idnits 2.17.1 draft-ietf-ipo-impairments-05.txt: -(673): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1026): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1060): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1061): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1144): 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 5 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.) ** There are 480 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 (May 2003) is 7652 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 67, but not defined == Missing Reference: 'Strand2001' is mentioned on line 67, but not defined == Unused Reference: 'Bra96' is defined on line 1186, but no explicit reference was found in the text == Unused Reference: 'CBD00' is defined on line 1189, but no explicit reference was found in the text == Unused Reference: 'Doverspike00' is defined on line 1193, but no explicit reference was found in the text == Unused Reference: 'Gerstel 2000' is defined on line 1197, but no explicit reference was found in the text == Unused Reference: 'KRB01a' is defined on line 1201, but no explicit reference was found in the text == Unused Reference: 'KRB01b' is defined on line 1205, but no explicit reference was found in the text == Unused Reference: 'Moy98' is defined on line 1209, but no explicit reference was found in the text -- 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') -- Possible downref: Non-RFC (?) normative reference: 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: 8 errors (**), 0 flaws (~~), 15 warnings (==), 10 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-05.txt Angela Chiu (Editor) 4 Informational Track AT&T 5 Expiration Date: November 2003 7 May 2003 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 On Optical Layer Routing 51 challenges for GMPLS. Optical technology is fundamentally an analog 52 rather than digital technology; and the optical layer is lowest in 53 the transport hierarchy and hence has an intimate relationship with 54 the physical geography of the network. 56 GMPLS already has incorporated extensions to deal with some of the 57 unique aspects of the optical layer. This contribution surveys some 58 of the aspects of optical networks which impact routing and 59 identifies possible GMPLS responses for each. Routing constraints 60 and/or complications arising from the design of network elements, 61 the accumulation of signal impairments, and from the need to 62 guarantee the physical diversity of some circuits are discussed. 64 Since the purpose of this draft is to further the specification of 65 GMPLS, alternative approaches to controlling an optical network are 66 not discussed. For discussions of some broader issues, see 67 [Gerstel2000] and [Strand2001]. 69 The organization of the contribution is as follows: 71 - Section 2 is a section requested by the sub-IP Area management 72 for all new drafts. It explains how this document fits into the 73 Area and into the IPO WG, and why it is appropriate for these 74 groups. 75 - Section 3 describes constraints arising from the design of new 76 software controllable network elements. 77 - Section 4 addresses the constraints in a single all-optical 78 domain without wavelength conversion. 79 - Section 5 extends the discussion to more complex networks 80 incorporating both all-optical and opaque architectures. 81 - Section 6 discusses the impacts of diversity constraints. 82 - Section 7 deals with security requirements. 83 - Section 8 contains acknowledgments. 84 - Section 9 contains references. 85 - Section 10 contains contributing authors' addresses. 86 - Section 11 contains editors' addresses. 88 2. Sub-IP Area Summary And Justification Of Work 89 This draft merges and extends two previous drafts, draft-chiu- 90 strand-unique-olcp-02.txt and draft-banerjee-routing-impairments- 91 00.txt. These two drafts were made IPO working group documents to 92 form a basis for a design team at the Minneapolis meeting, where it 93 was also requested that they be merged to create a requirements 94 document for the WG. 96 In the larger sub-IP Area structure, this merged document describes 97 specific characteristics of optical technology and the requirements 98 On Optical Layer Routing 100 they place on routing and path selection. It is appropriate for the 101 IPO working group because the material is specific to optical 102 networks. It identifies and documents the characteristics of the 103 optical transport network that are important for selecting paths for 104 optical channels, which is a work area for the IPO WG. It is 105 appropriate work for this WG because the material covered is 106 directly aimed at establishing a framework and requirements for 107 routing in an optical network. 109 Related documents are: 110 draft-banerjee-routing-impairments-00.txt 111 draft-parent-obgp-01.txt 112 draft-bernstein-optical-bgp-00.txt 113 draft-hayata-ipo-carrier-needs-00.txt 114 draft-many-carrier-framework-uni-01.txt 115 draft-papadimitriou-ipo-non-linear-routing-impairm-01.txt 117 3. Reconfigurable Network Elements 119 3.1 Technology Background 121 Control plane architectural discussions (e.g., [Awduche99]) usually 122 assume that the only software reconfigurable network element is an 123 optical layer cross-connect (OLXC). There are however other software 124 reconfigurable elements on the horizon, specifically tunable lasers and 125 receivers and reconfigurable optical add-drop multiplexers (OADM's). 126 These elements are illustrated in the following simple example, which 127 is modeled on announced Optical Transport System (OTS) products: 128 + + 129 ---+---+ |\ /| +---+--- 130 ---| A |----|D| X Y |D|----| A |--- 131 ---+---+ |W| +--------+ +--------+ |W| +---+--- 132 : |D|-----| OADM |-----| OADM |-----|D| : 133 ---+---+ |M| +--------+ +--------+ |M| +---+--- 134 ---| A |----| | | | | | | |----| A |--- 135 ---+---+ |/ | | | | \| +---+--- 136 + +---+ +---+ +---+ +---+ + 137 D | A | | A | | A | | A | E 138 +---+ +---+ +---+ +---+ 139 | | | | | | | | 141 Figure 3-1: An OTS With OADM's - Functional Architecture 143 In Fig.3-1, the part that is on the inner side of all boxes labeled 144 "A" defines an all-optical subnetwork. From a routing perspective 145 two aspects are critical: 147 On Optical Layer Routing 149 - Adaptation: These are the functions done at the edges of the 150 subnetwork that transform the incoming optical channel into the 151 physical wavelength to be transported through the subnetwork. 152 - Connectivity: This defines which pairs of edge Adaptation 153 functions can be interconnected through the subnetwork. 155 In Fig. 3-1, D and E are DWDM's and X and Y are OADM's. The boxes 156 labeled "A" are adaptation functions. They map one or more input 157 optical channels assumed to be standard short reach signals into a 158 long reach (LR) wavelength or wavelength group which will pass 159 transparently to a distant adaptation function. Adaptation 160 functionality which affects routing includes: 161 - Multiplexing: Either electrical or optical TDM may be used to 162 combine the input channels into a single wavelength. This is 163 done to increase effective capacity: A typical DWDM might be 164 able to handle 100 2.5 Gb/sec signals (250 Gb/sec total) or 50 165 10 Gb/sec (500 Gb/sec total); combining the 2.5 Gb/sec signals 166 together thus effectively doubles capacity. After multiplexing 167 the combined signal must be routed as a group to the distant 168 adaptation function. 169 - Adaptation Grouping: In this technique, groups of k (e.g., 4) 170 wavelengths are managed as a group within the system and must be 171 added/dropped as a group. We will call such a group an 172 "adaptation grouping". Examples include so called "wave group" 173 and "waveband" [Passmore01]. Groupings on the same system may 174 differ in basics such as wavelength spacing, which constrain the 175 type of channels that can be accommodated. 