Debian Codegraph @

rfc1276.html

raw
1 <pre>Network Working Group S.E. Hardcastle-Kille
2 Requests for Comments 1276 University College London
3 November 1991
4
5
6
7
8
9
10 <span class="h1">Replication and Distributed Operations extensions</span>
11 <span class="h1">to provide an Internet Directory using X.500</span>
12
13
14
15
16
17
18
19 Status of this Memo
20 This RFC specifies an IAB standards track protocol for the
21 Internet community, and requests discussion and suggestions for
22 improvements. Please refer to the current edition of the ``IAB
23 Official Protocol Standards'' for the standardization state and
24 status of this protocol. Distribution of this memo is unlimited.
25
26 Abstract
27 Some requirements on extensions to X.500 are described in the
28 RFC[HK91b], in order to build an Internet Directory using
29 X.500(1988). This document specifies a set of solutions to the
30 problems raised. These solutions are based on some work done for
31 the QUIPU implementation, and demonstrated to be effective in a
32 number of directory pilots. By documenting a de facto standard,
33 rapid progress can be made towards a full-scale pilot. These
34 procedures are an INTERIM approach. There are known
35 deficiencies, both in terms of manageability and scalability.
36 Transition to standard approaches are planned when appropriate
37 standards are available. This RFCwill be obsoleted at this
38 point.</pre>
39 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
40
41
42 Contents
43
44 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a> Approach </span> 2
45
46 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a> Extensions to Distributed Operations </span> 3
47
48
49 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a> Alternative DSAs </span> 4
50
51 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a> Data Model </span> 5
52
53
54 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a> DSA Naming </span> 6
55
56 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a> Knowledge Representation </span> 6
57
58 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a> Replication Protocol </span> 9
59
60
61 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a> New Application Context </span> 12
62
63 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a> Policy on Replication Procedures </span> 12
64
65
66 <span class="h2"><a class="selflink" id="section-10" href="#section-10">10</a> Use of the Directory by Applications </span> 12
67
68 <span class="h2"><a class="selflink" id="section-11" href="#section-11">11</a> Migration and Scaling </span> 12
69
70 <span class="h2"><a class="selflink" id="section-12" href="#section-12">12</a> Security Considerations </span> 13
71
72
73 <span class="h2"><a class="selflink" id="section-13" href="#section-13">13</a> Author's Address </span> 13
74
75 A ASN.1 Summary and Object Identifier Allocation 14
76
77
78 List of Figures
79
80 1 Knowledge Attributes . . . . . . . . 8
81
82 2 Replication Protocol . . . . . . . . 10
83 3 Summary of the ASN.1 . . . . . . . . 17
84
85
86
87 Hardcastle-Kille Page 1</pre>
88 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
89
90
91 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a> Approach</span>
92
93 There are a number of non-negotiable requirements which must be met
94 before a directory can be deployed on the Internet [<a href="#ref-HK91b" title="Department of Computer Science">HK91b</a>]. These
95 problems are being tackled in the standards arena, but there is
96 currently no stable solution. One approach would be to attempt to
97 intercept the standard. Difficulties with this would be:
98
99
100 o Defining a coherent intercept would be awkward, and the effort
101 would probably be better devoted to working on the standard. It
102 is not even clear that such an intercept could be defined.
103
104 o The target is moving, and it is always tempting to track it, thus
105 causing more delay.
106
107 o There would be a delay involved with this approach. It would be
108 too late to be useful for a rapid start, and sufficiently close to
109 the timing of the final standard that many would choose not to
110 implement it.
111
112 Therefore, we choose to take a simple approach. This is a good deal
113 simpler than the full X.500 approach, and is based on operational
114 experience. The advantages of this approach are:
115
116
117 o It is proven in operation. This RFCis simply documenting what is
118 being done already.
119
120 o There will be a minimum of delay in starting to use the approach.
121
122 o The approach is simpler, and so the cost of implementation is much
123 less. It will therefore be much more attractive to add into an
124 implementation, as it is less effort, and can be further ahead of
125 the standard.
