idnits 2.17.1 draft-ietf-ipo-impairments-00.txt: -(170): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(188): 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 -(217): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(220): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(228): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(480): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(618): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(766): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(778): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(894): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(941): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(942): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(965): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1043): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1044): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1049): 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 23 instances of lines with non-ascii characters in the document. == No 'Intended status' indicated for this document; assuming Proposed Standard == It seems as if not all pages are separated by form feeds - found 0 form feeds but 23 pages Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an Introduction section. ** The document seems to lack a Security Considerations section. ** 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 403 instances of too long lines in the document, the longest one being 4 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == Line 454 has weird spacing: '... margin in OS...' -- 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.) -- Couldn't find a document date in the document -- date freshness check skipped. 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 section? 'GMPLS' on line 61 looks like a reference -- Missing reference section? 'Gerstel2000' on line 82 looks like a reference -- Missing reference section? 'Strand2001' on line 82 looks like a reference -- Missing reference section? 'Awduche99' on line 990 looks like a reference -- Missing reference section? 'Tkach98' on line 1056 looks like a reference -- Missing reference section? 'Yates99' on line 1060 looks like a reference -- Missing reference section? 'Ramaswami98' on line 1040 looks like a reference -- Missing reference section? 'ITU' on line 1023 looks like a reference -- Missing reference section? 'Strand01' on line 1043 looks like a reference -- Missing reference section? 'Kaminow97' on line 1026 looks like a reference -- Missing reference section? 'Goldstein94' on line 1018 looks like a reference -- Missing reference section? 'Chaudhuri00' on line 1003 looks like a reference -- Missing reference section? 'Strand01b' on line 1049 looks like a reference -- Missing reference section? 'ABB01' on line 982 looks like a reference -- Missing reference section? 'Ashwood00' on line 986 looks like a reference -- Missing reference section? 'Bra96' on line 995 looks like a reference -- Missing reference section? 'CBD00' on line 998 looks like a reference -- Missing reference section? 'Doverspike00' on line 1010 looks like a reference -- Missing reference section? 'Gerstel 2000' on line 1014 looks like a reference -- Missing reference section? 'KRB01b' on line 1033 looks like a reference -- Missing reference section? 'Moy98' on line 1037 looks like a reference Summary: 9 errors (**), 0 flaws (~~), 4 warnings (==), 23 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Draft 3 Document: draft-ietf-ipo-impairments-00.txt Angela Chiu 4 Expiration Date: November 2001 Robert Tkach 5 Celion Networks 7 James Luciani 8 Crescent Networks 10 Ayan Banerjee 11 John Drake 12 Dan Blumenthal 13 Calient Networks 15 Andre Fredette 16 Nan Froberg 17 PhotonEx 19 John Strand (Editor) 20 AT&T 22 Impairments And Other Constraints On Optical Layer Routing 24 Status of this Memo 26 This document is an Internet-Draft and is in full conformance with 27 all provisions of Section 10 of RFC2026. Internet-Drafts are 28 Working documents of the Internet Engineering Task Force (IETF), its 29 areas, and its working groups. Note that other groups may also 30 distribute working documents as Internet-Drafts. 32 Internet-Drafts are draft documents valid for a maximum of six 33 months and may be updated, replaced, or made obsolete by other 34 documents at any time. It is inappropriate to use Internet-Drafts as 35 reference material or to cite them other than as "work in progress." 