Debian Codegraph @

rfc2071.html

raw
1 <pre>Network Working Group P. Ferguson
2 Request for Comments: 2071 cisco Systems, Inc.
3 Category: Informational H. Berkowitz
4 PSC International
5 January 1997
6
7 <span class="h1">Network Renumbering Overview:</span>
8 <span class="h1">Why would I want it and what is it anyway?</span>
9
10 Status of this Memo
11
12 This memo provides information for the Internet community. This memo
13 does not specify an Internet standard of any kind. Distribution of
14 this memo is unlimited.
15
16 Abstract
17
18 The PIER [Procedures for Internet/Enterprise Renumbering] working
19 group is compiling a series of documents to assist and instruct
20 organizations in their efforts to renumber. However, it is becoming
21 apparent that, with the increasing number of new Internet Service
22 Providers (ISP's) and organizations getting connected to the Internet
23 for the first time, the concept of network renumbering needs to be
24 further defined. This document attempts to clearly define the
25 concept of network renumbering and discuss some of the more pertinent
26 reasons why an organization would have a need to do so.
27
28 Table of Contents
29
30 <a href="#section-1">1</a>. Introduction . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-2">2</a>
31 <a href="#section-2">2</a>. Background . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-2">2</a>
32 <a href="#section-3">3</a>. Network Renumbering Defined. . . . . . . . . . . . . . . . . <a href="#page-3">3</a>
33 <a href="#section-4">4</a>. Reasons for Renumbering. . . . . . . . . . . . . . . . . . . <a href="#page-3">3</a>
34 <a href="#section-5">5</a>. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-12">12</a>
35 <a href="#section-6">6</a>. Security Considerations . . . . . . . . . . . . . . . . . . <a href="#page-12">12</a>
36 <a href="#section-7">7</a>. Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . <a href="#page-12">12</a>
37 <a href="#section-8">8</a>. References . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
38 <a href="#section-9">9</a>. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . <a href="#page-14">14</a>
39
40
41
42
43
44
45
46
47
48
49
50
51
52 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 1]</span></pre>
53 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span>
54 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
55
56
57 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span>
58
59 The popularity of connecting to the global Internet over the course
60 of the past several years has spawned new problems; what most people
61 casually refer to as "growing pains" can be attributed to more basic
62 problems in understanding the requirements for Internet connectivity.
63 However, the reasons why organizations may need to renumber their
64 networks can greatly vary. We'll discuss these issues in some amount
65 of detail below. It is not within the intended scope of this
66 document to discuss renumbering methodologies, techniques, or tools.
67
68 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Background</span>
69
70 The ability for any network or interconnected devices, such as
71 desktop PCs or workstations, to obtain connectivity to any potential
72 destination in the global Internet is reliant upon the possession of
73 unique IP host addresses [<a href="#ref-1" title="&quot;Unique Addresses are Good&quot;">1</a>]. A duplicate host address that is being
74 used elsewhere in the Internet could best be described as
75 problematic, since the presence of duplicate addresses would cause
76 one of the destinations to be unreachable from some origins in the
77 Internet. It should be noted, however, that globally unique IP
78 addresses are not always necessary, and is dependent on the
79 connectivity requirements [<a href="#ref-2" title="&quot;To Be `On' the Internet&quot;">2</a>].
80
81 However, the recent popularity in obtaining Internet connectivity has
82 made these types of connectivity dependencies unpredictable, and
83 conventional wisdom in the Internet community dictates that the
84 various address allocation registries, such as the InterNIC, as well
85 as the ISP's, become more prudent in their address allocation
86 strategies. In that vein, the InterNIC has defined address
87 allocation policies [<a href="#ref-3" title="&quot;INTERNET REGISTRY IP ALLOCATION GUIDELINES&quot;">3</a>] wherein the majority of address allocations
88 for end-user networks are accommodated by their upstream ISP, except
89 in cases where dual- or multihoming and very large blocks of
90 addresses are required. With this allocation policy becoming
91 standard current practice, it presents unique problems regarding the
92 portability of addresses from one provider to another.