126
127 These procedures are an INTERIM approach. There are known
128 deficiencies, both in terms of manageability and scalability.
129 Transition to standard approaches are planned when appropriate
130 standards are available. This RFCwill be obsoleted at this point.
131
132
133
134
135
136 Hardcastle-Kille Page 2</pre>
137 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
138
139
140 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a> Extensions to Distributed Operations</span>
141
142 The distributed operations of X.500 assume that all DUAs and DSAs are
143 fully interconnected with a global network service. For the Internet
144 Pilot, this assumption is invalid. DSAs may be operated over TCP/IP,
145 TP4/CLNS, or TP0/CONS.
146 The extension to distributed operations to support this situation is
147 straightforward. We define the term community as an environment where
148 direct (network) communication is possible. Communities may be
149 separated because they operate different protocols, or because of lack
150 of physical connectivity. Example communities are the DARPA/NSF
151 Internet, and the Janet private X.25 network. A network entity in a
152 community is addressed by its Network Address. If two network
153 entities are in the same community, they can by definition
154 communicate. A community is identified by a set of network address
155 prefixes. For the approach to be useful, this set should be small
156 (typically 1). For TCP/IP Networks, and X.25 Networks not providing
157 CONS, the approach is described in [<a href="#ref-HK91a" title="Department of Computer Science">HK91a</a>] allows for communities to
158 be defined for the networks of operational interest.
159
160 This model can be used to determine whether a pair of application
161 entities can communicate. For each entity, determine the presentation
162 address (typically by directory lookup). Each network address in the
163 presentation address will have a single associated community. The set
164 of communities to which each application entity belongs can thus be
165 determined. If the two application entities have a common community,
166 then they can communicate directly.
167 Two extensions to the standard distributed operations are needed.
168
169
170 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Consider a DSA (the local DSA) which is contacted by either a DUA</span>
171 <span class="h2"> or DSA (the calling entity) to resolve a query. The local DSA</span>
172 determines that the query must be progressed by another DSA (the
173 referred-to DSA). The DSA will make a chain/referral choice. If
174 chaining is prohibited by service control, a referral will be
175 passed back. Otherwise, if the local DSA prefers to chain (e.g.,
176 for policy reasons) it will then chain. The remaining situation
177 is that the local DSA prefers to give a referral. It shall only
178 do so if it believes that the calling entity can directly connect
179 to the referred-to DSA. If the calling entity is a DUA, it should
180 be assumed to belong only to the community of the called network
181 address. If the calling entity is a DSA, its communities should
182 be determined by lookup of the DSA's presentation address in the
183 directory. The communities of the referred-to DSA can be
184
185 Hardcastle-Kille Page 3</pre>
186 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
187
188
189 determined from its presentation address, which will either be
190 present in the reference or can be looked up in the directory. If
191 the calling entity and the referred-to DSA do not have a common
192 community, then chaining shall be used. Otherwise, a referral may
193 be passed back to the calling entity.
194
195 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Consider that a DSA (or DUA), termed here the local entity is</span>
196 <span class="h2"> following a referral (to a referred-to DSA). In some cases, the</span>
197 local entity and referred-to DSA will not be able to communicate
198 directly (i.e., not have a common community). There are two
199 approaches to solve this:
200
201 (a) Pass the query to a DSA it would use to resolve a query for
202 the entry one level higher in the DIT. This will work,
203 provided that this DSA follows this specification. This
204 default mechanism will work without additional configuration.
205
206 (b) Use a ``relay DSA'' to access the community. A relay DSA is
207 one which can chain the query on to the remote community. The
208 relay DSA must belong to both the remote community and to at
209 least one community to which the local entity belongs. The
210 choice of relay DSA for a given community will be manually
211 configured by a DSA manager to enable access to a community to
212 which there is not direct connectivity. Typically this will
213 be used where the default DSA is a poor choice (e.g., because
214 relaying is not authorised through this DSA).
215
216 A DSA conforming to this specification shall follow these
217 procedures. A DUA may also follow these procedures, and this will
218 give improvements in some circumstances (i.e., the ability to
219 resolve certain queries without use of chaining). However, this
220 specification does not place requirements on DUAs.
