idnits 2.17.1 draft-ietf-ipo-impairments-04.txt: -(127): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(158): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(191): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(204): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(207): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(215): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(907): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(918): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1029): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1076): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1077): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1096): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1122): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing document type: Expected "INTERNET-DRAFT" in the upper left hand corner of the first page ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document is more than 15 pages and seems to lack a Table of Contents. == There are 20 instances of lines with non-ascii characters in the document. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack an Authors' Addresses Section. ** There are 498 instances of too long lines in the document, the longest one being 5 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (December 2002) is 7803 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'GMPLS' is mentioned on line 44, but not defined == Missing Reference: 'Gerstel2000' is mentioned on line 68, but not defined == Missing Reference: 'Strand2001' is mentioned on line 68, but not defined == Unused Reference: 'Bra96' is defined on line 1202, but no explicit reference was found in the text == Unused Reference: 'CBD00' is defined on line 1205, but no explicit reference was found in the text == Unused Reference: 'Doverspike00' is defined on line 1209, but no explicit reference was found in the text == Unused Reference: 'Gerstel 2000' is defined on line 1213, but no explicit reference was found in the text == Unused Reference: 'KRB01a' is defined on line 1217, but no explicit reference was found in the text == Unused Reference: 'KRB01b' is defined on line 1221, but no explicit reference was found in the text == Unused Reference: 'Moy98' is defined on line 1225, but no explicit reference was found in the text -- Possible downref: Normative reference to a draft: ref. 'Chaudhuri00' -- Possible downref: Non-RFC (?) normative reference: ref. 'Goldstein94' -- Possible downref: Non-RFC (?) normative reference: ref. 'ITU' -- Possible downref: Non-RFC (?) normative reference: ref. 'Kaminow97' == Outdated reference: A later version (-07) exists of draft-ietf-ccamp-gmpls-architecture-03 == Outdated reference: A later version (-05) exists of draft-ietf-ipo-framework-02 ** Downref: Normative reference to an Informational draft: draft-ietf-ipo-framework (ref. 'Rajagopalam02') == Outdated reference: A later version (-02) exists of draft-chiu-strand-unique-olcp-01 -- Possible downref: Normative reference to a draft: ref. 'Strand01' -- Possible downref: Non-RFC (?) normative reference: ref. 'Strand01b' -- Possible downref: Non-RFC (?) normative reference: ref. 'Yates99' == Outdated reference: A later version (-03) exists of draft-awduche-mpls-te-optical-01 -- No information found for draft-ietf-gmpls- - is the name correct? -- No information found for draft-ospf-generalized- - is the name correct? Summary: 9 errors (**), 0 flaws (~~), 16 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Draft John Strand (Editor) 3 Document: draft-ietf-ipo-impairments-04.txt Angela Chiu (Editor) 4 Informational Track AT&T 5 Expiration Date: May 2003 7 December 2002 9 Impairments And Other Constraints On Optical Layer Routing 11 Status of this Memo 13 This document is an Internet-Draft and is in full conformance with 14 all provisions of Section 10 of RFC2026. Internet-Drafts are 15 working documents of the Internet Engineering Task Force (IETF), its 16 areas, and its working groups. Note that other groups may also 17 distribute working documents as Internet-Drafts. 19 Internet-Drafts are draft documents valid for a maximum of six 20 months and may be updated, replaced, or made obsolete by other 21 documents at any time. It is inappropriate to use Internet-Drafts as 22 reference material or to cite them other than as "work in progress." 24 The list of current Internet-Drafts can be accessed at 25 http://www.ietf.org/ietf/1id-abstracts.txt. 26 The list of Internet-Draft Shadow Directories can be accessed at 27 http://www.ietf.org/shadow.html. 29 Abstract 30 Optical networking poses a number challenges for GMPLS. Optical 31 technology is fundamentally an analog rather than digital technology; 32 and the optical layer is lowest in the transport hierarchy and hence 33 has an intimate relationship with the physical geography of the 34 network. This contribution surveys some of the aspects of optical 35 networks which impact routing and identifies possible GMPLS responses 36 for each: (1) Constraints arising from the design of new software 37 controllable network elements, (2) Constraints in a single all-optical 38 domain without wavelength conversion, (3) Complications arising in more 39 complex networks incorporating both all-optical and opaque 40 architectures, and (4) Impacts of diversity constraints. 42 1. Introduction 44 GMPLS [GMPLS] aims to extend MPLS to encompass a number of transport 45 architectures. Included are optical networks incorporating a number 46 of all-optical and opto-electronic elements such as optical cross- 47 connects with both optical and electrical fabrics, transponders, and 48 optical add-drop multiplexers. Optical networking poses a number 49 Impairments And Other Constraints December 2002 50 On Optical Layer Routing 52 challenges for GMPLS. Optical technology is fundamentally an analog 53 rather than digital technology; and the optical layer is lowest in 54 the transport hierarchy and hence has an intimate relationship with 55 the physical geography of the network. 57 GMPLS already has incorporated extensions to deal with some of the 58 unique aspects of the optical layer. This contribution surveys some 59 of the aspects of optical networks which impact routing and 60 identifies possible GMPLS responses for each. Routing constraints 61 and/or complications arising from the design of network elements, 62 the accumulation of signal impairments, and from the need to 63 guarantee the physical diversity of some circuits are discussed. 65 Since the purpose of this draft is to further the specification of 66 GMPLS, alternative approaches to controlling an optical network are 67 not discussed. For discussions of some broader issues, see 68 [Gerstel2000] and [Strand2001]. 70 The organization of the contribution is as follows: 72 - Section 2 is a section requested by the sub-IP Area management 73 for all new drafts. It explains how this document fits into the 74 Area and into the IPO WG, and why it is appropriate for these 75 groups. 76 - Section 3 describes constraints arising from the design of new 77 software controllable network elements. 78 - Section 4 addresses the constraints in a single all-optical 79 domain without wavelength conversion. 80 - Section 5 extends the discussion to more complex networks. 81 incorporating both all-optical and opaque architectures. 82 - Section 6 discusses the impacts of diversity constraints. 83 - Section 7 deals with security requirements. 84 - Section 8 contains acknowledgments. 85 - Section 9 contains references. 86 - Section 10 contains contributing authors� addresses. 87 - Section 11 contains editors� addresses. 89 2. Sub-IP Area Summary And Justification Of Work 90 This draft merges and extends two previous drafts, draft-chiu- 91 strand-unique-olcp-02.txt and draft-banerjee-routing-impairments- 92 00.txt. These two drafts were made IPO working group documents to 93 form a basis for a design team at the Minneapolis meeting, where it 94 was also requested that they be merged to create a requirements 95 document for the WG. 97 In the larger sub-IP Area structure, this merged document describes 98 specific characteristics of optical technology and the requirements 99 Impairments And Other Constraints December 2002 100 On Optical Layer Routing 102 they place on routing and path selection. It is appropriate for the 103 IPO working group because the material is specific to optical 104 networks. It identifies and documents the characteristics of the 105 optical transport network that are important for selecting paths for 106 optical channels, which is a work area for the IPO WG. It is 107 appropriate work for this WG because the material covered is 108 directly aimed at establishing a framework and requirements for 109 routing in an optical network. 