93
94 As a practical matter, end users cannot assume they "own" address
95 allocations, if their intention is to be to have full connectivity to
96 the global Internet. Rather, end users will "borrow" part of the
97 address space of an upstream provider's allocation. The larger
98 provider block from which their space is suballocated will have been
99 assigned in a manner consistent with global Internet routing.
100
101 Not having "permanent" addresses does not mean users will not have
102 unique identifiers. Such identifiers are typically Domain Name System
103 (DNS) [<a href="#ref-4" title="&quot;Domain Names - Concepts and Facilities&quot;">4</a>] names for endpoints such as servers and workstations.
104 Mechanisms such as the Dynamic Host Configuration Protocol (DHCP) [<a href="#ref-5" title="&quot;Dynamic Host Configuration Protocol&quot;">5</a>]
105
106
107
108 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 2]</span></pre>
109 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span>
110 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
111
112
113 can help automate the assignment and maintenance of host names, as
114 well as the 'borrowed' addresses required for routing-level
115 connectivity.
116
117 The PIER Working Group is developing procedures and guidelines for
118 detailed renumbering of specific technologies, such as routers [<a href="#ref-6" title="&quot;Router Renumbering Guide&quot;">6</a>].
119 PIER WG documents are intended to suggest methods both for making
120 existing networks prepared for convenient renumbering, as well as for
121 operational transition to new addressing schemes.
122
123 Also, in many instances, organizations who have never connected to
124 the Internet, yet have been using arbitrary blocks of addresses since
125 their construction, have different and unique challenges.
126
127 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Network Renumbering Defined</span>
128
129 In the simplest of definitions, the exercise of renumbering a network
130 consists of changing the IP host addresses, and perhaps the network
131 mask, of each device within the network that has an address
132 associated with it. This activity may or may not consist of all
133 networks within a particular domain, such as FOO.EDU, or networks
134 which comprise an entire autonomous system.
135
136 Devices which may need to be renumbered, for example, are networked
137 PC's, workstations, printers, file servers, terminal servers, and
138 routers. Renumbering a network may involve changing host parameters
139 and configuration files which contain IP addresses, such as
140 configuration files which contain addresses of DNS and other servers,
141 addresses contained in SNMP [<a href="#ref-7" title="&quot;A Simple Network Management Protocol (SNMP)&quot;">7</a>] management stations, and addresses
142 configured in access control lists. While this is not an all-
143 inclusive list, the PIER working group is making efforts to compile
144 documentation to identify these devices in a more detailed fashion.
145
146 Network renumbering need not be sudden activity, either; in most
147 instances, an organization's upstream service provider(s) will allow
148 a grace period where both the "old" addresses and the "new" addresses
149 may be used in parallel.
150
151 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. Reasons for Renumbering</span>
152
153 The following sections discuss particular reasons which may
154 precipitate network renumbering, and are not presented in any
155 particular order of precedence. They are grouped into reasons that
156 primarily reflect decisions made in the past, operational
157 requirements of the present, or plans for the future.
158
159
160
161
162
163
164 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 3]</span></pre>
165 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span>
166 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
167
168
169 Some of these requirements reflect evolution in the organization's
170 mission, such as a need to communicate with business partners, or to
171 work efficiently in a global Internet. Other requirements reflect
172 changes in network technologies.
173
174 <span class="h3"><a class="selflink" id="section-4.1" href="#section-4.1">4.1</a> Past</span>
175
176 Many organizations implemented IP-based networks not for connectivity
177 to the Internet, but simply to make use of effective data
178 communications mechanisms. These organizations subsequently found
179 valid reasons to connect to other organizations or the Internet in
180 general, but found the address structures they chose incompatible
181 with overall Internet practice.
182
183 Other organizations connected early to the Internet, but did so at a
184 time when address space was not scarce. Yet other organizations
185 still have no requirement to connect to the Internet, but have legacy
186 addressing structures that do not scale to adequate size.
187
188 <span class="h4"><a class="selflink" id="section-4.1.1" href="#section-4.1.1">4.1.1</a> Initial addressing using non-unique addresses</span>
189
190 As recently as two years ago, many organizations had no intention of
191 connecting to the Internet, and constructed their corporate or
192 organizational network(s) using unregistered, non-unique network
193 addresses. Obviously, as most problems evolve, these same
194 organizations determined that Internet connectivity had become a
195 valuable asset, and subsequently discovered that they could no longer
196 use the same unregistered, non-unique network addresses that were
197 previously deployed throughout their organization. Thus, the labor
198 of renumbering to valid network addresses is now upon them, as they
199 move to connect to the global Internet.