37 The list of current Internet-Drafts can be accessed at 38 http://www.ietf.org/ietf/1id-abstracts.txt. 39 The list of Internet-Draft Shadow Directories can be accessed at 40 http://www.ietf.org/shadow.html. 42 Abstract 43 Optical networking poses a number challenges for GMPLS. Optical 44 technology is fundamentally an analog rather than digital technology; 46 Impairments And Other Constraints May 2001 47 On Optical Layer Routing 49 and the optical layer is lowest in the transport hierarchy and hence 50 has an intimate relationship with the physical geography of the 51 network. This contribution surveys some of the aspects of optical 52 networks which impact routing and identifies possible GMPLS responses 53 for each: (1) Constraints arising from the design of new software 54 controllable network elements, (2)Constraints in a single all-optical 55 domain without wavelength conversion (3)Complications arisin in more 56 complex networks incorporating both all-optical and opaque 57 architectures, and (5)Impacts of diversity constraints. 59 1. Introduction 61 GMPLS [GMPLS] aims to extend MPLS to encompass a number of transport 62 architectures. Included are optical networks incorporating a number 63 of all-optical and opto-electronic elements such as optical cross- 64 connects with both optical and electrical fabrics, transponders, and 65 optical add-drop multiplexers. Optical networking poses a number 66 challenges for GMPLS. Optical technology is fundamentally an analog 67 rather than digital technology; and the optical layer is lowest in 68 the transport hierarchy and hence has an intimate relationship with 69 the physical geography of the network. 71 GMPLS already has incorporated extensions to deal with some of the 72 unique aspects of the optical layer. This contribution surveys some 73 of the aspects of optical networks which impact routing and 74 identifies possible GMPLS responses for each. Routing constraints 75 and/or complications arising from the design of network elements, 76 the accumulation of signal impairments, and from the need to 77 guarantee the physical diversity of some circuits are discussed. 79 Since the purpose of this draft is to further the specification of 80 GMPLS, alternative approaches to controlling an optical network are 81 not discussed. For discussions of some broader issues, see 82 [Gerstel2000] and [Strand2001]. 84 The organization of the contribution is as follows: 86 - Section 2 is a section requested by the sub-IP Area 87 management for all new drafts. It explains how this document 88 fits into the Area and into the IPO WG, and why it is 89 appropriate for these groups. 90 - Section 3 describes constraints arising from the design of 91 new software controllable network elements. 92 - Section 4 addresses the constraints in a single all-optical 93 domain without wavelength conversion. 95 Impairments And Other Constraints May 2001 96 On Optical Layer Routing 98 - Section 5 extends the discussion to more complex networks 99 incorporating both all-optical and opaque architectures. 100 - Section 6 discusses the impacts of diversity constraints. 101 - Section 7 deals with security requirements. 102 - Section 8 contains acknowledgments. 104 2. Sub-IP Area Summary And Justification Of Work 105 This draft merges and extends two previous drafts, draft-chiu- 106 strand-unique-olcp-02.txt and draft-banerjee-routing-impairments- 107 00.txt. These two drafts were made IPO working group documents to 108 form a basis for a design team at the Minneapolis meeting, where it 109 was also requested that they be merged to create a requirements 110 document for the WG. 112 In the larger sub-IP Area structure, this merged document describes 113 specific characteristics of optical technology and the requirements 114 they place on routing and path selection. It is appropriate for the 115 IPO working group because the material is specific to optical 116 networks. It identifies and documents the characteristics of the 117 optical transport network that are important for selecting paths for 118 optical channels, which is a work area for the IPO WG. It is 119 appropriate work for this WG because the material covered is 120 directly aimed at establishing a framework and requirements for 121 routing in an optical network. 