176 - Laser Tunability: The lasers producing the LR wavelengths may 177 have a fixed frequency, may be tunable over a limited range, or 178 be tunable over the entire range of wavelengths supported by the 179 DWDM. Tunability speeds may also vary. 181 Connectivity between adaptation functions may also be limited: 182 - As pointed out above, TDM multiplexing and/or adaptation 183 grouping by the adaptation function forces groups of input 184 channels to be delivered together to the same distant adaptation 185 function. 186 - Only adaptation functions whose lasers/receivers are tunable to 187 compatible frequencies can be connected. 188 - The switching capability of the OADM's may also be constrained. 189 For example: 190 o There may be some wavelengths that can not be dropped at 191 all. 192 o There may be a fixed relationship between the frequency 193 dropped and the physical port on the OADM to which it is 194 dropped. 195 o OADM physical design may put an upper bound on the number 196 of adaptation groupings dropped at any single OADM. 198 On Optical Layer Routing 200 For a fixed configuration of the OADM's and adaptation functions 201 connectivity will be fixed: Each input port will essentially be 202 hard-wired to some specific distant port. However this connectivity 203 can be changed by changing the configurations of the OADM's and 204 adaptation functions. For example, an additional adaptation grouping 205 might be dropped at an OADM or a tunable laser retuned. In each case 206 the port-to-port connectivity is changed. 208 These capabilities can be expected to be under software control. 209 Today the control would rest in the vendor-supplied Element 210 Management system (EMS), which in turn would be controlled by the 211 operator's OS's. However in principle the EMS could participate in 212 the GMPLS routing process. 214 3.2 Implications For Routing 216 An OTS of the sort discussed in Sec. 3.1 is essentially a 217 geographically distributed but blocking cross-connect system. The 218 specific port connectivity is dependent on the vendor design and 219 also on exactly what line cards have been deployed. 221 One way for GMPLS to deal with this architecture would be to view 222 the port connectivity as externally determined. In this case the 223 links known to GMPLS would be groups of identically routed 224 wavebands. If these were reconfigured by the external EMS the 225 resulting connectivity changes would need to be detected and 226 advertised within GMPLS. If the topology shown in Fig. 3-1 became a 227 tree or a mesh instead of the linear topology shown, the 228 connectivity changes could result in SRLG changes. 230 Alternatively, GMPLS could attempt to directly control this port 231 connectivity. The state information needed to do this is likely to 232 be voluminous and vendor specific. 234 4. Wavelength Routed All-Optical Networks 236 The optical networks presently being deployed may be called "opaque" 237 ([Tkach98]): each link is optically isolated by transponders doing 238 O/E/O conversions. They provide regeneration with retiming and 239 reshaping, also called 3R, which eliminates transparency to bit 240 rates and frame format. These transponders are quite expensive and 241 their lack of transparency also constrains the rapid introduction of 242 new services. Thus there are strong motivators to introduce 243 "domains of transparency" - all-optical subnetworks - larger than an 244 OTS. 246 The routing of lightpaths through an all-optical network has 247 received extensive attention. (See [Yates99] or [Ramaswami98]). 249 On Optical Layer Routing 251 When discussing routing in an all-optical network it is usually 252 assumed that all routes have adequate signal quality. This may be 253 ensured by limiting all-optical networks to subnetworks of limited 254 geographic size which are optically isolated from other parts of the 255 optical layer by transponders. This approach is very practical and 256 has been applied to date, e.g. when determining the maximum length 257 of an Optical Transport System (OTS). Furthermore operational 258 considerations like fault isolation also make limiting the size of 259 domains of transparency attractive. 261 There are however reasons to consider contained domains of 262 transparency in which not all routes have adequate signal quality. 263 From a demand perspective, maximum bit rates have rapidly increased 264 from DS3 to OC-192 and soon OC-768 (40 Gb/sec). As bit rates 265 increase it is necessary to increase power. This makes impairments 266 and nonlinearities more troublesome. From a supply perspective, 267 optical technology is advancing very rapidly, making ever-larger 268 domains possible. In this section we assume that these 269 considerations will lead to the deployment of a domain of 270 transparency that is too large to ensure that all potential routes 271 have adequate signal quality for all circuits. Our goal is to 272 understand the impacts of the various types of impairments in this 273 environment. 275 Note that as we describe later in the section there are many types 276 of physical impairments. Which of these needs to be dealt with 277 explicitly when performing on-line distributed routing will vary 278 considerably and will depend on many variables, including: 279 - Equipment vendor design choices, 280 - Fiber characteristics, 281 - Service characteristics (e.g., circuit speeds), 282 - Network size, 283 - Network operator engineering and deployment strategies. 284 For example, a metropolitan network which does not intend to support 285 bit rates above 2.5 Gb/sec may not be constrained by any of these 286 impairments, while a continental or international network which 287 wished to minimize O/E/O regeneration investment and support 40 288 Gb/sec connections might have to explicitly consider many of them. 289 Also, a network operator may reduce or even eliminate their 290 constraint set by building a relatively small domain of transparency 291 to ensure that all the paths are feasible, or by using some 292 proprietary tools based on rules from the OTS vendor to pre-qualify 293 paths between node pairs and put them in a table that can be 294 accessed each time a routing decision has to be made through that 295 domain. 297 4.1 Problem Formulation 298 On Optical Layer Routing 300 We consider a single domain of transparency without wavelength 301 translation. Additionally due to the proprietary nature of DWDM 302 transmission technology, we assume that the domain is either single 303 vendor or architected using a single coherent design, particularly 304 with regard to the management of impairments. 306 We wish to route a unidirectional circuit from ingress client node X 307 to egress client node Y. At both X and Y, the circuit goes through 308 an O/E/O conversion which optically isolates the portion within our 309 domain. We assume that we know the bit rate of the circuit. Also, 310 we assume that the adaptation function at X may apply some Forward 311 Error Correction (FEC) method to the circuit. We also assume we know 312 the launch power of the laser at X. 314 Impairments can be classified into two categories, linear and 315 nonlinear. (See [Tkach98] for more on impairment constraints). 316 Linear effects are independent of signal power and affect 317 wavelengths individually. Amplifier spontaneous emission (ASE), 318 polarization mode dispersion (PMD), and chromatic dispersion are 319 examples. Nonlinearities are significantly more complex: they 320 generate not only impairments on each channel, but also crosstalk 321 between channels. 323 In the remainder of this section we first outline how two key linear 324 impairments (PMD and ASE) might be handled by a set of analytical 325 formulae as additional constraints on routing. We next discuss how 326 the remaining constraints might be approached. Finally we take a 327 broader perspective and discuss the implications of such constraints 328 on control plane architecture and also on broader constrained domain 329 of transparency architecture issues. 