221
222
223 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a> Alternative DSAs</span>
224
225 There is a need to give information on slave copies of data. This can
226 be done using the standard protocol, but modifying the semantics.
227 This relies on the fact that there may only be a single subordinate
228 reference or cross reference.
229
230 If there is a need to include references to master and slave data (EDB
231 copies) in a referral, then this should be done in a referral by
232 specifying a subordinate reference with multiple values. This cannot
233
234 Hardcastle-Kille Page 4</pre>
235 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
236
237
238 be a standard subordinate reference, which would only have a single
239 value. Therefore, this usage does not conflict with standard
240 references. The first reference is the master copy, and subsequent
241 references are slave copies.
242
243
244 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a> Data Model</span>
245
246 The X.500 data model takes the unit of mastering data as the entry. A
247 DSA may hold an arbitrary collection of entries. We restrict this
248 model so that for the replication protocol defined in this
249 specification the base unit of replication (shadowing) is the complete
250 set of immediate subordinate entries of a given entry, termed an Entry
251 Data Block (EDB). An EDB is named by its parent entry. It contains
252 the relative distinguished names of all of the children of the entry,
253 and each of the child entries. For each entry, this comprises all
254 attributes of the entry, the relative distinguished name, and
255 knowledge information associated with the entry. If a DSA holds
256 (non-cached) information on an entry, it will hold information on all
257 of its siblings. One DSA will hold a master EDB. This will contain
258 two types of entry:
259
260 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Entries for which this DSA is the master.</span>
261
262 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Slave copies of entries which are mastered in another DSA,</span>
263 <span class="h2"> indicated by a subordinate reference. This copy must be</span>
264 maintained automatically by the DSA holding the master EDB.
265
266
267 Thus the master EDB contains a mixture of master entries, and entries
268 which are mastered elsewhere and shadowed by the DSA holding the
269 master EDB on an entry by entry basis. Other DSAs may hold slave
270 copies of this EDB (slave EDBs), which are replicated in their
271 entirity directly or indirectly from the master EDB. This approach has
272 the following advantages.
273
274 o Name resolution is simplified, and performance improved.
275
276 o Single level searching and listing have good performance, and are
277 straightforward to implement. In a more general case of applying
278 the standard, without sophisticated replication, these operations
279 might require to access very many DSAs and be prohibitively
280 expensive.
281
282
283 Hardcastle-Kille Page 5</pre>
284 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
285
286
287 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a> DSA Naming</span>
288
289 All DSAs must be named in the DIT, and the master definition of the
290 presentation address stored in this entry. X.500 (including some of
291 the extension work) implies that the presentation address information
292 is extensively replicated (manually). The management overhead implied
293 by this is not acceptable.
294 Care must be taken to prevent deadlock in determining a DSAs address.
295 This is solved by:
296
297
298 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Use of a well known DSA with ``root knowledge''</span>
299
300 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Naming DSAs in a manner which prevents deadlocks. </span>Currently this
301 is done by giving DSAs names high in the DIT.
302
303 The Internet Pilot will need to define detailed policies for naming
304 DSAs, in conjunction with the replication policy. This will be
305 defined in a future RFC.
306
307
308 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a> Knowledge Representation</span>
309
310
311 Knowledge information is represented in the DIT. It seems unreasonable
312 to manage this by any other means. Knowledge information is
313 represented in an entry by use of knowledge attributes. These
314 attributes are considered separately from all the other attributes in
315 the entry which are termed ``user attributes''. Each entry in a
316 master EDB will be in one of four categories.