200
201 While obtaining valid, unique addresses is certainly required to
202 obtain full Internet connectivity in most circumstances, the number
203 of unique addresses required can be significantly reduced by the
204 implementation of Network Address Translation (NAT) devices [<a href="#ref-8" title="&quot;The IP Network Address Translator (NAT)&quot;">8</a>] and
205 the use of private address space, as specified in [<a href="#ref-9" title="&quot;Address Allocation for Private Internets&quot;">9</a>]. NAT reduces
206 not only the number of required unique addresses, but also localizes
207 the changes required by renumbering.
208
209 It should also be noted that NAT technology may not always be a
210 viable option, depending upon scale of addressing, performance or
211 topological constraints.
212
213
214
215
216
217
218
219
220 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 4]</span></pre>
221 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span>
222 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
223
224
225 <span class="h4"><a class="selflink" id="section-4.1.2" href="#section-4.1.2">4.1.2</a> Legacy address allocation</span>
226
227 There are also several instances where organizations were originally
228 allocated very large amounts of address space, such as traditional
229 "Class A" or "Class B" allocations, while the actual address
230 requirements are much less than the total amount of address space
231 originally allocated. In many cases, these organizations could
232 suffice with a smaller CIDR allocation, and utilize the allocated
233 address space in a more efficient manner. As allocation requirements
234 become more stringent, mechanisms to review how these organizations
235 are utilizing their address space could, quite possibly, result in a
236 request to return the original allocation to a particular registry
237 and renumber with a more appropriately sized address block.
238
239 <span class="h4"><a class="selflink" id="section-4.1.3" href="#section-4.1.3">4.1.3</a> Limitations of Bridged Internetworks</span>
240
241 Bridging has a long and distinguished history in legacy networks. As
242 networks grow, however, traditional bridged networks reach
243 performance- and stability-related limits, including (but not limited
244 to) broadcast storms.
245
246 Early routers did not have the speed to handle the needs of some
247 large networks. Some organizations were literally not able to move
248 to routers until router forwarding performance improved to be
249 comparable to bridges. Now that routers are of comparable or
250 superior speed, and offer more robust features, replacing bridged
251 networks becomes reasonable.
252
253 IP addresses assigned to pure bridged networks tend not to be
254 subnetted, yet subnetting is a basic approach for router networks.
255 Introducing subnetting is a practical necessity in moving from
256 bridging to routing.
257
258 Special cases of bridging are realized in workgroup switching
259 systems, discussed below.
260
261 <span class="h4"><a class="selflink" id="section-4.1.4" href="#section-4.1.4">4.1.4</a> Limitations of Legacy Routing Systems</span>
262
263 Other performance problems might come from routing mechanisms that
264 advertise excessive numbers of routing updates (e.g., RIP, IGRP).
265 Likewise, appropriate replacement protocols (e.g., OSPF, EIGRP, S-IS)
266 will work best with a structured addressing system that encourages
267 aggregation.
268
269
270
271
272
273
274
275
276 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 5]</span></pre>
277 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span>
278 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
279
280
281 <span class="h4"><a class="selflink" id="section-4.1.5" href="#section-4.1.5">4.1.5</a> Limitations of System Administration Methodologies</span>
282
283 There can be operational limits to growth based on the difficulty of
284 adds, moves and changes. As enterprise networks grow, it may be
285 necessary to delegate portions of address assignment and maintenance.
286 If address space has been assigned randomly or inefficiently, it may
287 be difficult to delegate portions of the address space.
288
289 It is not unusual for organizational networks to grow sporadically,
290 obtaining an address prefix here and there, in a non-contiguous
291 fashion. Depending on the number of prefixes that an organization
292 acquires over time, it may become increasingly unmanageable or demand
293 higher levels of maintenance and administration when individual
294 prefixes are acquired in this way.
295
296 Reasonable IP address management may in general simplify continuing
297 system administration; a good numbering plan is also a good
298 renumbering plan. Renumbering may force a discipline into system
299 administration that will reduce long-term support costs.