331 4.2 Polarization Mode Dispersion (PMD) 333 For a transparent fiber segment, the general PMD requirement is that 334 the time-average differential group delay (DGD) between two 335 orthogonal state of polarizations should be less than fraction a of 336 the bit duration, T=1/B, where B is the bit rate. The value of the 337 parameter a depends on three major factors: 1) margin allocated to 338 PMD, e.g. 1dB; 2) targeted outage probability, e.g. 4x10-5, and 3) 339 sensitivity of the receiver to DGD. A typical value for a is 10% 340 [ITU]. More aggressive designs to compensate for PMD may allow 341 values higher than 10%. (This would be a system parameter dependent 342 on the system design. It would need to be known to the routing 343 process.) 345 The PMD parameter (Dpmd) is measured in pico-seconds (ps) per 346 sqrt(km). The square of the PMD in a fiber span, denoted as span- 347 PMD-square is then given by the product of Dpmd**2 and the span 348 length. (A fiber span in a transparent network refers to a segment 349 between two optical amplifiers.) If Dpmd is constant, this results 350 On Optical Layer Routing 352 in a upper bound on the maximum length of an M-fiber-span 353 transparent segment, which is inversely proportional to the square 354 of the product of bit rate and Dpmd (the detailed equation is 355 omitted due to the format constraint - see [Strand01] for details). 357 For older fibers with a typical PMD parameter of 0.5 picoseconds per 358 square root of km, based on the constraint, the maximum length of 359 the transparent segment should not exceed 400km and 25km for bit 360 rates of 10Gb/s and 40Gb/s, respectively. Due to recent advances in 361 fiber technology, the PMD-limited distance has increased 362 dramatically. For newer fibers with a PMD parameter of 0.1 363 picosecond per square root of km, the maximum length of the 364 transparent segment (without PMD compensation) is limited to 10000km 365 and 625km for bit rates of 10Gb/s and 40Gb/, respectively. Still 366 lower values of PMD are attainable in commercially available fiber 367 today, and the PMD limit can be further extended if a larger value 368 of the parameter a (ratio of DGD to the bit period) can be 369 tolerated. In general, the PMD requirement is not an issue for most 370 types of fibers at 10Gb/s or lower bit rate. But it will become an 371 issue at bit rates of 40Gb/s and higher. 373 If the PMD parameter varies between spans, a slightly more 374 complicated equation results (see [Strand01]), but in any event the 375 only link dependent information needed by the routing algorithm is 376 the square of the link PMD, denoted as link-PMD-square. It is the 377 sum of the span-PMD-square of all spans on the link. 379 Note that when one has some viable PMD compensation devices and 380 deploy them ubiquitously on all routes with potential PMD issues in 381 the network, then the PMD constraint disappears from the routing 382 perspective. 384 4.3 Amplifier Spontaneous Emission 386 ASE degrades the optical signal to noise ratio (OSNR). An acceptable 387 optical SNR level (SNRmin) which depends on the bit rate, 388 transmitter-receiver technology (e.g., FEC), and margins allocated 389 for the impairments, needs to be maintained at the receiver. In 390 order to satisfy this requirement, vendors often provide some 391 general engineering rule in terms of maximum length of the 392 transparent segment and number of spans. For example, current 393 transmission systems are often limited to up to 6 spans each 80km 394 long. For larger transparent domains, more detailed OSNR 395 computations will be needed to determine whether the OSNR level 396 through a domain of transparency is acceptable. This would provide 397 flexibility in provisioning or restoring a lightpath through a 398 transparent subnetwork. 400 Assume that the average optical power launched at the transmitter is 401 P. The lightpath from the transmitter to the receiver goes through M 402 On Optical Layer Routing 404 optical amplifiers, with each introducing some noise power. Unity 405 gain can be used at all amplifier sites to maintain constant signal 406 power at the input of each span to minimize noise power and 407 nonlinearity. A constraint on the maximum number of spans can be 408 obtained [Kaminow97] which is proportional to P and inversely 409 proportional to SNRmin, optical bandwidth B, amplifier gain G-1 and 410 spontaneous emission factor n of the optical amplifier, assuming all 411 spans have identical gain and noise figure. (Again, the detailed 412 equation is omitted due to the format constraint - see [Strand01] 413 for details.) Let's take a typical example. Assuming P=4dBm, 414 SNRmin=20dB with FEC, B=12.5GHz, n=2.5, G=25dB, based on the 415 constraint, the maximum number of spans is at most 10. However, if 416 FEC is not used and the requirement on SNRmin becomes 25dB, the 417 maximum number of spans drops down to 3. 419 For ASE the only link-dependent information needed by the routing 420 algorithm is the noise of the link, denoted as link-noise, which is 421 the sum of the noise of all spans on the link. Hence the constraint 422 on ASE becomes that the aggregate noise of the transparent segment 423 which is the sum of the link-noise of all links can not exceed 424 P/SNRmin. 426 4.4 Approximating The Effects Of Some Other Impairment Constraints 428 There are a number of other impairment constraints that we believe 429 could be approximated with a domain-wide margin on the OSNR, plus in 430 some cases a constraint on the total number of networking elements 431 (OXC or OADM) along the path. Most impairments generated at OXCs or 432 OADMs, including polarization dependent loss, coherent crosstalk, 433 and effective passband width, could be dealt with using this 434 approach. In principle, impairments generated at the nodes can be 435 bounded by system engineering rules because the node elements can be 436 designed and specified in a uniform manner. This approach is not 437 feasible with PMD and noise because neither can be uniformly 438 specified. Instead, they depend on node spacing and the 439 characteristics of the installed fiber plant, neither of which are 440 likely to be under the system designer's control. 442 Examples of the constraints we propose to approximate with a domain- 443 wide margin are given in the remaining paragraphs in this section. 444 It should be kept in mind that as optical transport technology 445 evolves it may become necessary to include some of these impairments 446 explicitly in the routing process. Other impairments not mentioned 447 here at all may also become sufficiently important to require 448 incorporation either explicitly or via a domain-wide margin. 450 Other Polarization Dependent Impairments Other polarization- 451 dependent effects besides PMD influence system performance. For 452 example, many components have polarization-dependent loss (PDL) 453 [Ramaswami98], which accumulates in a system with many components on 454 On Optical Layer Routing 456 the transmission path. The state of polarization fluctuates with 457 time and its distribution is very important also. It is generally 458 required to maintain the total PDL on the path to be within some 459 acceptable limit, potentially by using some compensation technology 460 for relatively long transmission systems, plus a small built-in 461 margin in OSNR. Since the total PDL increases with the number of 462 components in the data path, it must be taken into account by the 463 system vendor when determining the maximum allowable number of 464 spans. 