317
318 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. The entry is a leaf entry mastered in this EDB, and so only</span>
319 <span class="h2"> contains user attributes</span>
320
321 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. The level below has an associated EDB (i.e., the DIT continues</span>
322 <span class="h2"> downwards to use the data model of this specification). All</span>
323 attributes of this entry will be mastered in this entry. The
324 entry will contain an attribute with the name of the DSA which
325 holds the master of the associated EDB. Optionally, it will
326 contain an attribute holding the names of DSAs which hold slave
327 EDBs. The entry may not hold a subordinate reference attribute.
328 The DIT is followed by use of the master and slave attributes.
329
330
331
332 Hardcastle-Kille Page 6</pre>
333 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
334
335
336 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. The entry is mastered in a DSA which does not follow this</span>
337 <span class="h2"> specification. The entry in the EDB will contain a master</span>
338 attribute, which holds a subordinate reference (or cross
339 reference) to the DSA which holds the master entry. The user
340 attributes of the entry will be mastered in the DSA pointed to by
341 the reference. The DSA holding the master EDB, which actually
342 acts as an intermediate shadow for this entry, will read these
343 attributes from the DSA indicated by the reference, so that it
344 will have a full copy of the entry, using a standared DSP Read
345 operation. This technique is called ``spot shadowing''. Any
346 access control on the entry being spot shadowed must be configured
347 so that all attributes can be copied by the DSA holding the master
348 EDB. DSAs taking slave copies of the EDB will not do spot
349 shadowing. However, the knowledge attributes will be copied, and
350 may be used by this DSA (e.g., for modify operations).
351
352 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. The entries at the level below are held in DSAs which do not</span>
353 <span class="h2"> follow this specification, and all of these are indicated by a set</span>
354 of NSSRs (Non Specific Subordinate Reference). The NSSRs are
355 stored as an attribute of the entry. The user attributes are
356 either mastered in the EDB.
357 It is important to note that NSSRs are stored at the level above
358 subordinate references. At a given point in the DIT, if there are
359 subordinate references, these are stored in shadow entries below
360 that point, and named by the RDN. If there are NSSRs, they are
361 stored in the entry itself, as there is no RDN associated with an
362 NSSR. This approach is cleanest where there are either NSSRs or
363 subordinate references, but not both. For example, consider an
364 Organisation HP, whose many OUs are stored in a set of DSAs
365 indicated by by NSSRs. Here, the NSSR attributes will be used to
366 identify these DSAs.
367 This model of replication is not tightly integrated with NSSRs.
368 Where there is a mixture of NSSRs and Subordinate references at a
369 given point in the DIT, this is handled by giving a single
370 subordinate reference to a DSA which follows standard X.500
371 distributed operations and can cleanly handle this mixture. In
372 practice, this is equivalent to not allowing a mixture of
373 subordinate references and NSSRs.
374
375 The information framework needed to support this is defined in
376 Figure_1.______________________________________________________________
377
378
379
380
381 Hardcastle-Kille Page 7</pre>
382 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
383
384
385
386 InternetDSNonLeafObject ::= OBJECT-CLASS
387 SUBCLASS OF top
388 MUST CONTAIN {masterDSA}
389 MAY CONTAIN {slaveDSA}
390
391 ExternalDSObject ::= OBJECT-CLASS
392 SUBCLASS OF top
393 MAY CONTAIN {SubordinateReference, CrossReference, 10
394 NonSpecificSubordinateReference}
395 -- will contain exactly one of these references
396
397 MasterDSA ::= ATTRIBUTE
398 WITH ATTRIBUTE-SYNTAX distinguishedNameSyntax
399 SINGLE VALUE
400
401 SlaveDSA ::= ATTRIBUTE
402 WITH ATTRIBUTE-SYNTAX distinguishedNameSyntax
403 20
404 SubordinateReference ::= ATTRIBUTE
405 WITH ATTRIBUTE-SYNTAX AccessPoint
406 SINGLE VALUE
407
408 CrossReference ::= ATTRIBUTE
409 WITH ATTRIBUTE-SYNTAX AccessPoint
410 SINGLE VALUE
411
412 NonSpecificSubordinateReference ::= ATTRIBUTE
413 WITH ATTRIBUTE-SYNTAX AccessPoint 30
414
415 AccessPoint ::= SET {
416 ae-title [0] Name,
417 address [2] PresentationAddress OPTIONAL }
418
419 -- Same definition as X.500 AccessPoint,
420 -- but presentation address is optional
421
422
423 ___________________Figure_1:__Knowledge_Attributes_____________________
424
425 Two object classes are defined to support this approach:
426
427
428
429
430 Hardcastle-Kille Page 8</pre>
431 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
432
433
434 InternetDSNonLeafObject This is for where the level below follows the
435 model defined here, and there is an Entry Data Block (EDB)
436 containing the sibling entries. The Entry itself contains master
437 data. The associated attributes are:
438
439 MasterDSA The name of the DSA where the master EDB is held.
440
441 SlaveDSA The names of DSAs which hold slave copies of the EDB for
442 public access.