111 Related documents are: 112 draft-banerjee-routing-impairments-00.txt 113 draft-parent-obgp-01.txt 114 draft-bernstein-optical-bgp-00.txt 115 draft-hayata-ipo-carrier-needs-00.txt 116 draft-many-carrier-framework-uni-01.txt 117 draft-papadimitriou-ipo-non-linear-routing-impairm-01.txt 119 3. Reconfigurable Network Elements 121 3.1 Technology Background 123 Control plane architectural discussions (e.g., [Awduche99]) usually 124 assume that the only software reconfigurable network element is an 125 optical layer cross-connect (OLXC). There are however other software 126 reconfigurable elements on the horizon, specifically tunable lasers and 127 receivers and reconfigurable optical add-drop multiplexers (OADM�s). 128 These elements are illustrated in the following simple example, which 129 is modeled on announced Optical Transport System (OTS) products: 130 + + 131 ---+---+ |\ /| +---+--- 132 ---| A |----|D| X Y |D|----| A |--- 133 ---+---+ |W| +--------+ +--------+ |W| +---+--- 134 : |D|-----| OADM |-----| OADM |-----|D| : 135 ---+---+ |M| +--------+ +--------+ |M| +---+--- 136 ---| A |----| | | | | | | |----| A |--- 137 ---+---+ |/ | | | | \| +---+--- 138 + +---+ +---+ +---+ +---+ + 139 D | A | | A | | A | | A | E 140 +---+ +---+ +---+ +---+ 141 | | | | | | | | 143 Figure 3-1: An OTS With OADM's - Functional Architecture 145 In Fig.3-1, the part that is on the inner side of all boxes labeled 146 "A" defines an all-optical subnetwork. From a routing perspective 147 two aspects are critical: 149 Impairments And Other Constraints December 2002 150 On Optical Layer Routing 152 - Adaptation: These are the functions done at the edges of the 153 subnetwork that transform the incoming optical channel into the 154 physical wavelength to be transported through the subnetwork. 155 - Connectivity: This defines which pairs of edge Adaptation 156 functions can be interconnected through the subnetwork. 158 In Fig. 3-1, D and E are DWDM�s and X and Y are OADM�s. The boxes 159 labeled "A" are adaptation functions. They map one or more input 160 optical channels assumed to be standard short reach signals into a 161 long reach (LR) wavelength or wavelength group which will pass 162 transparently to a distant adaptation function. Adaptation 163 functionality which affects routing includes: 164 - Multiplexing: Either electrical or optical TDM may be used to 165 combine the input channels into a single wavelength. This is 166 done to increase effective capacity: A typical DWDM might be 167 able to handle 100 2.5 Gb/sec signals (250 Gb/sec total) or 50 168 10 Gb/sec (500 Gb/sec total); combining the 2.5 Gb/sec signals 169 together thus effectively doubles capacity. After multiplexing 170 the combined signal must be routed as a group to the distant 171 adaptation function. 172 - Adaptation Grouping: In this technique, groups of k (e.g., 4) 173 wavelengths are managed as a group within the system and must be 174 added/dropped as a group. We will call such a group an 175 "adaptation grouping". Examples include so called "wave group" 176 and "waveband" [Passmore01]. Groupings on the same system may 177 differ in basics such as wavelength spacing, which constrain the 178 type of channels that can be accommodated. 179 - Laser Tunability: The lasers producing the LR wavelengths may 180 have a fixed frequency, may be tunable over a limited range, or 181 be tunable over the entire range of wavelengths supported by the 182 DWDM. Tunability speeds may also vary. 184 Connectivity between adaptation functions may also be limited: 185 - As pointed out above, TDM multiplexing and/or adaptation 186 grouping by the adaptation function forces groups of input 187 channels to be delivered together to the same distant adaptation 188 function. 189 - Only adaptation functions whose lasers/receivers are tunable to 190 compatible frequencies can be connected. 191 - The switching capability of the OADM�s may also be constrained. 192 For example: 193 o There may be some wavelengths that can not be dropped at 194 all. 195 o There may be a fixed relationship between the frequency 196 dropped and the physical port on the OADM to which it is 197 dropped. 198 o OADM physical design may put an upper bound on the number 199 of adaptation groupings dropped at any single OADM. 201 Impairments And Other Constraints December 2002 202 On Optical Layer Routing 204 For a fixed configuration of the OADM�s and adaptation functions 205 connectivity will be fixed: Each input port will essentially be 206 hard-wired to some specific distant port. However this connectivity 207 can be changed by changing the configurations of the OADM�s and 208 adaptation functions. For example, an additional adaptation grouping 209 might be dropped at an OADM or a tunable laser retuned. In each case 210 the port-to-port connectivity is changed. 212 These capabilities can be expected to be under software control. 213 Today the control would rest in the vendor-supplied Element 214 Management system (EMS), which in turn would be controlled by the 215 operator�s OS�s. However in principle the EMS could participate in 216 the GMPLS routing process. 218 3.2 Implications For Routing 220 An OTS of the sort discussed in Sec. 3.1 is essentially a 221 geographically distributed but blocking cross-connect system. The 222 specific port connectivity is dependent on the vendor design and 223 also on exactly what line cards have been deployed. 225 One way for GMPLS to deal with this architecture would be to view 226 the port connectivity as externally determined. In this case the 227 links known to GMPLS would be groups of identically routed 228 wavebands. If these were reconfigured by the external EMS the 229 resulting connectivity changes would need to be detected and 230 advertised within GMPLS. If the topology shown in Fig. 3-1 became a 231 tree or a mesh instead of the linear topology shown, the 232 connectivity changes could result in SRLG changes. 234 Alternatively, GMPLS could attempt to directly control this port 235 connectivity. The state information needed to do this is likely to 236 be voluminous and vendor specific. 238 4. Wavelength Routed All-Optical Networks 240 The optical networks presently being deployed may be called "opaque" 241 ([Tkach98]): each link is optically isolated by transponders doing 242 O/E/O conversions. They provide regeneration with retiming and 243 reshaping, also called 3R, which eliminates transparency to bit 244 rates and frame format. These transponders are quite expensive and 245 their lack of transparency also constrains the rapid introduction of 246 new services. Thus there are strong motivators to introduce 247 "domains of transparency" - all-optical subnetworks - larger than an 248 OTS. 250 The routing of lightpaths through an all-optical network has 251 received extensive attention. (See [Yates99] or [Ramaswami98]). 253 Impairments And Other Constraints December 2002 254 On Optical Layer Routing 256 When discussing routing in an all-optical network it is usually 257 assumed that all routes have adequate signal quality. This may be 258 ensured by limiting all-optical networks to subnetworks of limited 259 geographic size which are optically isolated from other parts of the 260 optical layer by transponders. This approach is very practical and 261 has been applied to date, e.g. when determining the maximum length 262 of an Optical Transport System (OTS). Furthermore operational 263 considerations like fault isolation also make limiting the size of 264 domains of transparency attractive. 266 There are however reasons to consider contained domains of 267 transparency in which not all routes have adequate signal quality. 268 From a demand perspective, maximum bit rates have rapidly increased 269 from DS3 to OC-192 and soon OC-768 (40 Gb/sec). As bit rates 270 increase it is necessary to increase power. This makes impairments 271 and nonlinearities more troublesome. From a supply perspective, 272 optical technology is advancing very rapidly, making ever-larger 273 domains possible. In this section we assume that these 274 considerations will lead to the deployment of a domain of 275 transparency that is too large to ensure that all potential routes 276 have adequate signal quality for all circuits. Our goal is to 277 understand the impacts of the various types of impairments in this 278 environment. 280 Note that as we describe later in the section there are many types 281 of physical impairments. Which of these needs to be dealt with 282 explicitly when performing on-line distributed routing will vary 283 considerably and will depend on many variables, including: 284 - Equipment vendor design choices, 285 - Fiber characteristics, 286 - Service characteristics (e.g., circuit speeds), 287 - Network size, 288 - Network operator engineering and deployment strategies. 