466 Chromatic Dispersion In general this impairment can be adequately 467 (but not optimally) compensated for on a per-link basis, and/or at 468 system initial setup time. Today most deployed compensation devices 469 are based on DCF (Dispersion Compensation Fiber). DCF provides per 470 fiber compensation by means of a spool of fiber with a CD coefficient 471 opposite to the fiber. Due to the imperfect matching between the CD 472 slope of the fiber and the DCF some lambdas can be over compensated 473 while others can be under compensated. Moreover DCF modules may only 474 be available in fixed lengths of compensating fiber; this means that 475 sometimes it is impossible to find a DCF module that exactly 476 compensates the CD introduced by the fiber. These effects introduce 477 what is known as residual CD. Residual CD varies with the frequency 478 of the wavelength. Knowing the characteristics of both of the fiber 479 and the DCF modules along the path, this can be calculated with a 480 sufficient degree of precision. However this is a very challenging 481 task. In fact the per-wavelength residual dispersion needs to be 482 combined with other information in the system (e.g. types fibers to 483 figure out the amount of nonlinearities) to obtain the net effect of 484 CD either by simulation or by some analytical approximation. It 485 appears that the routing/control plane should not be burdened by such 486 a large set of information while it can be handled at the system 487 design level. Therefore it will be assumed until proven otherwise 488 that residual dispersion should not be reported. For high bit rates, 489 dynamic dispersion compensation may be required at the receiver to 490 clean up any residual dispersion. 492 Crosstalk Optical crosstalk refers to the effect of other signals on 493 the desired signal. It includes both coherent (i.e. intrachannel) 494 crosstalk and incoherent (i.e. interchannel) crosstalk. Main 495 contributors of crosstalk are the OADM and OXC sites that use a DWDM 496 multiplexer/demultiplexer (MUX/DEMUX) pair. For a relatively sparse 497 network where the number of OADM/OXC nodes on a path is low, 498 crosstalk can be treated with a low margin in OSNR without being a 499 binding constraint. But for some relatively dense networks where 500 crosstalk might become a binding constraint, one needs to propagate 501 the per-link crosstalk information to make sure that the end-to-end 502 On Optical Layer Routing 504 path crosstalk which is the sum of the crosstalks on all the 505 corresponding links to be within some limit, e.g. -25dB threshold 506 with 1dB penalty ([Goldstein94]). Another way to treat it without 507 having to propagate per-link crosstalk information is to have the 508 system evaluate what the maximum number of OADM/OXC nodes that has a 509 MUX/DEMUX pair for the worst route in the transparent domain for a 510 low built-in margin. The latter one should work well where all the 511 OXC/OADM nodes have similar level of crosstalk. 513 Effective Passband As more and more DWDM components are cascaded, 514 the effective passband narrows. The number of filters along the 515 link, their passband width and their shape will determine the end- 516 to-end effective passband. In general, this is a system design 517 issue, i.e., the system is designed with certain maximum bit rate 518 using the proper modulation format and filter spacing. For linear 519 systems, the filter effect can be turned into a constraint on the 520 maximum number of narrow filters with the condition that filters in 521 the systems are at least as wide as the one in the receiver. 522 Because traffic at lower bit rates can tolerate a narrower passband, 523 the maximum allowable number of narrow filters will increase as the 524 bit rate decreases. 526 Nonlinear Impairments It seems unlikely that these can be dealt with 527 explicitly in a routing algorithm because they lead to constraints 528 that can couple routes together and lead to complex dependencies, 529 e.g. on the order in which specific fiber types are traversed 530 [Kaminow97]. Note that different fiber types (standard single mode 531 fiber, dispersion shifted fiber, dispersion compensated fiber, etc.) 532 have very different effects from nonlinear impairments. A full 533 treatment of the nonlinear constraints would likely require very 534 detailed knowledge of the physical infrastructure, including 535 measured dispersion values for each span, fiber core area and 536 composition, as well as knowledge of subsystem details such as 537 dispersion compensation technology. This information would need to 538 be combined with knowledge of the current loading of optical signals 539 on the links of interest to determine the level of nonlinear 540 impairment. Alternatively, one could assume that nonlinear 541 impairments are bounded and result in X dB margin in the required 542 OSNR level for a given bit rate, where X for performance reasons 543 would be limited to 1 or 2 dB, consequently setting a limit on the 544 maximum number of spans. For the approach described here to be 545 useful, it is desirable for this span length limit to be longer than 546 that imposed by the constraints which can be treated explicitly. 547 When designing a DWDM transport system, there are tradeoffs between 548 signal power launched at the transmitter, span length, and nonlinear 549 effects on BER that need to be considered jointly. Here, we assume 550 that an X dB margin is obtained after the transport system has been 551 designed with a fixed signal power and maximum span length for a 552 given bit rate. Note that OTSs can be designed in very different 553 ways, in linear, pseudo-linear, or nonlinear environments. The X-dB 554 On Optical Layer Routing 556 margin approach may be valid for some but not for others. However, 557 it is likely that there is an advantage in designing systems that 558 are less aggressive with respect to nonlinearities, and therefore 559 somewhat sub-optimal, in exchange for improved scalability, 560 simplicity and flexibility in routing and control plane design. 562 4.5 Other Impairment Considerations 564 There are many other types of impairments that can degrade 565 performance. In this section we briefly mention one other type of 566 impairment, which we propose be dealt with by either by the system 567 designer or by the transmission engineers at the time the system is 568 installed. If dealt with successfully in this manner they should not 569 need to be considered in the dynamic routing process. 571 Gain Nonuniformity and Gain Transients For simple noise estimates to 572 be of use, the amplifiers must be gain-flattened and must have 573 automatic gain control (AGC). Furthermore, each link should have 574 dynamic gain equalization (DGE) to optimize power levels each time 575 wavelengths are added or dropped. Variable optical attenuators on 576 the output ports of an OXC or OADM can be used for this purpose, and 577 in-line devices are starting to become commercially available. 578 Optical channel monitors are also required to provide feedback to 579 the DGEs. AGC must be done rapidly if signal degradation after a 580 protection switch or link failure is to be avoided. 582 Note that the impairments considered here are treated more or less 583 independently. By considering them jointly and varying the tradeoffs 584 between the effects from different components may allow more routes 585 to be feasible. If that is desirable or the system is designed such 586 that certain impairments (e.g. nonlinearities) need to be considered 587 by a centralized process, then distributed routing is not the one to 588 use. 590 4.6 An Alternative Approach - Using Maximum Distance As The only 591 Constraint 593 Today, carriers often use maximum distance to engineer point-to- 594 point OTS systems given a fixed per-span length based on the OSNR 595 constraint for a given bit rate. They may desire to keep the same 596 engineering rule when they move to all-optical networks. Here, we 597 discuss the assumptions that need to be satisfied to keep this 598 approach viable and how to treat the network elements between two 599 adjacent links. 601 On Optical Layer Routing 603 In order to use the maximum distance for a given bit rate to meet an 604 OSNR constraint as the only binding constraint, the operators need 605 to satisfy the following constraints in their all-optical networks: 607 - All the other non-OSNR constraints described in the previous 608 subsections are not binding factors as long as the maximum 609 distance constraint is met. 610 - Specifically for PMD, this means that the whole all-optical 611 network is built on top of sufficiently low-PMD fiber such that 612 the upper bound on the mean aggregate path DGD is always 613 satisfied for any path that does not exceed the maximum 614 distance, or PMD compensation devices might be used for routes 615 with high-PMD fibers. 