300
301 It has been observed "...there is no way to renumber a network
302 without an inventory of the hosts (absent DHCP). On a large network
303 that needs a database, plus tools and staff to maintain the
304 database."[<a href="#ref-10">10</a>] It can be argued that a detailed inventory of router
305 configurations is even more essential.
306
307 <span class="h3"><a class="selflink" id="section-4.2" href="#section-4.2">4.2</a> Present</span>
308
309 Organizations now face needs to connect to the global Internet, or at
310 a minimum to other organizations through bilateral private links.
311
312 Certain new transmission technologies have tended to redefine the
313 basic notion of an IP subnet. An IP numbering plan needs to work
314 with these new ideas. Legacy bridged networks and leading-edge
315 workgroup switched networks may very well need changes in the
316 subnetting structure. Renumbering needs may also develop due to the
317 characteristics of new WAN technologies, especially nonbroadcast
318 multi-access (NBMA) services such as Frame-Relay and Asynchronous
319 Transfer Mode (ATM).
320
321 Increased use of telecommuting by mobile workers, and in small and
322 home offices, need on-demand WAN connectivity, using modems or ISDN.
323 Effective use of demand media often requires changes in numbering and
324 routing.
325
326
327
328
329
330
331
332 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 6]</span></pre>
333 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span>
334 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
335
336
337 <span class="h4"><a class="selflink" id="section-4.2.1" href="#section-4.2.1">4.2.1</a> Change in organizational structure or network topology</span>
338
339 As companies grow, through mergers, acquisitions and reorganizations,
340 the need may arise for realignment and modification of the various
341 organizational network architectures. The connectivity of disparate
342 corporate networks present unique challenges in the realm of
343 renumbering, since one or more individual networks may have to be
344 blended into a much larger architecture consisting a different IP
345 address prefix altogether.
346
347 <span class="h4"><a class="selflink" id="section-4.2.2" href="#section-4.2.2">4.2.2</a> Inter-Enterprise Connectivity</span>
348
349 Even if they do not connect to the general Internet, enterprises may
350 interconnect to other organizations which have independent numbering
351 systems. Such connectivity can be as simple as bilateral dedicated
352 circuits. If both enterprises use unregistered or private address
353 space, they run the risk of using duplicate addresses.
354
355 In such cases, one or both organizations may need to renumber into
356 different parts of the private address space, or obtain unique
357 registered addresses.
358
359 <span class="h4"><a class="selflink" id="section-4.2.3" href="#section-4.2.3">4.2.3</a> Change of Internet Service Provider</span>
360
361 As mentioned previously in <a href="#section-2">Section 2</a>, it is increasingly becoming
362 current practice for organizations to have their IP addresses
363 allocated by their upstream ISP. Also, with the advent of Classless
364 Inter Domain Routing (CIDR) [<a href="#ref-11" title="&quot;Classless Inter-Domain Routing (CIDR): an Address Assignment and Aggregation Strategy&quot;">11</a>], and the considerable growth in the