443
444 ExternalDSObject This is for where the entry and levels below are
445 mastered according to X.500. There are attributes corresponding
446 to the standard knowledge references, which are used to resolve
447 queries. The presentation address is optional in these
448 attributes. If not present, it should be looked up in the DSAs
449 own entry. For NonSpecificSubordinateReference, the master of the
450 entry will be in the master EDB, For SubordinateReference or
451 CrossReference1 the DSA which masters the EDB will ``spot shadow''
452 the entry, by reading it at intervals. This will ensure that the
453 master EDB contains a copy of each entry. Single level searching
454 can then be done efficiently where it is not required to access
455 the master copy of the data. DSAs holding slave copies of the EDB
456 do not perform spot shadowing, but do receive copies of the
457 references.
458
459
460 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a> Replication Protocol</span>
461 _______________________________________________________________________
462 GetEntryDataBlock ABSTRACT-OPERATION
463 ARGUMENT GetEntryDataBlockArgument
464 RESULT GetEntryDataBlockResult
465 ERRORS {nameError,ServiceError,SecurityError,EDBVersionError}
466
467 EDBVersionError ABSTRACT-ERROR
468 PARAMETER versionHeld EDBVersion
469
470
471 GetEntryDataBlockArgument ::= SET { 10
472
473 ----------------------------
474 1. These references are really the same. The function and value
475 are the same. The name depends on where the reference is stored. It
476 may be preferable to have only one attribute.
477
478
479 Hardcastle-Kille Page 9</pre>
480 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
481
482
483 entry [0] DistinguishedName,
484 CHOICE {
485 sendIfMoreRecentThan [1] EDBVersion,
486 getVersionNumber [2] NULL,
487 getEDB [3] NULL, -- force retrieval
488 continuation [4] SEQUENCE {
489 EDBVersion,
490 nextEntryPosition INTEGER }
491 },
492 maxEntries [5] INTEGER OPTIONAL 20
493 -- if omitted return whole EDB in
494 -- one operation
495 }
496
497 GetEntryDataBlockResult ::= SEQUENCE {
498 versionHeld [0] EDBVersion,
499 [1] SEQUENCE OF RelativeEntry OPTIONAL,
500 -- if omitted, only version is returned
501 nextEntryPostion INTEGER OPTIONAL
502 -- if omitted there are no more entries 30
503 }
504
505
506
507 RelativeEntry ::= SEQUENCE {
508 RelativeDistinguishedName,
509 SET OF Attribute
510 }
511
512 EDBVersion ::= UTCTime 40
513
514 ___________________Figure_2:__Replication_Protocol_____________________
515
516 A ROS operation to support replication is defined in Figure 2. This
517 pulls an entire copy of the EDB. In normal use, the initiator
518 specifies the EDB Version held. If the responder has a more recent
519 version, then all of the entries in the EDB are returned. There are
520 options to rerequest only the version of EDB held, or to return the
521 full EDB irrespective of the version held by the initiator.
522 For large EDBs, transfer of an entire EDB in a single operation would
523 lead to very large ROS PDUs. This gives a definite scaling
524 limitation. To overcome this, the protocol allows an EDB to be
525 retrived in chunks of a size (in number of entries) specified by the
526
527
528 Hardcastle-Kille Page 10</pre>
529 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
530
531
532 initiator. The responder specifies a number which indicates the next
533 entry to be transferred. The same operation can be used to retrieve
534 the next chunk of the EDB, with EDBVersion and the same integer as
535 parameters.