616 - In terms of the ASE/OSNR constraint, in order to convert the ASE 617 constraint into a distance constraint directly, the network 618 needs to have a fixed fiber distance D for each span (so that 619 ASE can be directly mapped by the gain of the amplifier which 620 equals to the loss of the previous fiber span), e.g., 80km 621 spacing which is commonly chosen by carriers. However, when 622 spans have variable lengths, certain adjustment and compromise 623 need to be made in order to avoid treating ASE explicitly as in 624 section 4.3. These include: 1) If a span is shorter than a 625 typical span length D, unless certain mechanism is built in the 626 OTS to take advantages of shorter spans, it needs to be treated 627 as a span of length D instead of with its real length. 2) When 628 there are spans that are longer than D, it means that paths with 629 these longer spans would have higher average span loss. In 630 general, the maximum system reach decreases when the average 631 span loss increases. Thus, in order to accommodate longer spans 632 in the network, the maximum distance upper bound has to be set 633 with respect to the average span loss of the worst path in the 634 network. This sub-optimality may be acceptable for some networks 635 if the variance is not too large, but may be too conservative 636 for others. 638 If these assumptions are satisfied, the second issue we need to 639 address is how to treat a transparent network element (e.g., MEMS- 640 based switch) between two adjacent links in terms of a distance 641 constraint since it also introduces an insertion loss. If the 642 network element cannot somehow compensate for this OSNR degradation, 643 one approach is to convert each network element into an equivalent 644 length of fiber based on its loss/ASE contribution. Hence, in 645 general, introducing a set of transparent network elements would 646 On Optical Layer Routing 648 effectively result in reducing the overall actual transmission 649 distance between the OEO edges. 651 With this approach, the link-specific state information is link- 652 distance, the length of a link. It equals to the distance sum of all 653 fiber spans on the link and the equivalent length of fiber for the 654 network element(s) on the link. The constraint is that the sum of 655 all the link-distance over all links of a path should be less than 656 the maximum-path-distance, the upper bound of all paths. 658 4.7 Other Considerations 660 Routing in an all-optical network without wavelength conversion 661 raises several additional issues: 663 - Since the route selected must have the chosen wavelength 664 available on all links, this information needs to be considered 665 in the routing process. One approach is to propagate 666 information throughout the network about the state of every 667 wavelength on every link in the network. However, the state 668 required and the overhead involved in processing and 669 maintaining this information is proportional to the total 670 number of links (thus, number of nodes squared), maximum number 671 of wavelengths which keeps doubling every couple of years), and 672 the frequency of wavelength availability changes, which can be 673 very high. Instead [Hj�lmt�sson00] proposes an alternative 674 method which probes along a chosen path to determine which 675 wavelengths (if any) are available. This would require a 676 significant addition to the routing logic normally used in 677 OSPF. Others have proposed simultaneously probing along 678 multiple paths. 680 - Choosing a path first and then a wavelength along the path is 681 known to give adequate results in simple topologies such as 682 rings and trees ([Yates99]). This does not appear to be true in 683 large mesh networks under realistic provisioning scenarios, 684 however. Instead significantly better results are achieved if 685 wavelength and route are chosen simultaneously ([Strand01b]). 686 This approach would however also have a significant effect on 687 OSPF. 689 4.8 Implications For Routing and Control Plane Design 691 If distributed routing is desired, additional state information will 692 be required by the routing to deal with the impairments described in 693 Sections 4.2 - 4.4: 695 On Optical Layer Routing 697 - As mentioned earlier, an operator who wants to avoid having to 698 provide impairment-related parameters to the control plane may 699 elect not to deal with them at the routing level, instead 700 treating them at the system design and planning level if that is 701 a viable approach for their network. In this approach the 702 operator can pre-qualify all or a set of feasible end-to-end 703 optical paths through the domain of transparency for each bit 704 rate. This approach may work well with relatively small and 705 sparse networks, but it may not be scalable for large and dense 706 networks where the number of feasible paths can be very large. 708 - If the optical paths are not pre-qualified, additional link- 709 specific state information will be required by the routing 710 algorithm for each type of impairment that has the potential of 711 being limiting for some routes. Note that for one operator, PMD 712 might be the only limiting constraint while for another, ASE 713 might be the only one, or it could be both plus some other 714 constraints considered in this document. Some networks might not 715 be limited by any of these constraints. 717 - For an operator needing to deal explicitly with these 718 constraints, the link-dependent information identified above for 719 PMD is link-PMD-square which is the square of the total PMD on a 720 link. For ASE the link-dependent information identified is link- 721 noise which is the total noise on a link. Other link-dependent 722 information includes link-span-length which is the total number 723 of spans on a link, link-crosstalk or OADM-OXC-number which is 724 the total crosstalk or the number of OADM/OXC nodes on a link, 725 respectively, and filter-number which is the number of narrow 726 filters on a link. When the alternative distance-only approach 727 is chosen, the link-specific information is link-distance. 729 - In addition to the link-specific information, bounds on each of 730 the impairments need to be quantified. Since these bounds are 731 determined by the system designer's impairment allocations, 732 these will be system dependent. For PMD, the constraint is that 733 the sum of the link-PMD-square of all links on the transparent 734 segment is less than the square of (a/B) where B is the bit 735 rate. Hence, the required information is the parameter "a". For 736 ASE, the constraint is that the sum of the link-noise of all 737 links is no larger than P/SNRmin. Thus, the information needed 738 include the launch power P and OSNR requirement SNRmin. The 739 minimum acceptable OSNR, in turn, depends on the strength of the 740 FEC being used and the margins reserved for other types of 741 impairments. Other bounds include the maximum span length of the 742 transmission system, the maximum path crosstalk or the maximum 743 number of OADM/OXC nodes, and the maximum number of narrow 744 filters, all are bit rate dependent. With the alternative 745 distance-only approach, the upper bound is the maximum-path- 746 distance. In single-vendor "islands" some of these parameters 747 On Optical Layer Routing 749 may be available in a local or EMS database and would not need 750 to be advertised 752 - It is likely that the physical layer parameters do not change 753 value rapidly and could be stored in some database; however 754 these are physical layer parameters that today are frequently 755 not known at the granularity required. If the ingress node of a 756 lightpath does path selection these parameters would need to be 757 available at this node. 759 - The specific constraints required in a given situation will 760 depend on the design and engineering of the domain of 761 transparency; for example it will be essential to know whether 762 chromatic dispersion has been dealt with on a per-link basis, 763 and whether the domain is operating in a linear or nonlinear 764 regime. 