365 size of the global Internet table, Internet Service Providers are
366 becoming more and more reluctant to allow customers to continue using
367 addresses which were allocated by the ISP, when the customer
368 terminates service and moves to another ISP. The prevailing reason
369 is that the ISP was previously issued a CIDR block of contiguous
370 address space, which can be announced to the remainder of the
371 Internet community as a single prefix. (A prefix is what is referred
372 to in classless terms as a contiguous block of IP addresses.) If a
373 non-customer advertises a specific component of the CIDR block, then
374 this adds an additional routing entry to the global Internet routing
375 table. This is what is commonly referred to as "punching holes" in a
376 CIDR block. Consequently, there are usually no routing anomalies in
377 doing this since a specific prefix is always preferred over an
378 aggregate route. However, if this practice were to happen on a large
379 scale, the growth of the global routing table would become much
380 larger, and perhaps too large for current backbone routers to
381 accommodate in an acceptable fashion with regards to performance of
382 recalculating routing information and sheer size of the routing table
383 itself. For obvious reasons, this practice is highly discouraged by
384 ISP's with CIDR blocks, and some ISP's are making this a contractual
385
386
387
388 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 7]</span></pre>
389 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span>
390 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
391
392
393 issue, so that customers understand that addresses allocated by the
394 ISP are non-portable.
395
396 It is noteworthy to mention that the likelihood of being forced to
397 renumber in this situation is inversely proportional to the size of
398 the customer's address space. For example, an organization with a
399 /16 allocation may be allowed to consider the address space
400 "portable", while an organization with multiple non-contiguous /24
401 allocations may not. While the scenarios may be vastly different in
402 scope, it becomes an issue to be decided at the discretion of the
403 initial allocating entity, and the ISP's involved; the major deciding
404 factor being whether or not the change will fragment an existing CIDR
405 block and whether it will significantly contribute to the overall
406 growth of the global Internet routing tables.
407
408 It should also be noted that (contrary to opinions sometimes voiced)
409 this form of renumbering is a technically necessary consequence of
410 changing ISP's, rather than a commercial or political mandate.
411
412 <span class="h4"><a class="selflink" id="section-4.2.3" href="#section-4.2.3">4.2.3</a> Internet Global Routing</span>
413
414 Even large organizations, now connected to the Internet with
415 "portable" address space, may find their address allocation too
416 small. Current registry guidelines require that address space usage
417 be justified by an engineering plan. Older networks may not have
418 efficiently utilized existing address space, and may need to make
419 their existing structures more efficient before new address
420 allocations can be made.
421
422 <span class="h4"><a class="selflink" id="section-4.2.4" href="#section-4.2.4">4.2.4</a> Internal Use of LAN Switching</span>
423
424 Introducing workgroup switches may introduce subtle renumbering
425 needs. Fundamentally, workgroup switches are specialized, high-
426 performance bridges, which make their main forwarding decisions based
427 on Layer 2 (MAC) address information. Even so, they rarely are
428 independent of Layer 3 (IP) address structure. Pure Layer 2
429 switching has a "flat" address space that will need to be renumbered
430 into a hierarchical, subnetted space consistent with routing.
431
432 Introducing single switches or stacks of switches may not have
433 significant impact on addressing, as long as it is understood that
434 each system of switches is a single broadcast domain. Each broadcast
435 domain should map to a single IP subnetwork.
436
437 Virtual LANs (VLANs) further extend the complexity of the role of
438 workgroup switches. It is generally true that moving an end station
439 from one switch port to another within the same VLAN will not cause
440 major changes in addressing. Many overview presentations of this
441
442
443
444 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 8]</span></pre>
445 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span>
446 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
447
448
449 technology do not make it clear that moving the same end station
450 between different VLANs will move the end station into another IP
451 subnet, requiring a significant address change.
452
453 Switches are commonly managed by SNMP applications. These network
454 management applications communicate with managed devices using IP.
455 Even if the switch does not do IP forwarding, it will itself need IP
456 addresses if it is to be managed. Also, if the clients and servers in
457 the workgroup are managed by SNMP, they will also require IP
458 addresses. The workgroup, therefore, will need to appear as one or
459 more IP subnetworks.
460
461 Increasingly, internetworking products are not purely Layer 2 or
462 Layer 3 devices. A workgroup switch product often includes a routing
463 function, so the numbering plan must support both flat Layer 2 and
464 hierarchical Layer 3 addressing.
465
466 <span class="h4"><a class="selflink" id="section-4.2.4" href="#section-4.2.4">4.2.4</a> Internal Use of NBMA Cloud Services</span>
467
468 "Cloud" services such as frame relay often are more economical than
469 traditional services. At first glance, when converting existing
470 enterprise networks to NBMA, it might appear that the existing subnet
471 structure should be preserved, but this is often not the case.
472
473 Many organizations often began by treating the "cloud" as a single
474 subnet, but experience has shown it is often better to treat the
475 individual virtual circuits as separate subnets, which appear as
476 traditional point-to-point circuits. When the individual point-to-
477 point VCs become separate subnets, efficient address utilization
478 requires the use of long prefixes (i.e., 30 bit) for these subnets.
