zone
2174 contains an SOA, NS, and a single A record at the apex:
2175 </p>
2177 example.nil. 3600 IN SOA example.nil. hostmaster.example.nil. (
2178 123 900 600 86400 3600
2179 )
2180 example.nil. 3600 IN NS example.nil.
2182 </pre>
2183 <p>
2184 The sample driver is capable of retrieving information about the
2185 querying client, and altering its response on the basis of this
2186 information. To demonstrate this feature, the example driver
2188 with the source address of the query. Note, however, that this
2189 record will *not* be included in AXFR or ANY responses. Normally,
2190 this feature would be used to alter responses in some other fashion,
2191 e.g., by providing different address records for a particular name
2192 depending on the network from which the query arrived.
2193 </p>
2194 <p>
2195 Documentation of the DLZ module API can be found in
2198 defines the API and should be included by any dynamically-linkable
2199 DLZ module.
2200 </p>
2201 </div>
2202 </div>
2205 <a name="id2571523"></a>IPv6 Support in <acronym class="acronym">BIND</acronym> 9</h2></div></div></div>
2206 <p>
2208 defined forms of IPv6 name to address and address to name
2209 lookups. It will also use IPv6 addresses to make queries when
2210 running on an IPv6 capable system.
2211 </p>
2212 <p>
2214 only AAAA records. RFC 3363 deprecated the use of A6 records,
2215 and client-side support for A6 records was accordingly removed
2218 load zone files containing A6 records correctly, answer queries
2219 for A6 records, and accept zone transfer for a zone containing A6
2220 records.
2221 </p>
2222 <p>
2224 the traditional "nibble" format used in the
2229 but support of binary labels has been completely removed per
2230 RFC 3363.
2232 the binary label format at all any more, and will return an
2233 error if given.
2235 name server will not load a zone file containing binary labels.
2236 </p>
2237 <p>
2238 For an overview of the format and structure of IPv6 addresses,
2239 see <a href="Bv9ARM.ch11.html#ipv6addresses" title="IPv6 addresses (AAAA)">the section called “IPv6 addresses (AAAA)”</a>.
2240 </p>
2244 <p>
2245 The IPv6 AAAA record is a parallel to the IPv4 A record,
2246 and, unlike the deprecated A6 record, specifies the entire
2247 IPv6 address in a single record. For example,
2248 </p>
2250 $ORIGIN example.com.
2252 </pre>
2253 <p>
2254 Use of IPv4-in-IPv6 mapped addresses is not recommended.
2255 If a host has an IPv4 address, use an A record, not
2257 the address.
2258 </p>
2259 </div>
2263 <p>
2264 When looking up an address in nibble format, the address
2265 components are simply reversed, just as in IPv4, and
2267 resulting name.
2268 For example, the following would provide reverse name lookup for
2269 a host with address
2271 </p>
2274 1.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0 14400 IN PTR (
2275 host.example.com. )
2276 </pre>
2277 </div>
2278 </div>
2279 </div>
2281 <hr>
2283 <tr>
2288 </td>
2289 </tr>
2290 <tr>
2293 <td width="40%" align="right" valign="top"> Chapter 5. The <acronym class="acronym">BIND</acronym> 9 Lightweight Resolver</td>
2294 </tr>
2295 </table>
2296 </div>
2298 </body>
2299 </html>