766 - As optical transport technology evolves, the set of constraints 767 that will need to be considered either explicitly or via a 768 domain-wide margin may change. The routing and control plane 769 design should therefore be as open as possible, allowing 770 parameters to be included as necessary. 772 - In the absence of wavelength conversion, the necessity of 773 finding a single wavelength that is available on all links 774 introduces the need to either advertise detailed information on 775 wavelength availability, which probably doesn't scale, or have 776 some mechanism for probing potential routes with or without 777 crankback to determine wavelength availability. Choosing the 778 route first, and then the wavelength, may not yield acceptable 779 utilization levels in mesh-type networks. 781 5. More Complex Networks 783 Mixing optical equipment in a single domain of transparency that has 784 not been explicitly designed to interwork is beyond the scope of 785 this document. This includes most multi-vendor all-optical networks. 787 An optical network composed of multiple domains of transparency 788 optically isolated from each other by O/E/O devices (transponders) 789 is more plausible. A network composed of both "opaque" (optically 790 isolated) OLXC's and one or more all-optical "islands" isolated by 791 transponders is of particular interest because this is most likely 792 how all-optical technologies (such as that described in Sec. 2) are 793 going to be introduced. (We use the term "island" in this discussion 794 rather than a term like "domain" or "area" because these terms are 795 associated with specific approaches like BGP or OSPF.) 796 On Optical Layer Routing 798 We consider the complexities raised by these alternatives now. 800 The first requirement for routing in a multi-island network is that 801 the routing process needs to know the extent of each island. There 802 are several reasons for this: 803 - When entering or leaving an all-optical island, the regeneration 804 process cleans up the optical impairments discussed in Sec. 3. 805 - Each all-optical island may have its own bounds on each 806 impairment. 807 - The routing process needs to be sensitive to the costs 808 associated with "island-hopping". 810 This last point needs elaboration. It is extremely important to 811 realize that, at least in the short to intermediate term, the 812 resources committed by a single routing decision can be very 813 significant: The equipment tied up by a single coast-to-coast OC-192 814 can easily have a first cost of $10**6, and the holding times on a 815 circuit once established is likely to be measured in months. 816 Carriers will expect the routing algorithms used to be sensitive to 817 these costs. Simplistic measures of cost such as the number of 818 "hops" are not likely to be acceptable. 820 Taking the case of an all-optical island consisting of an "ultra 821 long-haul" system like that in Fig. 3-1 embedded in an OEO network 822 of electrical fabric OLXC's as an example: It is likely that the ULH 823 system will be relatively expensive for short hops but relatively 824 economical for longer distances. It is therefore likely to be 825 deployed as a sort of "express backbone". In this scenario a carrier 826 is likely to expect the routing algorithm to balance OEO costs 827 against the additional costs associated with ULH technology and 828 route circuitously to make maximum use of the backbone where 829 appropriate. Note that the metrics used to do this must be 830 consistent throughout the routing domain if this expectation is to 831 be met. 833 The first-order implications for GMPLS seem to be: 834 - Information about island boundaries needs to be advertised. 835 - The routing algorithm needs to be sensitive to island 836 transitions and to the connectivity limitations and impairment 837 constraints particular to each island. 838 - The cost function used in routing must allow the balancing of 839 transponder costs, OXC and OADM costs, and line haul costs 840 across the entire routing domain. 842 Several distributed approaches to multi-island routing seem worth 843 investigating: 844 - Advertise the internal topology and constraints of each island 845 globally; let the ingress node compute an end-to-end strict 846 explicit route sensitive to all constraints and wavelength 847 availabilities. In this approach the routing algorithm used by 848 On Optical Layer Routing 850 the ingress node must be able to deal with the details of 851 routing within each island. 852 - Have the EMS or control plane of each island determine and 853 advertise the connectivity between its boundary nodes together 854 with additional information such as costs and the bit rates and 855 formats supported. As the spare capacity situation changes, 856 updates would be advertised. In this approach impairment 857 constraints are handled within each island and impairment- 858 related parameters need not be advertised outside of the island. 859 The ingress node would then do a loose explicit route and leave 860 the routing and wavelength selection within each island to the 861 island. 862 - Have the ingress node send out probes or queries to nearby 863 gateway nodes or to an NMS to get routing guidance. 865 6. Diversity 867 6.1 Background On Diversity 869 "Diversity" is a relationship between lightpaths. Two lightpaths are 870 said to be diverse if they have no single point of failure. In 871 traditional telephony the dominant transport failure mode is a 872 failure in the interoffice plant, such as a fiber cut inflicted by a 873 backhoe. 875 Why is diversity a unique problem that needs to be considered for 876 optical networks? So far, data network operators have relied on 877 their private line providers to ensure diversity and so have not had 878 to deal directly with the problem. GMPLS makes the complexities 879 handled by the private line provisioning process, including 880 diversity, part of the common control plane and so visible to all. 882 To determine whether two lightpath routings are diverse it is 883 necessary to identify single points of failure in the interoffice 884 plant. To do so we will use the following terms: A fiber cable is a 885 uniform group of fibers contained in a sheath. An Optical Transport 886 System will occupy fibers in a sequence of fiber cables. Each fiber 887 cable will be placed in a sequence of conduits - buried honeycomb 888 structures through which fiber cables may be pulled - or buried in a 889 right of way (ROW). A ROW is land in which the network operator has 890 the right to install his conduit or fiber cable. It is worth noting 891 that for economic reasons, ROW's are frequently obtained from 892 railroads, pipeline companies, or thruways. It is frequently the 893 case that several carriers may lease ROW from the same source; this 894 makes it common to have a number of carriers' fiber cables in close 895 proximity to each other. Similarly, in a metropolitan network, 896 On Optical Layer Routing 898 several carriers might be leasing duct space in the same RBOC 899 conduit. There are also "carrier's carriers" - optical networks 900 which provide fibers to multiple carriers, all of whom could be 901 affected by a single failure in the "carrier's carrier" network. 903 In a typical intercity facility network there might be on the order 904 of 100 offices that are candidates for OLXC's. To represent the 905 inter-office fiber network accurately a network with an order of 906 magnitude more nodes is required. In addition to Optical Amplifier 907 (OA) sites, these additional nodes include: 908 - Places where fiber cables enter/leave a conduit or right of way; 909 - Locations where fiber cables cross; 910 Locations where fiber splices are used to interchange fibers between 911 fiber cables. 