479 In practice, obtaining 30 bit prefixes means the logical network
480 should support variable length subnet masks (VLSM). VLSMs are the
481 primary method in which an assigned prefix can be subnetted
482 efficiently for different media types. This is accomplished by
483 establishing one or more prefix lengths for LAN media with more than
484 two hosts, and subdividing one or more of these shorter prefixes into
485 longer /30 prefixes that minimize address loss.
486
487 There are alternative ways to configure routing over NBMA, using
488 special mechanisms to exploit or simulate point-to-multipoint VCs.
489 These often have a significant performance impact, and may be less
490 reliable because a single routing point of failure is created.
491 Motivations for such alternatives tend to include:
492
493
494
495
496
497
498
499
500 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 9]</span></pre>
501 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span>
502 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
503
504
505 1. A desire not to use VLSM. This is often founded in fear
506 rather than technology.
507
508 2. Router implementation issues that limit the number of subnets
509 or interfaces a given router can support.
510
511 3. An inherently point-to-multipoint application (e.g., remote
512 hosts to a data center). In such cases, some of the
513 limitations are due to the dynamic routing protocol in use.
514 In such "hub-and-spoke" implementations, static routing can
515 be preferable from a performance and flexibility standpoint,
516 since it does not produce routing protocol chatter and is
517 unaffected by split horizon constraints (namely, the inability
518 to build an adjacency with a peer within the same IP
519 subnetwork).
520
521 <span class="h4"><a class="selflink" id="section-4.2.5" href="#section-4.2.5">4.2.5</a> Expansion of Dialup Services</span>
522
523 Dialup services, especially public Internet access providers, are
524 experiencing explosive growth. This success represents a particular
525 drain on the available address space, especially with a commonly used
526 practice of assigning unique addresses to each customer.
527
528 In this case, individual users announce their address to the access
529 server using PPP's IP control protocol (IPCP) [<a href="#ref-12" title="&quot;The PPP Internet Protocol Control Protocol (IPCP)&quot;">12</a>]. The server may
530 validate the proposed address against some type of user
531 identification, or simply make the address active in a subnet to
532 which the access server (or set of bridged access servers) belongs.
533
534 The preferred technique is to allocate dynamic addresses to the user
535 from a pool of addresses available to the access server.
536
537 <span class="h4"><a class="selflink" id="section-4.2.6" href="#section-4.2.6">4.2.6</a> Returning non-contiguous prefixes for an aggregate</span>
538
539 In many instances, an organization can return their current, non-
540 contiguous prefix allocations for a contiguous block of address space
541 of equal or greater size, which can be accommodated with CIDR. Also,
542 many organizations have begun to deploy classless interior routing
543 protocols within their domains that make use of route summarization
544 and other optimized routing features, effectively reducing the total
545 number of routes being propagated within their internal network(s),
546 and making it much easier to administer and maintain.
547
548 Hierarchical routing protocols such as OSPF scale best when the
549 address assignment of a given network reflects the topology, and the
550 topology of the network can often be fluid. Given that the network is
551 fluid, even the best planned address assignment scheme, given time,
552 will diverge from the actual topology. While not required, some
553
554
555
556 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 10]</span></pre>
557 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span>
558 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
559
560
561 organization may choose to gain the benefit of both technical and
562 administrative scalability of their IGP by periodically renumbering
563 to have address assignments reflect the network topology. Patrick
564 Henry once said "the tree of liberty must from time to time be
565 watered with the blood of patriots." In the Internet, routing trees
566 of the best-planned networks need from time to time be watered with
567 at least the sweat of network administrators. Improving aggregation
568 is also highly encouraged to reduce the size of not only the global
569 Internet routing table, but also the size and scalability of interior
570 routing within the enterprise.
571
572 <span class="h3"><a class="selflink" id="section-4.3" href="#section-4.3">4.3</a> Future</span>
573
574 Emerging new protocols will most definitely affect addressing plans
575 and numbering schemes.
576
577 <span class="h4"><a class="selflink" id="section-4.3.1" href="#section-4.3.1">4.3.1</a> Internal Use of Switched Virtual Circuit Services</span>
578
579 Services such as ATM virtual circuits, switched frame relay, etc.,
580 present challenges not considered in the original IP design. The
581 basic IP decision in forwarding a packet is whether the destination
582 is local or remote, in relation to the source host's subnet. Address
583 resolution mechanisms are used to find the medium address of the
584 destination in the case of local destinations, or to find the medium
585 address of the router in the case of remote routers.