289 For example, a metropolitan network which does not intend to support 290 bit rates above 2.5 Gb/sec may not be constrained by any of these 291 impairments, while a continental or international network which 292 wished to minimize O/E/O regeneration investment and support 40 293 Gb/sec connections might have to explicitly consider many of them. 294 Also, a network operator may reduce or even eliminate their 295 constraint set by building a relatively small domain of transparency 296 to ensure that all the paths are feasible, or by using some 297 proprietary tools based on rules from the OTS vendor to pre-qualify 298 paths between node pairs and put them in a table that can be 299 accessed each time a routing decision has to be made through that 300 domain. 302 4.1 Problem Formulation 303 Impairments And Other Constraints December 2002 304 On Optical Layer Routing 306 We consider a single domain of transparency without wavelength 307 translation. Additionally due to the proprietary nature of DWDM 308 transmission technology, we assume that the domain is either single 309 vendor or architected using a single coherent design, particularly 310 with regard to the management of impairments. 312 We wish to route a unidirectional circuit from ingress client node X 313 to egress client node Y. At both X and Y, the circuit goes through 314 an O/E/O conversion which optically isolates the portion within our 315 domain. We assume that we know the bit rate of the circuit. Also, 316 we assume that the adaptation function at X may apply some Forward 317 Error Correction (FEC) method to the circuit. We also assume we know 318 the launch power of the laser at X. 320 Impairments can be classified into two categories, linear and 321 nonlinear. (See [Tkach98] for more on impairment constraints). 322 Linear effects are independent of signal power and affect 323 wavelengths individually. Amplifier spontaneous emission (ASE), 324 polarization mode dispersion (PMD), and chromatic dispersion are 325 examples. Nonlinearities are significantly more complex: they 326 generate not only impairments on each channel, but also crosstalk 327 between channels. 329 In the remainder of this section we first outline how two key linear 330 impairments (PMD and ASE) might be handled by a set of analytical 331 formulae as additional constraints on routing. We next discuss how 332 the remaining constraints might be approached. Finally we take a 333 broader perspective and discuss the implications of such constraints 334 on control plane architecture and also on broader constrained domain 335 of transparency architecture issues. 337 4.2 Polarization Mode Dispersion (PMD) 339 For a transparent fiber segment, the general PMD requirement is that 340 the time-average differential group delay (DGD) between two 341 orthogonal state of polarizations should be less than fraction a of 342 the bit duration, T=1/B, where B is the bit rate. The value of the 343 parameter a depends on three major factors: 1) margin allocated to 344 PMD, e.g. 1dB; 2) targeted outage probability, e.g. 4x10-5, and 3) 345 sensitivity of the receiver to DGD. A typical value for a is 10% 346 [ITU]. More aggressive designs to compensate for PMD may allow 347 values higher than 10%. (This would be a system parameter dependent 348 on the system design. It would need to be known to the routing 349 process.) 351 The PMD parameter (Dpmd) is measured in pico-seconds (ps) per 352 sqrt(km). The square of the PMD in a fiber span, denoted as span- 353 PMD-square is then given by the product of Dpmd**2 and the span 354 length. (A fiber span in a transparent network refers to a segment 355 between two optical amplifiers.) If Dpmd is constant, this results 356 Impairments And Other Constraints December 2002 357 On Optical Layer Routing 359 in a upper bound on the maximum length of an M-fiber-span 360 transparent segment, which is inversely proportional to the square 361 of the product of bit rate and Dpmd (the detailed equation is 362 omitted due to the format constraint - see [Strand01] for details). 364 For older fibers with a typical PMD parameter of 0.5 picoseconds per 365 square root of km, based on the constraint, the maximum length of 366 the transparent segment should not exceed 400km and 25km for bit 367 rates of 10Gb/s and 40Gb/s, respectively. Due to recent advances in 368 fiber technology, the PMD-limited distance has increased 369 dramatically. For newer fibers with a PMD parameter of 0.1 370 picosecond per square root of km, the maximum length of the 371 transparent segment (without PMD compensation) is limited to 10000km 372 and 625km for bit rates of 10Gb/s and 40Gb/, respectively. Still 373 lower values of PMD are attainable in commercially available fiber 374 today, and the PMD limit can be further extended if a larger value 375 of the parameter a (ratio of DGD to the bit period) can be 376 tolerated. In general, the PMD requirement is not an issue for most 377 types of fibers at 10Gb/s or lower bit rate. But it will become an 378 issue at bit rates of 40Gb/s and higher. 380 If the PMD parameter varies between spans, a slightly more 381 complicated equation results (see [Strand01]), but in any event the 382 only link dependent information needed by the routing algorithm is 383 the square of the link PMD, denoted as link-PMD-square. It is the 384 sum of the span-PMD-square of all spans on the link. 386 Note that when one has some viable PMD compensation devices and 387 deploy them ubiquitously on all routes with potential PMD issues in 388 the network, then the PMD constraint disappears from the routing 389 perspective. 391 4.3 Amplifier Spontaneous Emission 393 ASE degrades the optical signal to noise ratio (OSNR). An acceptable 394 optical SNR level (SNRmin) which depends on the bit rate, 395 transmitter-receiver technology (e.g., FEC), and margins allocated 396 for the impairments, needs to be maintained at the receiver. In 397 order to satisfy this requirement, vendors often provide some 398 general engineering rule in terms of maximum length of the 399 transparent segment and number of spans. For example, current 400 transmission systems are often limited to up to 6 spans each 80km 401 long. For larger transparent domains, more detailed OSNR 402 computations will be needed to determine whether the OSNR level 403 through a domain of transparency is acceptable. This would provide 404 flexibility in provisioning or restoring a lightpath through a 405 transparent subnetwork. 407 Assume that the average optical power launched at the transmitter is 408 P. The lightpath from the transmitter to the receiver goes through M 409 Impairments And Other Constraints December 2002 410 On Optical Layer Routing 412 optical amplifiers, with each introducing some noise power. Unity 413 gain can be used at all amplifier sites to maintain constant signal 414 power at the input of each span to minimize noise power and 415 nonlinearity. A constraint on the maximum number of spans can be 416 obtained [Kaminow97] which is proportional to P and inversely 417 proportional to SNRmin, optical bandwidth B, amplifier gain G-1 and 418 spontaneous emission factor n of the optical amplifier, assuming all 419 spans have identical gain and noise figure. (Again, the detailed 420 equation is omitted due to the format constraint - see [Strand01] 421 for details.) Let�s take a typical example. Assuming P=4dBm, 422 SNRmin=20dB with FEC, B=12.5GHz, n=2.5, G=25dB, based on the 423 constraint, the maximum number of spans is at most 10. However, if 424 FEC is not used and the requirement on SNRmin becomes 25dB, the 425 maximum number of spans drops down to 3. 427 For ASE the only link-dependent information needed by the routing 428 algorithm is the noise of the link, denoted as link-noise, which is 429 the sum of the noise of all spans on the link. Hence the constraint 430 on ASE becomes that the aggregate noise of the transparent segment 431 which is the sum of the link-noise of all links can not exceed 432 P/SNRmin. 434 4.4 Approximating The Effects Of Some Other Impairment Constraints 436 There are a number of other impairment constraints that we believe 437 could be approximated with a domain-wide margin on the OSNR, plus in 438 some cases a constraint on the total number of networking elements 439 (OXC or OADM) along the path. Most impairments generated at OXCs or 440 OADMs, including polarization dependent loss, coherent crosstalk, 441 and effective passband width, could be dealt with using this 442 approach. In principle, impairments generated at the nodes can be 443 bounded by system engineering rules because the node elements can be 444 designed and specified in a uniform manner. This approach is not 445 feasible with PMD and noise because neither can be uniformly 446 specified. Instead, they depend on node spacing and the 447 characteristics of the installed fiber plant, neither of which are 448 likely to be under the system designer�s control. 