912 An example of the first might be: 913 A B 914 A-------------B \ / 915 \ / 916 X-----Y 917 / \ 918 C-------------D / \ 919 C D 921 (a) Fiber Cable Topology (b) Right-Of-Way/Conduit Topology 923 Figure 6-1: Fiber Cable vs. ROW Topologies 925 Here the A-B fiber cable would be physically routed A-X-Y-B and the 926 C-D cable would be physically routed C-X-Y-D. This topology might 927 arise because of some physical bottleneck: X-Y might be the Lincoln 928 Tunnel, for example, or the Bay Bridge. 930 Fiber route crossing (the second case) is really a special case of 931 this, where X and Y coincide. In this case the crossing point may 932 not even be a manhole; the fiber routes might just be buried at 933 different depths. 935 Fiber splicing (the third case) often occurs when a major fiber 936 route passes near to a small office. To avoid the expense and 937 additional transmission loss only a small number of fibers are 938 spliced out of the major route into a smaller route going to the 939 small office. This might well occur in a manhole or hut. An 940 example is shown in Fig. 6-2(a), where A-X-B is the major route, X 941 the manhole, and C the smaller office. The actual fiber topology 942 On Optical Layer Routing 944 would then look like Fig. 6-2(b), where there would typically be 945 many more A-B fibers than A-C or C-B fibers, and where A-C and C-B 946 might have different numbers of fibers. (One of the latter might 947 even be missing.) 949 C C 950 | / \ 951 | / \ 952 | / \ 953 A------X------B A---------------B 955 (a) Fiber Cable Topology (b) Fiber Topology 957 Figure 6-2. Fiber Cable vs Fiber Topologies 959 The imminent deployment of ultra-long (>1000 km) Optical Transport 960 Systems introduces a further complexity: Two OTS's could interact a 961 number of times. To make up a hypothetical example: A New York - 962 Atlanta OTS and a Philadelphia - Orlando OTS might ride on the same 963 right of way for x miles in Maryland and then again for y miles in 964 Georgia. They might also cross at Raleigh or some other intermediate 965 node without sharing right of way. 967 Diversity is often equated to routing two lightpaths between a 968 single pair of points, or different pairs of points so that no 969 single route failure will disrupt them both. This is too simplistic, 970 for a number of reasons: 972 - A sophisticated client of an optical network will want to derive 973 diversity needs from his/her end customers' availability 974 requirements. These often lead to more complex diversity 975 requirements than simply providing diversity between two 976 lightpaths. For example, a common requirement is that no single 977 failure should isolate a node or nodes. If a node A has single 978 lightpaths to nodes B and C, this requires A-B and A-C to be 979 diverse. In real applications, a large data network with N 980 lightpaths between its routers might describe their needs in an 981 NxN matrix, where (i,j) defines whether lightpaths i and j must 982 be diverse. 984 - Two circuits that might be considered diverse for one 985 application might not be considered diverse for in another 986 situation. Diversity is usually thought of as a reaction to 987 interoffice route failures. High reliability applications may 988 On Optical Layer Routing 990 require other types of failures to be taken into account. Some 991 examples: 992 o Office Outages: Although less frequent than route failures, 993 fires, power outages, and floods do occur. Many network 994 managers require that diverse routes have no (intermediate) 995 nodes in common. In other cases an intermediate node might 996 be acceptable as long as there is power diversity within 997 the office. 998 o Shared Rings: Many applications are willing to allow 999 "diverse" circuits to share a SONET ring-protected link; 1000 presumably they would allow the same for optical layer 1001 rings. 1002 o Disasters: Earthquakes and floods can cause failures over 1003 an extended area. Defense Department circuits might need 1004 to be routed with nuclear damage radii taken into account. 1005 - Conversely, some networks may be willing to take somewhat larger 1006 risks. Taking route failures as an example: Such a network 1007 might be willing to consider two fiber cables in heavy duty 1008 concrete conduit as having a low enough chance of simultaneous 1009 failure to be considered "diverse". They might also be willing 1010 to view two fiber cables buried on opposite sides of a railroad 1011 track as being diverse because there is minimal danger of a 1012 single backhoe disrupting them both even though a bad train 1013 wreck might jeopardize them both. A network seeking N mutually 1014 diverse paths from an office with less than N diverse ROW's will 1015 need to live with some level of compromise in the immediate 1016 vicinity of the office. 1018 These considerations strongly suggest that the routing algorithm 1019 should be sensitive to the types of threat considered unacceptable 1020 by the requester. Note that the impairment constraints described in 1021 the previous section may eliminate some of the long circuitous 1022 routes sometimes needed to provide diversity. This would make it 1023 harder to find many diverse paths through an all-optical network 1024 than an opaque one. 1026 [Hj�lmt�sson00] introduced the term "Shared Risk Link Group" (SRLG) 1027 to describe the relationship between two non-diverse links. The 1028 above examples and discussion given at the start of this section 1029 suggests that an SRLG should be characterized by 2 parameters: 1030 - Type of Compromise: Examples would be shared fiber cable, shared 1031 conduit, shared ROW, shared optical ring, shared office without 1032 power sharing, etc.) 1033 On Optical Layer Routing 1035 - Extent of Compromise: For compromised outside plant, this would 1036 be the length of the sharing. 1037 A CSPF algorithm could then penalize a diversity compromise by an 1038 amount dependent on these two parameters. 1040 Two links could be related by many SRLG's (AT&T's experience 1041 indicates that a link may belong to over 100 SRLG's, each 1042 corresponding to a separate fiber group. Each SRLG might relate a 1043 single link to many other links. For the optical layer, similar 1044 situations can be expected where a link is an ultra-long OTS). 1046 The mapping between links and different types of SRLG's is in 1047 general defined by network operators based on the definition of each 1048 SRLG type. Since SRLG information is not yet ready to be 1049 discoverable by a network element and does not change dynamically, 1050 it need not be advertised with other resource availability 1051 information by network elements. It could be configured in some 1052 central database and be distributed to or retrieved by the nodes, or 1053 advertised by network elements at the topology discovery stage. 1055 6.2 Implications For Routing 1057 Dealing with diversity is an unavoidable requirement for routing in 1058 the optical layer. It requires dealing with constraints in the 1059 routing process but most importantly requires additional state 1060 information � the SRLG relationships and also the routings of any 1061 existing circuits from the new circuit is to be diverse � to be 1062 available to the routing process. 1064 At present SRLG information cannot be self-discovered. Indeed, in a 1065 large network it is very difficult to maintain accurate SRLG 1066 information. The problem becomes particularly daunting whenever 1067 multiple administrative domains are involved, for instance after the 1068 acquisition of one network by another, because there normally is a 1069 likelihood that there are diversity violations between the domains. 1070 It is very unlikely that diversity relationships between carriers 1071 will be known any time in the near future. 1073 Considerable variation in what different customers will mean by 1074 acceptable diversity should be anticipated. Consequently we suggest 1075 that an SRLG should be defined as follows: (i) It is a relationship 1076 between two or more links, and (ii) it is characterized by two 1077 parameters, the type of compromise (shared conduit, shared ROW, 1078 shared optical ring, etc.) and the extent of the compromise (e.g., 1079 the number of miles over which the compromise persisted). This will 1080 allow the SRLG's appropriate to a particular routing request to be 1081 easily identified. 