586
587 In these new services, there are cases where it is far more effective
588 to "cut-through" a new virtual circuit to the destination. If the
589 destination is on a different subnet than the source, the cut-through
590 typically is to the egress router that serves the destination subnet.
591 The advantage of cut-through in such a case is that it avoids the
592 latency of multiple router hops, and reduces load on "backbone"
593 routers. The cut-through decision is usually made by an entry router
594 that is aware of both the routed and switched environments.
595
596 This entry router communicates with a address resolution server using
597 the Next Hop Resolution Protocol (NHRP) [<a href="#ref-13" title="&quot;NBMA Next Hop Resolution Protocol (NHRP)&quot;">13</a>]. This server maps the
598 destination network address to either a next-hop router (where cut-
599 through is not appropriate) or to an egress router reached over the
600 switched service. Obviously, the data base in such a server may be
601 affected by renumbering. Clients may have a hard-coded address of the
602 server, which again may need to change. While the NHRP protocol
603 specifications are still evolving at the time of this writing,
604 commercial implementations based on drafts of the protocol standard
605 are in use.
606
607
608
609
610
611
612 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 11]</span></pre>
613 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span>
614 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
615
616
617 <span class="h4"><a class="selflink" id="section-4.3.2" href="#section-4.3.2">4.3.2</a> Transitioning to IP version 6</span>
618
619 Of course, when IPv6 [<a href="#ref-14" title="&quot;Internet Protocol, Version 6 (IPv6) Specification&quot;">14</a>] deployment is set in motion, and as
620 methodologies are developed to transition to IPv6, renumbering will
621 also be necessary, but perhaps not immediately mandatory. To aid in
622 the transition to IPv6, mechanisms to deploy dual- IPv4/IPv6 stacks
623 on network hosts should also become available. It is also envisioned
624 that Network Address Translation (NAT) devices will be developed to
625 assist in the IPv4 to IPv6 transition, or perhaps supplant the need
626 to renumber the majority of interior networks altogether, but that is
627 beyond the scope of this document. At the very least, DNS hosts will
628 need to be reconfigured to resolve new host names and addresses, and
629 routers will need to be reconfigured to advertise new prefixes.
630
631 IPv6 address allocation will be managed by the Internet Assigned
632 Numbers Authority (IANA) as set forth in [<a href="#ref-15" title="&quot;IPv6 Address Allocation Management&quot;">15</a>].
633
634 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Summary</span>
635
636 As indicated by the Internet Architecture Board (IAB) in [<a href="#ref-16" title="&quot;Renumbering Needs Work&quot;">16</a>], the
637 task of renumbering networks is becoming more widespread and
638 commonplace. Although there are numerous reasons why an organization
639 would desire, or be required to renumber, there are equally as many
640 reasons why address allocation should be done with great care and
641 forethought at the onset, in order to minimize the impact that
642 renumbering would have on the organization. Even with the most
643 forethought and vision, however, an organization cannot foresee the
644 possibility for renumbering. The best advice, in this case, is to be
645 prepared, and get ready for renumbering.
646
647 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Security Considerations</span>
648
649 Although no obvious security issues are discussed in this document,
650 it stands to reason that renumbering certain devices can defeat
651 security systems designed and based on static IP host addresses.
652 Care should be exercised by the renumbering entity to ensure that all
653 security systems deployed with the network(s) which may need to be
654 renumbered be given special consideration and significant forethought
655 to provide continued functionality and adequate security.
656
657 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. Acknowledgments</span>
658
659 Special acknowledgments to Yakov Rekhter [cisco Systems, Inc.], Tony
660 Bates [cisco Systems, Inc.] and Brian Carpenter [CERN] for their
661 contributions and editorial critique.
662
663
664
665
666
667
668 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 12]</span></pre>
669 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span>
670 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
671
672
673 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. References</span>
674
675 [<a id="ref-1">1</a>] Gerich, E., "Unique Addresses are Good", <a href="./rfc1814">RFC 1814</a>, IAB, July 1995.