450 Examples of the constraints we propose to approximate with a domain- 451 wide margin are given in the remaining paragraphs in this section. 452 It should be kept in mind that as optical transport technology 453 evolves it may become necessary to include some of these impairments 454 explicitly in the routing process. Other impairments not mentioned 455 here at all may also become sufficiently important to require 456 incorporation either explicitly or via a domain-wide margin. 458 Other Polarization Dependent Impairments Other polarization- 459 dependent effects besides PMD influence system performance. For 460 example, many components have polarization-dependent loss (PDL) 461 [Ramaswami98], which accumulates in a system with many components on 462 Impairments And Other Constraints December 2002 463 On Optical Layer Routing 465 the transmission path. The state of polarization fluctuates with 466 time and its distribution is very important also. It is generally 467 required to maintain the total PDL on the path to be within some 468 acceptable limit, potentially by using some compensation technology 469 for relatively long transmission systems, plus a small built-in 470 margin in OSNR. Since the total PDL increases with the number of 471 components in the data path, it must be taken into account by the 472 system vendor when determining the maximum allowable number of 473 spans. 475 Chromatic Dispersion In general this impairment can be adequately 476 (but not optimally) compensated for on a per-link basis, and/or at 477 system initial setup time. Today most deployed compensation devices 478 are based on DCF (Dispersion Compensation Fiber). DCF provides per 479 fiber compensation by means of a spool of fiber with a CD coefficient 480 opposite to the fiber. Due to the imperfect matching between the CD 481 slope of the fiber and the DCF some lambdas can be over compensated 482 while others can be under compensated. Moreover DCF modules may only 483 be available in fixed lengths of compensating fiber; this means that 484 sometimes it is impossible to find a DCF module that exactly 485 compensates the CD introduced by the fiber. These effects introduce 486 what is known as residual CD. Residual CD varies with the frequency 487 of the wavelength. Knowing the characteristics of both of the fiber 488 and the DCF modules along the path, this can be calculated with a 489 sufficient degree of precision. However this is a very challenging 490 task. In fact the per-wavelength residual dispersion needs to be 491 combined with other information in the system (e.g. types fibers to 492 figure out the amount of nonlinearities) to obtain the net effect of 493 CD either by simulation or by some analytical approximation. It 494 appears that the routing/control plane should not be burdened by such 495 a large set of information while it can be handled at the system 496 design level. Therefore it will be assumed until proven otherwise 497 that residual dispersion should not be reported. For high bit rates, 498 dynamic dispersion compensation may be required at the receiver to 499 clean up any residual dispersion. 501 Crosstalk Optical crosstalk refers to the effect of other signals on 502 the desired signal. It includes both coherent (i.e. intrachannel) 503 crosstalk and incoherent (i.e. interchannel) crosstalk. Main 504 contributors of crosstalk are the OADM and OXC sites that use a DWDM 505 multiplexer/demultiplexer (MUX/DEMUX) pair. For a relatively sparse 506 network where the number of OADM/OXC nodes on a path is low, 507 crosstalk can be treated with a low margin in OSNR without being a 508 binding constraint. But for some relatively dense networks where 509 crosstalk might become a binding constraint, one needs to propagate 510 the per-link crosstalk information to make sure that the end-to-end 511 Impairments And Other Constraints December 2002 512 On Optical Layer Routing 514 path crosstalk which is the sum of the crosstalks on all the 515 corresponding links to be within some limit, e.g. -25dB threshold 516 with 1dB penalty ([Goldstein94]). Another way to treat it without 517 having to propagate per-link crosstalk information is to have the 518 system evaluate what the maximum number of OADM/OXC nodes that has a 519 MUX/DEMUX pair for the worst route in the transparent domain for a 520 low built-in margin. The latter one should work well where all the 521 OXC/OADM nodes have similar level of crosstalk. 523 Effective Passband As more and more DWDM components are cascaded, 524 the effective passband narrows. The number of filters along the 525 link, their passband width and their shape will determine the end- 526 to-end effective passband. In general, this is a system design 527 issue, i.e., the system is designed with certain maximum bit rate 528 using the proper modulation format and filter spacing. For linear 529 systems, the filter effect can be turned into a constraint on the 530 maximum number of narrow filters with the condition that filters in 531 the systems are at least as wide as the one in the receiver. 532 Because traffic at lower bit rates can tolerate a narrower passband, 533 the maximum allowable number of narrow filters will increase as the 534 bit rate decreases. 536 Nonlinear Impairments It seems unlikely that these can be dealt with 537 explicitly in a routing algorithm because they lead to constraints 538 that can couple routes together and lead to complex dependencies, 539 e.g. on the order in which specific fiber types are traversed 540 [Kaminow97]. Note that different fiber types (standard single mode 541 fiber, dispersion shifted fiber, dispersion compensated fiber, etc.) 542 have very different effects from nonlinear impairments. A full 543 treatment of the nonlinear constraints would likely require very 544 detailed knowledge of the physical infrastructure, including 545 measured dispersion values for each span, fiber core area and 546 composition, as well as knowledge of subsystem details such as 547 dispersion compensation technology. This information would need to 548 be combined with knowledge of the current loading of optical signals 549 on the links of interest to determine the level of nonlinear 550 impairment. Alternatively, one could assume that nonlinear 551 impairments are bounded and result in X dB margin in the required 552 OSNR level for a given bit rate, where X for performance reasons 553 would be limited to 1 or 2 dB, consequently setting a limit on the 554 maximum number of spans. For the approach described here to be 555 useful, it is desirable for this span length limit to be longer than 556 that imposed by the constraints which can be treated explicitly. 557 When designing a DWDM transport system, there are tradeoffs between 558 signal power launched at the transmitter, span length, and nonlinear 559 effects on BER that need to be considered jointly. Here, we assume 560 that an X dB margin is obtained after the transport system has been 561 designed with a fixed signal power and maximum span length for a 562 given bit rate. Note that OTSs can be designed in very different 563 ways, in linear, pseudo-linear, or nonlinear environments. The X-dB 564 Impairments And Other Constraints December 2002 565 On Optical Layer Routing 567 margin approach may be valid for some but not for others. However, 568 it is likely that there is an advantage in designing systems that 569 are less aggressive with respect to nonlinearities, and therefore 570 somewhat sub-optimal, in exchange for improved scalability, 571 simplicity and flexibility in routing and control plane design. 573 4.5 Other Impairment Considerations 575 There are many other types of impairments that can degrade 576 performance. In this section we briefly mention one other type of 577 impairment, which we propose be dealt with by either by the system 578 designer or by the transmission engineers at the time the system is 579 installed. If dealt with successfully in this manner they should not 580 need to be considered in the dynamic routing process. 