1083 On Optical Layer Routing 1085 7. Security Considerations 1087 We are assuming OEO interfaces to the domain(s) covered by our 1088 discussion (see, e.g., Sec. 4.1 above). If this assumption were to 1089 be relaxed and externally generated optical signals allowed into the 1090 domain, network security issues would arise. Specifically, 1091 unauthorized usage in the form of signals at improper wavelengths or 1092 with power levels or impairments inconsistent with those assumed by 1093 the domain would be possible. With OEO interfaces, these types of 1094 layer one threats should be controllable. 1096 A key layer one security issue is resilience in the face of physical 1097 attack. Diversity, as describe in Sec. 6, is a part of the 1098 solution. However, it is ineffective if there is not sufficient 1099 spare capacity available to make the network whole after an attack. 1100 Several major related issues are: 1101 - Defining the threat: If, for example, an electro-magnetic 1102 interference (EMI) burst is an in-scope threat, then (in the 1103 terminology of Sec. 6) all of the links sufficiently close 1104 together to be disrupted by such a burst must be included in a 1105 single SRLG. Similarly for other threats: For each in-scope 1106 threat, SRLG's must be defined so that all links vulnerable to a 1107 single incident of the threat must be grouped together in a 1108 single SRLG. 1109 - Allocating responsibility for responding to a layer one failure 1110 between the various layers (especially the optical and IP 1111 layers): This must be clearly specified to avoid churning and 1112 unnecessary service interruptions. 1114 The whole proposed process depends on the integrity of the 1115 impairment characterization information (PMD parameters, etc.) and 1116 also the SRLG definitions. Security of this information, both when 1117 stored and when distributed, is essential. 1119 This document does not address control plane issues, and so control- 1120 plane security is out of scope. IPO control plane security 1121 considerations are discussed in [Rajagopalam02]. Security 1122 considerations for GMPLS, a likely control plane candidate, are 1123 discussed in [Mannie02]. 1125 8. Acknowledgments 1127 This document has benefited from discussions with Michael Eiselt, 1128 Jonathan Lang, Mark Shtaif, Jennifer Yates, Dongmei Wang, Guangzhi 1129 Li, Robert Doverspike, Albert Greenberg, Jim Maloney, John Jacob, 1130 Katie Hall, Diego Caviglia, D. Papadimitriou, O. Audouin, J. P. 1131 Faure, L. Noirie, and with our OIF colleagues. 1133 On Optical Layer Routing 1135 9. References 1137 9.1 Normative References 1139 [Goldstein94] Goldstein, E. L., Eskildsen, L., and Elrefaie, A. F., 1140 Performance Implications of Component Crosstalk in Transparent 1141 Lightwave Networks", IEEE Photonics Technology Letters, Vol.6, No.5, 1142 May 1994. 1144 [Hj�lmt�sson00] G�sli Hj�lmt�sson, Jennifer Yates, Sid Chaudhuri and 1145 Albert 1146 Greenberg, "Smart Routers - Simple Optics: An Architecture for the 1147 Optical Internet, IEEE/OSA Journal of Lightwave Technology, December 1148 2000,, Vo 18, Issue 12 , Dec. 2000 , pp. 1880 -1891. 1150 [ITU] ITU-T Doc. G.663, Optical Fibers and Amplifiers, Section 1151 II.4.1.2. 1153 [Kaminow97] Kaminow, I. P. and Koch, T. L., editors, Optical Fiber 1154 Telecommunications IIIA, Academic Press, 1997. 1156 [Mannie02] Mannie, E. (ed.), "Generalized Multi-Protocol Label 1157 Switching (GMPLS) Architecture", Interned Draft, draft-ietf-ccamp- 1158 gmpls-architecture-03.txt, August, 2002. 1160 [Rajagopalam02] Rajagopalam, B., et. al., "IP over Optical Networks: 1161 A Framework", Internet Draft, draft-ietf-ipo-framework-02.txt June, 1162 2002. 1164 [Strand01] J. Strand, A. Chiu, and R. Tkach, "Issues for Routing in 1165 the Optical Layer", IEEE Communications Magazine, Feb. 2001, vol. 39 1166 No. 2, pp. 81-88. 1168 [Strand01b] J. Strand, R. Doverspike, and G. Li, "Importance of 1169 Wavelength Conversion In An Optical Network", Optical Networks 1170 Magazine, May/June 2001, pp. 33-44. 1172 [Yates99] Yates, J. M., Rumsewicz, M. P. and Lacey, J. P. R., 1173 "Wavelength Converters in Dynamically-Reconfigurable WDM Networks", 1174 IEEE Communications Surveys, 2Q1999 (online at 1175 www.comsoc.org/pubs/surveys/2q99issue/yates.html). 1177 On Optical Layer Routing 1179 9.2 Informative References 1181 [Awduche99] Awduche, D. O., Rekhter, Y., Drake, J., and Coltun, R., 1182 "Multi-Protocol Lambda Switching: Combining MPLS Traffic Engineering 1183 Control With Optical Crossconnects", Work in Progress, draft- 1184 awduche-mpls-te-optical-01.txt. 1186 [Bra96] Bradner, S., "The Internet Standards Process -- Revision 3," 1187 BCP 9, RFC 2026, October 1996. 1189 [CBD00] Ceuppens, L., Blumenthal, D., Drake, J., Chrostowski, J., 1190 Edwards, W., "Performance Monitoring in Photonic Networks in Support 1191 of MPL(ambda)S", Internet draft, work in progress, March 2000. 1193 [Doverspike00] Doverspike, R. and Yates, J., "Challenges For MPLS in 1194 Optical Network Restoration", IEEE Communication Magazine, February, 1195 2001. 1197 [Gerstel 2000] O. Gorstel, "Optical Layer Signaling: How Much Is 1198 Really Needed?" IEEE Communications Magazine, vol. 38 no. 10, Oct. 1199 2000, pp. 154-160 1201 [KRB01a] Kompella, K., et.al., "IS-IS extensions in support of 1202 Generalized MPLS," Internet Draft, draft-ietf-gmpls- extensions- 1203 01.txt, work in progress, 2001. 1205 [KRB01b] Kompella, K., et. al., "OSPF extensions in support of 1206 Generalized MPLS," Internet draft, draft-ospf-generalized- mpls- 1207 00.txt, work in progress, March 2001. 1209 [Moy98] Moy, John T., OSPF: Anatomy of an Internet Routing Protocol, 1210 Addison-Wesley, 1998. 1212 [Passmore01] Passmore, D., "Managing Fatter Pipes," Business 1213 Communications Review, August 2001, pp. 20-21. 1215 [Ramaswami98] Ramaswami, R. and Sivarajan, K. N., Optical Networks: 1216 A Practical Perspective, Morgan Kaufmann Publishers, 1998. 1218 [Tkach98] Tkach, R., Goldstein, E., Nagel, J., and Strand, J., 1219 "Fundamental Limits of Optical Transparency", Optical Fiber 1220 Communication Conf., Feb. 1998, pp. 161-162. 1222 On Optical Layer Routing 1224 10. Contributing Authors 1226 This document was a collective work of a number of people. The text 1227 and content of this document was contributed by the editors and the 1228 co-authors listed below. 1230 Ayan Banerjee 1231 Calient Networks 1232 5853 Rue Ferrari 1233 San Jose, CA 95138 1234 Email: abanerjee@calient.net 1236 Dan Blumenthal 1237 Calient Networks 1238 5853 Rue Ferrari 1239 San Jose, CA 95138 1240 Email: dblumenthal@calient.net 1242 John Drake 1243 Calient Networks 1244 5853 Rue Ferrari 1245 San Jose, CA 95138 1246 Email: jdrake@calient.net 1248 Andre Fredette 1249 Hatteras Networks 1250 PO Box 110025 1251 Research Triangle Park, NC 27709 1252 Email: afredette@hatterasnetworks.com 1254 Nan Froberg 1255 PhotonEx Corporation 1256 200 Metrowest Technology Dr. 1257 Maynard, MA 01754 1258 Email: nfroberg@photonex.com 1260 Taha Landolsi 1261 WorldCom, Inc. 1262 2400 North Glenville Drive 1263 Richardson, TX 75082 1264 Email: taha.landolsi@wcom.com 1266 James V. Luciani 1267 900 Chelmsford St. 1268 Lowell, MA 01851 1269 Email: james_luciani@mindspring.com 1271 Robert Tkach 1272 On Optical Layer Routing 1274 Celion Networks 1275 1 Sheila Dr., Suite 2 1276 Tinton Falls, NJ 07724 1277 Email: bob.tkach@celion.com 1279 Yong Xue 1280 WorldCom, Inc. 1281 22001 Loudoun County Parkway 1282 Ashburn, VA 20147 1283 Email: yxue@cox.com 1285 11. Editors' Addresses 1287 Angela Chiu 1288 AT&T Labs 1289 200 Laurel Ave., Rm A5-1F13 1290 Middletown, NJ 07748 1291 Phone:(732) 420-9061 1292 Email: chiu@research.att.com 1294 John Strand 1295 AT&T Labs 1296 200 Laurel Ave., Rm A5-1D33 1297 Middletown, NJ 07748 1298 Phone:(732) 420-9036 1299 Email: jls@research.att.com