536 This approach is simple to implement. It is less efficient than an
537 incremental technique. When scaling dictates that an incremental
538 technique must be used, it is expected that a suitable standard will
539 be available.
540 An implementation issue that must be noted is how to deal with updates
541 whilst a multi-operation transfer is in progress. There are two
542 possible approaches:
543
544
545 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Refuse/block updates until the EDB is transferred. </span>This may cause
546 problems where the rate of update and transfer is high, as this
547 may make update very difficult (for the manager).
548
549 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Create a new version of the EDB, whilst retaining the old EDB to</span>
550 <span class="h2"> complete the bulk transfer. A suitable retentions strategy would</span>
551 be to hold an EDB version as long as the association on which it
552 is being pulled it remains active.
553
554 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Allow the update and fail subsequent transfer requests for the</span>
555 <span class="h2"> EDB. This may cause both transfer failure and excessive waste of</span>
556 bandwidth due to retries if the rate of update and transfer is
557 high.
558
559 If option 1. or 3. is chosen, for a widely replicated EDB where the
560 update rate is greater than a few changes per day, it is recommended
561 to configure the master EDB in a DSA which only replicates to one
562 other DSA. This second DSA can then control its update rate, and
563 safely perform a large fanout of replications (option 3). The first
564 DSA will have reasonable availability for modifications (option 1).
565
566 This protocol will be used by DSAs to obtain copies of EDBs high in
567 the tree (typically root and national EDBs). DSAs which need these
568 copies should establish bilateral agreements to access them2.
569 This protocol should only transfer user attributes. In particular,
570 implementation specific attributes such as those needed to support
571
572 ----------------------------
573 2. QUIPU defines some attributes to register such agreements, but
574 these are probably not appropriate for this specification.
575
576
577 Hardcastle-Kille Page 11</pre>
578 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
579
580
581 private access control should not be transferred. There may be
582 bilateral agreements on access control policy of the information
583 (e.g., size limits on listing), which are implemented by (different)
584 system specific techniques.
585
586
587 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a> New Application Context</span>
588
589 A DSA which follows these procedures will support a new
590 ApplicationContext ``Internet DSP'' defined in <a href="#appendix-A">Appendix A</a>. This will
591 be stored in the DSAs entry, so that support of the extensions defined
592 here can easily be determined.
593
594
595 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a> Policy on Replication Procedures</span>
596
597 To be effective, a directory configuration must be laid out. These
598 protocols will need to be used in the framework of a pilot, and
599 service providers making available data for replication.
600 There is a requirement to manage the replication process. This can be
601 done by a combination of local configuration (to register shadowing
602 agreements) and directory operations to set pointers to master and
603 slave copies of the data.
604
605
606 <span class="h2"><a class="selflink" id="section-10" href="#section-10">10</a> Use of the Directory by Applications</span>
607
608
609 Care must be taken by users of the directory when replication is
610 available. This is not a change from current use of X.500, but is
611 noted here as it is important. Normal read requests should allow use
612 of copy information. If the user of the directory believes that
613 information may be out of date (e.g., because an association could not
614 be established), then the request should be repeated and use of copy
615 data prohibited by service controls.
616
617
618 <span class="h2"><a class="selflink" id="section-11" href="#section-11">11</a> Migration and Scaling</span>
619
620 The major scaling limit of this approach is the non-incremental
621 update. This will put a limit on the maximum DIT fanout which can be
622 supported. Given an average entry size of around a thousand bytes,
623 and a maximum reasonable transfer size is tens of megabytes, then the
624
625
626 Hardcastle-Kille Page 12</pre>
627 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
628
629
630 fanout limit of this approach is of order 10 000. Note that smaller
631 organisations will tend to be registered geographically (e.g., in the
632 US, by State), so that the limit of the number of Organisations is
633 somewhat larger. It should be noted that although the replication
634 technique described here is general, it is only intended for high
635 levels of the DIT. These figures assume this.
636 These techniques do not preclude use of other techniques for
637 replication. It would be quite reasonable to replicate data using
638 this approach, and that which will be defined in X.500(92).