582 Gain Nonuniformity and Gain Transients For simple noise estimates to 583 be of use, the amplifiers must be gain-flattened and must have 584 automatic gain control (AGC). Furthermore, each link should have 585 dynamic gain equalization (DGE) to optimize power levels each time 586 wavelengths are added or dropped. Variable optical attenuators on 587 the output ports of an OXC or OADM can be used for this purpose, and 588 in-line devices are starting to become commercially available. 589 Optical channel monitors are also required to provide feedback to 590 the DGEs. AGC must be done rapidly if signal degradation after a 591 protection switch or link failure is to be avoided. 593 Note that the impairments considered here are treated more or less 594 independently. By considering them jointly and varying the tradeoffs 595 between the effects from different components may allow more routes 596 to be feasible. If that is desirable or the system is designed such 597 that certain impairments (e.g. nonlinearities) need to be considered 598 by a centralized process, then distributed routing is not the one to 599 use. 601 4.6 An Alternative Approach - Using Maximum Distance As The only 602 Constraint 604 Today, carriers often use maximum distance to engineer point-to- 605 point OTS systems given a fixed per-span length based on the OSNR 606 constraint for a given bit rate. They may desire to keep the same 607 engineering rule when they move to all-optical networks. Here, we 608 discuss the assumptions that need to be satisfied to keep this 609 approach viable and how to treat the network elements between two 610 adjacent links. 612 Impairments And Other Constraints December 2002 613 On Optical Layer Routing 615 In order to use the maximum distance for a given bit rate to meet an 616 OSNR constraint as the only binding constraint, the operators need 617 to satisfy the following constraints in their all-optical networks: 619 - All the other non-OSNR constraints described in the previous 620 subsections are not binding factors as long as the maximum 621 distance constraint is met. 622 - Specifically for PMD, this means that the whole all-optical 623 network is built on top of sufficiently low-PMD fiber such that 624 the upper bound on the mean aggregate path DGD is always 625 satisfied for any path that does not exceed the maximum 626 distance, or PMD compensation devices might be used for routes 627 with high-PMD fibers. 628 - In terms of the ASE/OSNR constraint, in order to convert the ASE 629 constraint into a distance constraint directly, the network 630 needs to have a fixed fiber distance D for each span (so that 631 ASE can be directly mapped by the gain of the amplifier which 632 equals to the loss of the previous fiber span), e.g., 80km 633 spacing which is commonly chosen by carriers. However, when 634 spans have variable lengths, certain adjustment and compromise 635 need to be made in order to avoid treating ASE explicitly as in 636 section 4.3. These include: 1) If a span is shorter than a 637 typical span length D, unless certain mechanism is built in the 638 OTS to take advantages of shorter spans, it needs to be treated 639 as a span of length D instead of with its real length. 2) When 640 there are spans that are longer than D, it means that paths with 641 these longer spans would have higher average span loss. In 642 general, the maximum system reach decreases when the average 643 span loss increases. Thus, in order to accommodate longer spans 644 in the network, the maximum distance upper bound has to be set 645 with respect to the average span loss of the worst path in the 646 network. This sub-optimality may be acceptable for some networks 647 if the variance is not too large, but may be too conservative 648 for others. 650 If these assumptions are satisfied, the second issue we need to 651 address is how to treat a transparent network element (e.g., MEMS- 652 based switch) between two adjacent links in terms of a distance 653 constraint since it also introduces an insertion loss. If the 654 network element cannot somehow compensate for this OSNR degradation, 655 one approach is to convert each network element into an equivalent 656 length of fiber based on its loss/ASE contribution. Hence, in 657 general, introducing a set of transparent network elements would 658 Impairments And Other Constraints December 2002 659 On Optical Layer Routing 661 effectively result in reducing the overall actual transmission 662 distance between the OEO edges. 664 With this approach, the link-specific state information is link- 665 distance, the length of a link. It equals to the distance sum of all 666 fiber spans on the link and the equivalent length of fiber for the 667 network element(s) on the link. The constraint is that the sum of 668 all the link-distance over all links of a path should be less than 669 the maximum-path-distance, the upper bound of all paths. 671 4.7 Other Considerations 673 Routing in an all-optical network without wavelength conversion 674 raises several additional issues: 676 - Since the route selected must have the chosen wavelength 677 available on all links, this information needs to be considered 678 in the routing process. This is discussed in [Chaudhuri00], 679 where it is concluded that advertising detailed wavelength 680 availabilities on each link is not likely to scale. Instead 681 they propose an alternative method which probes along a chosen 682 path to determine which wavelengths (if any) are available. 683 This would require a significant addition to the routing logic 684 normally used in OSPF. Others have proposed simultaneously 685 probing along multiple paths. 687 - Choosing a path first and then a wavelength along the path is 688 known to give adequate results in simple topologies such as 689 rings and trees ([Yates99]). This does not appear to be true in 690 large mesh networks under realistic provisioning scenarios, 691 however. Instead significantly better results are achieved if 692 wavelength and route are chosen simultaneously ([Strand01b]). 693 This approach would however also have a significant effect on 694 OSPF. 696 4.8 Implications For Routing and Control Plane Design 698 If distributed routing is desired, additional state information will 699 be required by the routing to deal with the impairments described in 700 Sections 4.2 - 4.4: 702 - As mentioned earlier, an operator who wants to avoid having to 703 provide impairment-related parameters to the control plane may 704 elect not to deal with them at the routing level, instead 705 treating them at the system design and planning level if that is 706 a viable approach for their network. In this approach the 707 operator can pre-qualify all or a set of feasible end-to-end 708 Impairments And Other Constraints December 2002 709 On Optical Layer Routing 711 optical paths through the domain of transparency for each bit 712 rate. This approach may work well with relatively small and 713 sparse networks, but it may not be scalable for large and dense 714 networks where the number of feasible paths can be very large. 716 - If the optical paths are not pre-qualified, additional link- 717 specific state information will be required by the routing 718 algorithm for each type of impairment that has the potential of 719 being limiting for some routes. Note that for one operator, PMD 720 might be the only limiting constraint while for another, ASE 721 might be the only one, or it could be both plus some other 722 constraints considered in this document. Some networks might not 723 be limited by any of these constraints. 725 - For an operator needing to deal explicitly with these 726 constraints, the link-dependent information identified above for 727 PMD is link-PMD-square which is the square of the total PMD on a 728 link. For ASE the link-dependent information identified is link- 729 noise which is the total noise on a link. Other link-dependent 730 information includes link-span-length which is the total number 731 of spans on a link, link-crosstalk or OADM-OXC-number which is 732 the total crosstalk or the number of OADM/OXC nodes on a link, 733 respectively, and filter-number which is the number of narrow 734 filters on a link. When the alternative distance-only approach 735 is chosen, the link-specific information is link-distance. 