676
677 [<a id="ref-2">2</a>] Crocker, D., "To Be `On' the Internet", <a href="./rfc1775">RFC 1775</a>, March 1995.
678
679 [<a id="ref-3">3</a>] Hubbard, K., Kosters, M., Conrad, D., Karrenberg, D., and J.
680 Postel, "INTERNET REGISTRY IP ALLOCATION GUIDELINES",
681 <a href="https://www.rfc-editor.org/bcp/bcp12">BCP 12</a>, <a href="./rfc2050">RFC 2050</a>, November 1996.
682
683 [<a id="ref-4">4</a>] Mockapetris, P., "Domain Names - Concepts and Facilities",
684 and "Domain Names - Implementation and Specification",
685 STD 13, RFCs 1034, 1035, November 1987.
686
687 [<a id="ref-5">5</a>] Droms, R., "Dynamic Host Configuration Protocol", <a href="./rfc1541">RFC 1541</a>,
688 October 1993.
689
690 [<a id="ref-6">6</a>] Berkowitz, H., "Router Renumbering Guide", <a href="./rfc2072">RFC 2072</a>,
691 January 1997.
692
693 [<a id="ref-7">7</a>] Case, J., Fedor, M., Schoffstall, M., and J. Davin, "A Simple
694 Network Management Protocol (SNMP)", STD 15, <a href="./rfc1157">RFC 1157</a>,
695 May 1990.
696
697 [<a id="ref-8">8</a>] Egevang,, K., and P. Francis, "The IP Network Address Translator
698 (NAT)", <a href="./rfc1631">RFC 1631</a>, May 1994.
699
700 [<a id="ref-9">9</a>] Rekhter, Y., Moskowitz, R., Karrenberg, D., de Groot, G-J., and E.
701 Lear, "Address Allocation for Private Internets", <a href="./rfc1918">RFC 1918</a>,
702 February 1996.
703
704 [<a id="ref-10">10</a>] Messages to PIER list on CERN renumbering; Brian Carpenter, CERN.
705 Available in PIER WG mailing list archives.
706
707 [<a id="ref-11">11</a>] Fuller, V., Li, T., Yu, J., and K. Varadhan, "Classless
708 Inter-Domain Routing (CIDR): an Address Assignment and
709 Aggregation Strategy", <a href="./rfc1519">RFC 1519</a>, October 1993.
710
711 [<a id="ref-12">12</a>] McGregor, G., "The PPP Internet Protocol Control Protocol
712 (IPCP)", <a href="./rfc1332">RFC 1332</a>, May 1992.
713
714 [<a id="ref-13">13</a>] Luciani, J., Katz, D., Piscitello, D., and Cole, B., "NBMA Next
715 Hop Resolution Protocol (NHRP)", Work in Progress.
716
717 [<a id="ref-14">14</a>] Deering, S., and R. Hinden, "Internet Protocol, Version 6 (IPv6)
718 Specification", <a href="./rfc1883">RFC 1883</a>, December 1995.
719
720
721
722
723
724 <span class="grey">Ferguson &amp; Berkowitz Informational [Page 13]</span></pre>
725 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span>
726 <span class="grey"><a href="./rfc2071">RFC 2071</a> Network Renumbering Overview January 1997</span>
727
728
729 [<a id="ref-15">15</a>] IAB and IESG, "IPv6 Address Allocation Management", <a href="./rfc1881">RFC 1881</a>,
730 December 1995.
731
732 [<a id="ref-16">16</a>] Carpenter, B., and Y. Rekhter, "Renumbering Needs Work", <a href="./rfc1900">RFC 1900</a>,
733 February 1996.
734
735 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a>. Authors' Addresses</span>
736
737 Paul Ferguson
738 cisco Systems, Inc.
739 1875 Campus Commons Road
740 Suite 210
741 Reston, VA 22091
742
743 Phone: (703) 716-9538
744 Fax: (703) 716-9599
745 EMail: [email protected]
746
747
748 Howard C. Berkowitz
749 PSC International
750 1600 Spring Hill Road
751 Vienna, VA 22182
752
753 Phone (703) 998-5819
754 Fax: (703) 998-5058
755 EMail: [email protected]
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780 Ferguson &amp; Berkowitz Informational [Page 14]
781 </pre>