123 Related documents are: 124 draft-banerjee-routing-impairments-00.txt 125 draft-parent-obgp-01.txt 126 draft-bernstein-optical-bgp-00.txt 127 draft-hayata-ipo-carrier-needs-00.txt 128 draft-many-carrier-framework-uni-01.txt 130 3. Reconfigurable Network Elements 132 3.1 Technology Background 134 Control plane architectural discussions (e.g., [Awduche99]) usually 135 assume that the only software reconfigurable network element is an 136 optical layer cross-connect (OLXC). There are however other 137 software reconfigurable elements on the horizon, specifically 138 tunable lasers and receivers and reconfigurable optical add-drop 139 multiplexers (OADM�s). These elements are illustrated in the 140 following simple example, which is modeled on announced Optical 141 Transport System (OTS) products: 143 Impairments And Other Constraints May 2001 144 On Optical Layer Routing 146 + + 147 ---+---+ |\ /| +---+--- 148 ---| A |----|D| X Y |D|----| A |--- 149 ---+---+ |W| +--------+ +--------+ |W| +---+--- 150 : |D|-----| OADM |-----| OADM |-----|D| : 151 ---+---+ |M| +--------+ +--------+ |M| +---+--- 152 ---| A |----| | | | | | | |----| A |--- 153 ---+---+ |/ | | | | \| +---+--- 154 + +---+ +---+ +---+ +---+ + 155 D | A | | A | | A | | A | E 156 +---+ +---+ +---+ +---+ 157 | | | | | | | | 159 Figure 3-1: An OTS With OADM's - Functional Architecture 161 In Fig.3-1, the part that is on the inner side of all boxes labeled 162 "A" defines an all-optical subnetwork. From a routing perspective 163 two aspects are critical: 164 - Adaptation: These are the functions done at the edges of the 165 subnetwork that transform the incoming optical channel into the 166 physical wavelength to be transported through the subnetwork. 167 - Connectivity: This defines which pairs of edge Adaptation 168 functions can be interconnected through the subnetwork. 170 In Fig. 3-1, D and E are DWDM�s and X and Y are OADM�s. The boxes 171 labeled "A" are adaptation functions. They map one or more input 172 optical channels assumed to be standard short reach signals into a 173 long reach (LR) wavelength or wavelength group which will pass 174 transparently to a distant adaptation function. Adaptation 175 functionality which affects routing includes: 176 - Multiplexing: Either electrical or optical TDM may be used to 177 combine the input channels into a single wavelength. This is 178 done to increase effective capacity: A typical DWDM might be 179 able to handle 100 2.5 Gb/sec signals (250 Gb/sec total) or 50 180 10 Gb/sec (500 Gb/sec total); combining the 2.5 Gb/sec signals 181 together thus effectively doubles capacity. After multiplexing 182 the combined signal must be routed as a group to the distant 183 adaptation function. 184 - Adaptation Grouping: In this technique, groups of k (e.g., 4) 185 wavelengths are managed as a group within the system and must be 186 added/dropped as a group. We will call such a group an 187 "adaptation grouping". Other terms frequently used are "wave 188 group" and �waveband�. Groupings on the same system may differ 189 in basics such as wavelength spacing, which constrain the type 190 of channels which they can accommodate. 191 - Laser Tunability: The lasers producing the LR wavelengths may 192 have a fixed frequency, may be tunable over a limited range, or 194 Impairments And Other Constraints May 2001 195 On Optical Layer Routing 197 be tunable over the entire range of wavelengths supported by the 198 DWDM. Tunability speeds may also vary. 200 Connectivity between adaptation functions may also be limited: 201 - As pointed out above, TDM multiplexing and/or adaptation 202 grouping by the adaptation function forces groups of input 203 channels to be delivered together to the same distant adaptation 204 function. 205 - Only adaptation functions whose lasers/receivers are tunable to 206 compatible frequencies can be connected. 207 - The switching capability of the OADM�s may also be constrained. 208 For example: 209 o There may be some wavelengths that can not be dropped at 210 all. 211 o There may be a fixed relationship between the frequency 212 dropped and the physical port on the OADM to which it is 213 dropped. 214 o OADM physical design may put an upper bound on the number 215 of adaptation groupings dropped at any single OADM. 217 For a fixed configuration of the OADM�s and adaptation functions 218 connectivity will be fixed: Each input port will essentially be 219 hard-wired to some specific distant port. However this connectivity 220 can be changed by changing the configurations of the OADM�s and 221 adaptation functions. For example, an additional adaptation grouping 222 might be dropped at an OADM or a tunable laser retuned. In each case 223 the port-to-port connectivity is changed. 225 These capabilities can be expected to be under software control. 226 Today the control would rest in the vendor-supplied Element 227 Management system (EMS), which in turn would be controlled by the 228 operator�s OS�s. However in principle the EMS could participate in 229 the GMPLS routing process. 