737 - In addition to the link-specific information, bounds on each of 738 the impairments need to be quantified. Since these bounds are 739 determined by the system designer's impairment allocations, 740 these will be system dependent. For PMD, the constraint is that 741 the sum of the link-PMD-square of all links on the transparent 742 segment is less than the square of (a/B) where B is the bit 743 rate. Hence, the required information is the parameter "a". For 744 ASE, the constraint is that the sum of the link-noise of all 745 links is no larger than P/SNRmin. Thus, the information needed 746 include the launch power P and OSNR requirement SNRmin. The 747 minimum acceptable OSNR, in turn, depends on the strength of the 748 FEC being used and the margins reserved for other types of 749 impairments. Other bounds include the maximum span length of the 750 transmission system, the maximum path crosstalk or the maximum 751 number of OADM/OXC nodes, and the maximum number of narrow 752 filters, all are bit rate dependent. With the alternative 753 distance-only approach, the upper bound is the maximum-path- 754 distance. In single-vendor "islands" some of these parameters 755 may be available in a local or EMS database and would not need 756 to be advertised 758 - It is likely that the physical layer parameters do not change 759 value rapidly and could be stored in some database; however 760 these are physical layer parameters that today are frequently 761 Impairments And Other Constraints December 2002 762 On Optical Layer Routing 764 not known at the granularity required. If the ingress node of a 765 lightpath does path selection these parameters would need to be 766 available at this node. 768 - The specific constraints required in a given situation will 769 depend on the design and engineering of the domain of 770 transparency; for example it will be essential to know whether 771 chromatic dispersion has been dealt with on a per-link basis, 772 and whether the domain is operating in a linear or nonlinear 773 regime. 775 - As optical transport technology evolves, the set of constraints 776 that will need to be considered either explicitly or via a 777 domain-wide margin may change. The routing and control plane 778 design should therefore be as open as possible, allowing 779 parameters to be included as necessary. 781 - In the absence of wavelength conversion, the necessity of 782 finding a single wavelength that is available on all links 783 introduces the need to either advertise detailed information on 784 wavelength availability, which probably doesn't scale, or have 785 some mechanism for probing potential routes with or without 786 crankback to determine wavelength availability. Choosing the 787 route first, and then the wavelength, may not yield acceptable 788 utilization levels in mesh-type networks. 790 5. More Complex Networks 792 Mixing optical equipment in a single domain of transparency that has 793 not been explicitly designed to interwork is beyond the scope of 794 this document. This includes most multi-vendor all-optical networks. 796 An optical network composed of multiple domains of transparency 797 optically isolated from each other by O/E/O devices (transponders) 798 is more plausible. A network composed of both "opaque" (optically 799 isolated) OLXC's and one or more all-optical "islands" isolated by 800 transponders is of particular interest because this is most likely 801 how all-optical technologies (such as that described in Sec. 2) are 802 going to be introduced. (We use the term "island" in this discussion 803 rather than a term like "domain" or "area" because these terms are 804 associated with specific approaches like BGP or OSPF.) 806 We consider the complexities raised by these alternatives now. 808 The first requirement for routing in a multi-island network is that 809 the routing process needs to know the extent of each island. There 810 are several reasons for this: 812 Impairments And Other Constraints December 2002 813 On Optical Layer Routing 815 - When entering or leaving an all-optical island, the regeneration 816 process cleans up the optical impairments discussed in Sec. 3. 817 - Each all-optical island may have its own bounds on each 818 impairment. 819 - The routing process needs to be sensitive to the costs 820 associated with "island-hopping". 822 This last point needs elaboration. It is extremely important to 823 realize that, at least in the short to intermediate term, the 824 resources committed by a single routing decision can be very 825 significant: The equipment tied up by a single coast-to-coast OC-192 826 can easily have a first cost of $10**6, and the holding times on a 827 circuit once established is likely to be measured in months. 828 Carriers will expect the routing algorithms used to be sensitive to 829 these costs. Simplistic measures of cost such as the number of 830 "hops" are not likely to be acceptable. 832 Taking the case of an all-optical island consisting of an "ultra 833 long-haul" system like that in Fig. 3-1 embedded in an OEO network 834 of electrical fabric OLXC's as an example: It is likely that the ULH 835 system will be relatively expensive for short hops but relatively 836 economical for longer distances. It is therefore likely to be 837 deployed as a sort of "express backbone". In this scenario a carrier 838 is likely to expect the routing algorithm to balance OEO costs 839 against the additional costs associated with ULH technology and 840 route circuitously to make maximum use of the backbone where 841 appropriate. Note that the metrics used to do this must be 842 consistent throughout the routing domain if this expectation is to 843 be met. 845 The first-order implications for GMPLS seem to be: 846 - Information about island boundaries needs to be advertised. 847 - The routing algorithm needs to be sensitive to island 848 transitions and to the connectivity limitations and impairment 849 constraints particular to each island. 850 - The cost function used in routing must allow the balancing of 851 transponder costs, OXC and OADM costs, and line haul costs 852 across the entire routing domain. 854 Several distributed approaches to multi-island routing seem worth 855 investigating: 856 - Advertise the internal topology and constraints of each island 857 globally; let the ingress node compute an end-to-end strict 858 explicit route sensitive to all constraints and wavelength 859 availabilities. In this approach the routing algorithm used by 860 the ingress node must be able to deal with the details of 861 routing within each island. 862 - Have the EMS or control plane of each island determine and 863 advertise the connectivity between its boundary nodes together 864 with additional information such as costs and the bit rates and 865 Impairments And Other Constraints December 2002 866 On Optical Layer Routing 868 formats supported. As the spare capacity situation changes, 869 updates would be advertised. In this approach impairment 870 constraints are handled within each island and impairment- 871 related parameters need not be advertised outside of the island. 872 The ingress node would then do a loose explicit route and leave 873 the routing and wavelength selection within each island to the 874 island. 875 - Have the ingress node send out probes or queries to nearby 876 gateway nodes or to an NMS to get routing guidance. 878 6. Diversity 880 6.1 Background On Diversity 882 "Diversity" is a relationship between lightpaths. Two lightpaths are 883 said to be diverse if they have no single point of failure. In 884 traditional telephony the dominant transport failure mode is a 885 failure in the interoffice plant, such as a fiber cut inflicted by a 886 backhoe. 888 Why is diversity a unique problem that needs to be considered for 889 optical networks? So far, data network operators have relied on 890 their private line providers to ensure diversity and so have not had 891 to deal directly with the problem. GMPLS makes the complexities 892 handled by the private line provisioning process, including 893 diversity, part of the common control plane and so visible to all. 895 To determine whether two lightpath routings are diverse it is 896 necessary to identify single points of failure in the interoffice 897 plant. To do so we will use the following terms: A fiber cable is a 898 uniform group of fibers contained in a sheath. An Optical Transport 899 System will occupy fibers in a sequence of fiber cables. Each fiber 900 cable will be placed in a sequence of conduits - buried honeycomb 901 structures through which fiber cables may be pulled - or buried in a 902 right of way (ROW). A ROW is land in which the network operator has 903 the right to install his conduit or fiber cable. It is worth noting 904 that for economic reasons, ROW�s are frequently obtained from 905 railroads, pipeline companies, or thruways. It is frequently the 906 case that several carriers may lease ROW from the same source; this 907 makes it common to have a number of carriers� fiber cables in close 908 proximity to each other. Similarly, in a metropolitan network, 909 several carriers might be leasing duct space in the same RBOC 910 conduit. There are also "carrier's carriers" - optical networks 911 which provide fibers to multiple carriers, all of whom could be 912 affected by a single failure in the "carrier's carrier" network. 