639
640
641 References
642
643 [<a id="ref-HK91a">HK91a</a>] S.E. Hardcastle-Kille. Encoding network addresses to support
644 operation over non-osi lower layers. Request for Comments
645 <a href="./rfc1277">RFC 1277</a>, Department of Computer Science, University College
646 London, November 1991.
647
648 [<a id="ref-HK91b">HK91b</a>] S.E. Hardcastle-Kille. Replication requirement to provide an
649 internet directory using X.500. Request for Comments
650 <a href="./rfc1275">RFC 1275</a>, Department of Computer Science, University College
651 London, November 1991.
652
653
654 <span class="h2"><a class="selflink" id="section-12" href="#section-12">12</a> Security Considerations</span>
655
656 Security considerations are not discussed in this memo.
657
658
659 <span class="h2"><a class="selflink" id="section-13" href="#section-13">13</a> Author's Address</span>
660
661 Steve Hardcastle-Kille
662 Department of Computer Science
663 University College London
664 Gower Street
665 WC1E 6BT
666 England
667
668
669 Phone: +44-71-380-7294
670
671 EMail: [email protected]
672
673
674
675 Hardcastle-Kille Page 13</pre>
676 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
677
678
679 A ASN.1 Summary and Object Identifier Allocation
680
681 There_are_a_few_object_identifiers_needed.__These_are_defined_here.____
682
683 InternetDSP TAGS ::=
684 BEGIN
685
686 IMPORTS
687 APPLICATION-SERVICE-ELEMENT, PORT, APPLICATION-CONTEXT,
688 aCSE, ABSTRACT OPERATION
689 FROM Remote-Operations-Notation-extension {joint-iso-ccitt
690 remote-operations(4) notation-extension(2)}
691
692 10
693 id-as-mrse, id-as-mase, id-as-ms
694 FROM MTSAccessProtocol {joint-iso-ccitt mhs-motis(6)
695 protocols(0) modules(0) object-identifiers(0)}
696
697 chainedReadASE, chainedSearchASE, chainedModifyASE
698 FROM DirectorySystemProtocol {joint-iso-ccitt ds(5)
699 modules(1) dsp(12)}
700
701 DistinguishedName, RelativeDistinguishedName, Attribute
702 FROM InformationFramework {joint-iso-ccitt ds(5) 20
703 modules(1) InformationFramework(1)}
704
705
706 ATTRIBUTE, OBJECT-CLASS
707 FROM InformationFramework {joint-iso-ccitt ds(5)
708 modules(1) informationFramework(1)};
709
710
711
712 internet-dsp OBJECT IDENTIFIER ::= {ccitt data(9) pss(2342) 30
713 ucl(19200300) internet-dsp(107)}
714
715 -- General
716
717 at OBJECT IDENTIFIER ::= {internet-dsp at(1)}
718 oc OBJECT IDENTIFIER ::= {internet-dsp oc(2)}
719
720
721 -- Object Classes needed for association
722
723
724 Hardcastle-Kille Page 14</pre>
725 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
726
727
728 40
729 id-ac-idsp OBJECT IDENTIFIER ::= {internet-dsp ac-idsp(3))}
730 id-as-idsp OBJECT IDENTIFIER ::= {internet-dsp as-idsp(4))}
731 id-ase-replication OBJECT IDENTIFIER ::= {internet-dsp ase-replication(5))}
732
733
734 -- Attribute Types
735
736 master-dsa MasterDSA ::= {at 1}
737 slave-dsa SlaveDSA ::= {at 2}
738 subordinate-reference SubordinateReference ::= {at 3} 50
739 cross-reference CrossReference ::= {at 4}
740 nssr NonSpecificSubordinateReference ::= {at 5}
741
742 -- Object Classes
743
744 internet-ds-non-leaf-object InternetDSNonLeafObject ::= {oc 1}