914 Impairments And Other Constraints December 2002 915 On Optical Layer Routing 917 In a typical intercity facility network there might be on the order 918 of 100 offices that are candidates for OLXC�s. To represent the 919 inter-office fiber network accurately a network with an order of 920 magnitude more nodes is required. In addition to Optical Amplifier 921 (OA) sites, these additional nodes include: 922 - Places where fiber cables enter/leave a conduit or right of way; 923 - Locations where fiber cables cross; 924 Locations where fiber splices are used to interchange fibers between 925 fiber cables. 926 An example of the first might be: 927 A B 928 A-------------B \ / 929 \ / 930 X-----Y 931 / \ 932 C-------------D / \ 933 C D 935 (a) Fiber Cable Topology (b) Right-Of-Way/Conduit Topology 937 Figure 6-1: Fiber Cable vs. ROW Topologies 939 Here the A-B fiber cable would be physically routed A-X-Y-B and the 940 C-D cable would be physically routed C-X-Y-D. This topology might 941 arise because of some physical bottleneck: X-Y might be the Lincoln 942 Tunnel, for example, or the Bay Bridge. 944 Fiber route crossing (the second case) is really a special case of 945 this, where X and Y coincide. In this case the crossing point may 946 not even be a manhole; the fiber routes might just be buried at 947 different depths. 949 Fiber splicing (the third case) often occurs when a major fiber 950 route passes near to a small office. To avoid the expense and 951 additional transmission loss only a small number of fibers are 952 spliced out of the major route into a smaller route going to the 953 small office. This might well occur in a manhole or hut. An 954 example is shown in Fig. 6-2(a), where A-X-B is the major route, X 955 the manhole, and C the smaller office. The actual fiber topology 956 would then look like Fig. 6-2(b), where there would typically be 957 many more A-B fibers than A-C or C-B fibers, and where A-C and C-B 958 might have different numbers of fibers. (One of the latter might 959 even be missing.) 960 Impairments And Other Constraints December 2002 961 On Optical Layer Routing 963 C C 964 | / \ 965 | / \ 966 | / \ 967 A------X------B A---------------B 969 (a) Fiber Cable Topology (b) Fiber Topology 971 Figure 6-2. Fiber Cable vs Fiber Topologies 973 The imminent deployment of ultra-long (>1000 km) Optical Transport 974 Systems introduces a further complexity: Two OTS's could interact a 975 number of times. To make up a hypothetical example: A New York - 976 Atlanta OTS and a Philadelphia - Orlando OTS might ride on the same 977 right of way for x miles in Maryland and then again for y miles in 978 Georgia. They might also cross at Raleigh or some other intermediate 979 node without sharing right of way. 981 Diversity is often equated to routing two lightpaths between a 982 single pair of points, or different pairs of points so that no 983 single route failure will disrupt them both. This is too simplistic, 984 for a number of reasons: 986 - A sophisticated client of an optical network will want to derive 987 diversity needs from his/her end customers' availability 988 requirements. These often lead to more complex diversity 989 requirements than simply providing diversity between two 990 lightpaths. For example, a common requirement is that no single 991 failure should isolate a node or nodes. If a node A has single 992 lightpaths to nodes B and C, this requires A-B and A-C to be 993 diverse. In real applications, a large data network with N 994 lightpaths between its routers might describe their needs in an 995 NxN matrix, where (i,j) defines whether lightpaths i and j must 996 be diverse. 998 - Two circuits that might be considered diverse for one 999 application might not be considered diverse for in another 1000 situation. Diversity is usually thought of as a reaction to 1001 interoffice route failures. High reliability applications may 1002 require other types of failures to be taken into account. Some 1003 examples: 1004 o Office Outages: Although less frequent than route failures, 1005 fires, power outages, and floods do occur. Many network 1006 Impairments And Other Constraints December 2002 1007 On Optical Layer Routing 1009 managers require that diverse routes have no (intermediate) 1010 nodes in common. In other cases an intermediate node might 1011 be acceptable as long as there is power diversity within 1012 the office. 1013 o Shared Rings: Many applications are willing to allow 1014 "diverse" circuits to share a SONET ring-protected link; 1015 presumably they would allow the same for optical layer 1016 rings. 1017 o Disasters: Earthquakes and floods can cause failures over 1018 an extended area. Defense Department circuits might need 1019 to be routed with nuclear damage radii taken into account. 1020 - Conversely, some networks may be willing to take somewhat larger 1021 risks. Taking route failures as an example: Such a network 1022 might be willing to consider two fiber cables in heavy duty 1023 concrete conduit as having a low enough chance of simultaneous 1024 failure to be considered "diverse". They might also be willing 1025 to view two fiber cables buried on opposite sides of a railroad 1026 track as being diverse because there is minimal danger of a 1027 single backhoe disrupting them both even though a bad train 1028 wreck might jeopardize them both. A network seeking N mutually 1029 diverse paths from an office with less than N diverse ROW�s will 1030 need to live with some level of compromise in the immediate 1031 vicinity of the office. 1033 These considerations strongly suggest that the routing algorithm 1034 should be sensitive to the types of threat considered unacceptable 1035 by the requester. Note that the impairment constraints described in 1036 the previous section may eliminate some of the long circuitous 1037 routes sometimes needed to provide diversity. This would make it 1038 harder to find many diverse paths through an all-optical network 1039 than an opaque one. 1041 [Chaudhuri00] introduced the term "Shared Risk Link Group" (SRLG) to 1042 describe the relationship between two non-diverse links. The above 1043 discussion suggests that an SRLG should be characterized by 2 1044 parameters: 1045 - Type of Compromise: Examples would be shared fiber cable, shared 1046 conduit, shared ROW, shared optical ring, shared office without 1047 power sharing, etc.) 1048 - Extent of Compromise: For compromised outside plant, this would 1049 be the length of the sharing. 1050 A CSPF algorithm could then penalize a diversity compromise by an 1051 amount dependent on these two parameters. 1053 Impairments And Other Constraints December 2002 1054 On Optical Layer Routing 1056 Two links could be related by many SRLG's (AT&T's experience 1057 indicates that a link may belong to over 100 SRLG's, each 1058 corresponding to a separate fiber group. Each SRLG might relate a 1059 single link to many other links. For the optical layer, similar 1060 situations can be expected where a link is an ultra-long OTS). 1062 The mapping between links and different types of SRLG�s is in 1063 general defined by network operators based on the definition of each 1064 SRLG type. Since SRLG information is not yet ready to be 1065 discoverable by a network element and does not change dynamically, 1066 it need not be advertised with other resource availability 1067 information by network elements. It could be configured in some 1068 central database and be distributed to or retrieved by the nodes, or 1069 advertised by network elements at the topology discovery stage. 1071 6.2 Implications For Routing 1073 Dealing with diversity is an unavoidable requirement for routing in 1074 the optical layer. It requires dealing with constraints in the 1075 routing process but most importantly requires additional state 1076 information � the SRLG relationships and also the routings of any 1077 existing circuits from the new circuit is to be diverse � to be 1078 available to the routing process. 1080 At present SRLG information cannot be self-discovered. Indeed, in a 1081 large network it is very difficult to maintain accurate SRLG 1082 information. The problem becomes particularly daunting whenever 1083 multiple administrative domains are involved, for instance after the 1084 acquisition of one network by another, because there normally is a 1085 likelihood that there are diversity violations between the domains. 1086 It is very unlikely that diversity relationships between carriers 1087 will be known any time in the near future. 1089 Considerable variation in what different customers will mean by 1090 acceptable diversity should be anticipated. Consequently we suggest 1091 that an SRLG should be defined as follows: (i) It is a relationship 1092 between two or more links, and (ii) it is characterized by two 1093 parameters, the type of compromise (shared conduit, shared ROW, 1094 shared optical ring, etc.) and the extent of the compromise (e.g., 1095 the number of miles over which the compromise persisted). This will 1096 allow the SRLG�s appropriate to a particular routing request to be 1097 easily identified. 1099 7. Security Considerations 1100 Impairments And Other Constraints December 2002 1101 On Optical Layer Routing 1103 We are assuming OEO interfaces to the domain(s) covered by our 1104 discussion (see, e.g., Sec. 4.1 above). If this assumption were to 1105 be relaxed and externally generated optical signals allowed into the 1106 domain, network security issues would arise. Specifically, 1107 unauthorized usage in the form of signals at improper wavelengths or 1108 with power levels or impairments inconsistent with those assumed by 1109 the domain would be possible. With OEO interfaces, these types of 1110 layer one threats should be controllable. 1112 A key layer one security issue is resilience in the face of physical 1113 attack. Diversity, as describe in Sec. 6, is a part of the 1114 solution. However, it is ineffective if there is not sufficient 1115 spare capacity available to make the network whole after an attack. 1116 Several major related issues are: 1117 - Defining the threat: If, for example, an electro-magnetic 1118 interference (EMI) burst is an in-scope threat, then (in the 1119 terminology of Sec. 6) all of the links sufficiently close 1120 together to be disrupted by such a burst must be included in a 1121 single SRLG. Similarly for other threats: For each in-scope 1122 threat, SRLG�s must be defined so that all links vulnerable to a 1123 single incident of the threat must be grouped together in a 1124 single SRLG. 1125 - Allocating responsibility for responding to a layer one failure 1126 between the various layers (especially the optical and IP 1127 layers): This must be clearly specified to avoid churning and 1128 unnecessary service interruptions. 1130 The whole proposed process depends on the integrity of the 1131 impairment characterization information (PMD parameters, etc.) and 1132 also the SRLG definitions. Security of this information, both when 1133 stored and when distributed, is essential. 1135 This document does not address control plane issues, and so control- 1136 plane security is out of scope. IPO control plane security 1137 considerations are discussed in [Rajagopalam02]. Security 1138 considerations for GMPLS, a likely control plane candidate, are 1139 discussed in [Mannie02]. 1141 8. Acknowledgments 1143 This document has benefited from discussions with Michael Eiselt, 1144 Jonathan Lang, Mark Shtaif, Jennifer Yates, Dongmei Wang, Guangzhi 1145 Li, Robert Doverspike, Albert Greenberg, Jim Maloney, John Jacob, 1146 Katie Hall, Diego Caviglia, D. Papadimitriou, O. Audouin, J. P. 1147 Faure, L. Noirie, and with our OIF colleagues. 1149 9. References 1150 Impairments And Other Constraints December 2002 1151 On Optical Layer Routing 1153 9.1 Normative References 1155 [Chaudhuri00] Chaudhuri, S., Hjalmtysson, G., and Yates, J., 1156 "Control of Lightpaths in an Optical Network", Work in Progress, 1157 draft-chaudhuri-ip-olxc-control-00.txt. 1159 [Goldstein94] Goldstein, E. L., Eskildsen, L., and Elrefaie, A. F., 1160 Performance Implications of Component Crosstalk in Transparent 1161 Lightwave Networks", IEEE Photonics Technology Letters, Vol.6, No.5, 1162 May 1994. 1164 [ITU] ITU-T Doc. G.663, Optical Fibers and Amplifiers, Section 1165 II.4.1.2. 1167 [Kaminow97] Kaminow, I. P. and Koch, T. L., editors, Optical Fiber 1168 Telecommunications IIIA, Academic Press, 1997. 1170 [Mannie02] Mannie, E. (ed.), "Generalized Multi-Protocol Label 1171 Switching (GMPLS) Architecture", Interned Draft, draft-ietf-ccamp- 1172 gmpls-architecture-03.txt, August, 2002. 1174 [Rajagopalam02] Rajagopalam, B., et. al., "IP over Optical Networks: 1175 A Framework", Internet Draft, draft-ietf-ipo-framework-02.txt June, 1176 2002. 1178 [Strand01] J. Strand, A. Chiu, and R. Tkach, "Issues for Routing in 1179 the Optical Layer", IEEE Communications Magazine, Feb. 2001, vol. 39 1180 No. 2, pp. 81-88; also see "Unique Features and Requirements for The 1181 Optical Layer Control Plane", Internet Draft, draft-chiu-strand- 1182 unique-olcp-01.txt, work in progress, November 2000. 1184 [Strand01b] J. Strand, R. Doverspike, and G. Li, "Importance of 1185 Wavelength Conversion In An Optical Network", Optical Networks 1186 Magazine, May/June 2001, pp. 33-44. 1188 [Yates99] Yates, J. M., Rumsewicz, M. P. and Lacey, J. P. R., 1189 "Wavelength Converters in Dynamically-Reconfigurable WDM Networks", 1190 IEEE Communications Surveys, 2Q1999 (online at 1191 www.comsoc.org/pubs/surveys/2q99issue/yates.html). 1193 9.2 Informative References 1194 Impairments And Other Constraints December 2002 1195 On Optical Layer Routing 1197 [Awduche99] Awduche, D. O., Rekhter, Y., Drake, J., and Coltun, R., 1198 "Multi-Protocol Lambda Switching: Combining MPLS Traffic Engineering 1199 Control With Optical Crossconnects", Work in Progress, draft- 1200 awduche-mpls-te-optical-01.txt. 1202 [Bra96] Bradner, S., "The Internet Standards Process -- Revision 3," 1203 BCP 9, RFC 2026, October 1996. 1205 [CBD00] Ceuppens, L., Blumenthal, D., Drake, J., Chrostowski, J., 1206 Edwards, W., "Performance Monitoring in Photonic Networks in Support 1207 of MPL(ambda)S", Internet draft, work in progress, March 2000. 1209 [Doverspike00] Doverspike, R. and Yates, J., "Challenges For MPLS in 1210 Optical Network Restoration", IEEE Communication Magazine, February, 1211 2001. 1213 [Gerstel 2000] O. Gorstel, "Optical Layer Signaling: How Much Is 1214 Really Needed?" IEEE Communications Magazine, vol. 38 no. 10, Oct. 1215 2000, pp. 154-160 1217 [KRB01a] Kompella, K., et.al., "IS-IS extensions in support of 1218 Generalized MPLS," Internet Draft, draft-ietf-gmpls- extensions- 1219 01.txt, work in progress, 2001. 1221 [KRB01b] Kompella, K., et. al., "OSPF extensions in support of 1222 Generalized MPLS," Internet draft, draft-ospf-generalized- mpls- 1223 00.txt, work in progress, March 2001. 1225 [Moy98] Moy, John T., OSPF: Anatomy of an Internet Routing Protocol, 1226 Addison-Wesley, 1998. 1228 [Passmore01] Passmore, D., "Managing Fatter Pipes," Business 1229 Communications Review, August 2001, pp. 20-21. 1231 [Ramaswami98] Ramaswami, R. and Sivarajan, K. N., Optical Networks: 1232 A Practical Perspective, Morgan Kaufmann Publishers, 1998. 1234 [Tkach98] Tkach, R., Goldstein, E., Nagel, J., and Strand, J., 1235 "Fundamental Limits of Optical Transparency", Optical Fiber 1236 Communication Conf., Feb. 1998, pp. 161-162. 1238 10. Contributing Authors 1239 Impairments And Other Constraints December 2002 1240 On Optical Layer Routing 1242 This document was a collective work of a number of people. The text 1243 and content of this document was contributed by the editors and the 1244 co-authors listed below. 1246 Ayan Banerjee 1247 Calient Networks 1248 5853 Rue Ferrari 1249 San Jose, CA 95138 1250 Email: abanerjee@calient.net 1252 Dan Blumenthal 1253 Calient Networks 1254 5853 Rue Ferrari 1255 San Jose, CA 95138 1256 Email: dblumenthal@calient.net 1258 John Drake 1259 Calient Networks 1260 5853 Rue Ferrari 1261 San Jose, CA 95138 1262 Email: jdrake@calient.net 1264 Andre Fredette 1265 Hatteras Networks 1266 PO Box 110025 1267 Research Triangle Park, NC 27709 1268 Email: afredette@hatterasnetworks.com 1270 Nan Froberg 1271 PhotonEx Corporation 1272 200 Metrowest Technology Dr. 1273 Maynard, MA 01754 1274 Email: nfroberg@photonex.com 1276 Taha Landolsi 1277 WorldCom, Inc. 1278 2400 North Glenville Drive 1279 Richardson, TX 75082 1280 Email: taha.landolsi@wcom.com 1282 James V. Luciani 1283 900 Chelmsford St. 1284 Lowell, MA 01851 1285 Email: james_luciani@mindspring.com 1287 Robert Tkach 1288 Celion Networks 1289 1 Sheila Dr., Suite 2 1290 Tinton Falls, NJ 07724 1291 Impairments And Other Constraints December 2002 1292 On Optical Layer Routing 1294 Email: bob.tkach@celion.com 1296 Yong Xue 1297 WorldCom, Inc. 1298 22001 Loudoun County Parkway 1299 Ashburn, VA 20147 1300 Email: yxue@cox.com 1302 11. Editors� Addresses 1304 Angela Chiu 1305 AT&T Labs 1306 200 Laurel Ave., Rm A5-1F13 1307 Middletown, NJ 07748 1308 Phone:(732) 420-9061 1309 Email: chiu@research.att.com 1311 John Strand 1312 AT&T Labs 1313 200 Laurel Ave., Rm A5-1D33 1314 Middletown, NJ 07748 1315 Phone:(732) 420-9036 1316 Email: jls@research.att.com