745 external-ds-object ExternalDSObject ::= {oc 2}
746
747
748 -- Operation and Error bindings 60
749
750 getEntryDataBlock GetEntryDataBlock ::= 10
751
752 eDBVersionError EDBVersionError ::= 10
753
754
755 -- Protocol Definitions
756
757 replicationASE APPLICATION-SERVICE-ELEMENT
758 OPERATIONS {getEntryDataBlock} 70
759 ::= id-ase-replication
760
761 internet-dsp APPLICATION-CONTEXT
762 APPLICATION SERVICE ELEMENTS {aCSE}
763 BIND MSBind
764 UNBIND MSUnbind
765 REMOTE OPERATIONS {rOSE}
766 OPERATIONS OF { chainedReadADSm chainedSearchASE,
767 chainedModifyASE, replicationASE }
768 ABSTRACT SYNTAXES { 80
769 id-as-acse,
770 id-as-idsp }
771 ::= id-ac-idsp
772
773 Hardcastle-Kille Page 15</pre>
774 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
775
776
777
778
779
780
781
782
783 90
784 InternetDSNonLeafObject ::= OBJECT-CLASS
785 SUBCLASS OF top
786 MUST CONTAIN {masterDSA}
787 MAY CONTAIN {slaveDSA}
788
789 ExternalDSObject ::= OBJECT-CLASS
790 SUBCLASS OF top
791 MAY CONTAIN {SubordinateReference, CrossReference,
792 NonSpecificSubordinateReference}
793 -- will contain exactly one of these references100
794
795 MasterDSA ::= ATTRIBUTE
796 WITH ATTRIBUTE-SYNTAX distinguishedNameSyntax
797 SINGLE VALUE
798
799 SlaveDSA ::= ATTRIBUTE
800 WITH ATTRIBUTE-SYNTAX distinguishedNameSyntax
801
802 SubordinateReference ::= ATTRIBUTE
803 WITH ATTRIBUTE-SYNTAX AccessPoint 110
804 SINGLE VALUE
805
806 CrossReference ::= ATTRIBUTE
807 WITH ATTRIBUTE-SYNTAX AccessPoint
808 SINGLE VALUE
809
810 NonSpecificSubordinateReference ::= ATTRIBUTE
811 WITH ATTRIBUTE-SYNTAX AccessPoint
812
813 AccessPoint ::= SET { 120
814 ae-title [0] Name,
815 address [2] PresentationAddress OPTIONAL }
816
817 -- Same definition as X.500 AccessPoint,
818 -- but presentation address is optional
819
820 GetEntryDataBlock ABSTRACT-OPERATION
821
822 Hardcastle-Kille Page 16</pre>
823 <hr class='noprint'/><!--NewPage--><pre class='newpage'><a href="./rfc1276">RFC 1276</a> Internet Directory Replication November 1991
824
825
826 ARGUMENT GetEntryDataBlockArgument
827 RESULT GetEntryDataBlockResult
828 ERRORS {nameError,ServiceError,SecurityError,EDBVersionError}130
829
830 EDBVersionError ABSTRACT-ERROR
831 PARAMETER versionHeld EDBVersion
832
833
834 GetEntryDataBlockArgument ::= SET {
835 entry [0] DistinguishedName,
836 CHOICE {
837 sendIfMoreRecentThan [1] EDBVersion,
838 getVersionNumber [2] NULL, 140
839 getEDB [3] NULL, -- force retrieval
840 continuation [4] SEQUENCE {
841 EDBVersion,
842 nextEntryPosition INTEGER }
843 },
844 maxEntries [5] INTEGER OPTIONAL
845 -- if omitted return whole EDB in
846 -- one operation
847 }
848 150
849 GetEntryDataBlockResult ::= SEQUENCE {
850 versionHeld [0] EDBVersion,
851 [1] SEQUENCE OF RelativeEntry OPTIONAL,
852 -- if omitted, only version is returned
853 nextEntryPostion INTEGER OPTIONAL
854 -- if omitted there are no more entries
855 }
856
857
858 160
859 RelativeEntry ::= SEQUENCE {
860 RelativeDistinguishedName,
861 SET OF Attribute
862 }
863
864 EDBVersion ::= UTCTime
865 END
866
867 ___________________Figure_3:__Summary_of_the_ASN.1_____________________
868
869
870
871 Hardcastle-Kille Page 17
872 </pre>