231 3.2 Implications For Routing 233 An OTS of the sort discussed in Sec. 3.1 is essentially a 234 geographically distributed but blocking cross-connect system. The 235 specific port connectivity is dependent on the vendor design and 236 also on exactly what line cards have been deployed. 238 One way for GMPLS to deal with this architecture would be to view 239 the port connectivity as externally determined. In this case the 240 links known to GMPLS would be groups of identically routed 241 wavebands. If these were reconfigured by the external EMS the 242 resulting connectivity changes would need to be detected and 243 advertised within GMPLS. If the topology shown in Fig. 2-1 became a 244 tree or a mesh instead of the linear topology shown, the 245 connectivity changes could result in SRLG changes. 247 Impairments And Other Constraints May 2001 248 On Optical Layer Routing 250 Alternatively, GMPLS could attempt to directly control this port 251 connectivity. The state information needed to do this is likely to 252 be voluminous and vendor specific. 254 4. Wavelength Routed All-Optical Networks 256 The optical networks presently being deployed may be called "opaque" 257 ([Tkach98]): each link is optically isolated by transponders doing 258 O/E/O conversions. They provide regeneration with retiming and 259 reshaping, also called 3R, which eliminates transparency to bit 260 rates and frame format. These transponders are quite expensive and 261 their lack of transparency also constrains the rapid introduction of 262 new services. Thus there are strong motivators to introduce 263 "domains of transparency" - all-optical subnetworks - larger than an 264 OTS. 266 The routing of lightpaths through an all-optical network has 267 received extensive attention. (See [Yates99] or [Ramaswami98]). 268 When discussing routing in an all-optical network it is usually 269 assumed that all routes have adequate signal quality. This may be 270 ensured by limiting all-optical networks to subnetworks of limited 271 geographic size which are optically isolated from other parts of the 272 optical layer by transponders. This approach is very practical and 273 has been applied to date, e.g. when determining the maximum length 274 of an Optical Transport System (OTS). Furthermore operational 275 considerations like fault isolation also make limiting the size of 276 domains of transparency attractive. 278 There are however reasons to consider contained domains of 279 transparency in which not all routes have adequate signal quality. 280 From a demand perspective, maximum bit rates have rapidly increased 281 from DS3 to OC-192 and soon OC-768 (40 Gb/sec). As bit rates 282 increase it is necessary to increase power. This makes impairments 283 and nonlinearities more troublesome. From a supply perspective, 284 optical technology is advancing very rapidly, making ever-larger 285 domains possible. In this section we assume that these 286 considerations will lead to the deployment of a domain of 287 transparency that is too large to ensure that all potential routes 288 have adequate signal quality for all circuits. Our goal is to 289 understand the impacts of the various types of impairments in this 290 environment. 292 4.1 Problem Formulation 294 We consider a single domain of transparency without wavelength 295 translation. Additionally due to proprietary natures of DWDM 296 transmission technology, we assume that the domain is either single 297 vendor or architected using a single coherent design philosophy, 298 particularly with regard to the management of impairments. 300 Impairments And Other Constraints May 2001 301 On Optical Layer Routing 303 We wish to route a unidirectional circuit from ingress client node X 304 to egress client node Y. At both X and Y, the circuit goes through 305 an O/E/O conversion which optically isolates the portion within our 306 domain. We assume that we know the bit rate of the circuit. Also, 307 we assume that the adaptation function at X may apply some Forward 308 Error Correction (FEC) method to the circuit. We also assume we know 309 the launch power of the laser at X. 311 Impairments can be classified into two categories, linear and 312 nonlinear. (See [Tkach98] for more on impairment constraints). 313 Linear effects are independent of signal power and affect 314 wavelengths individually. Amplifier spontaneous emission (ASE), 315 polarization mode dispersion (PMD), and chromatic dispersion are 316 examples. Nonlinearities are significantly more complex: they 317 generate not only impairments on each channel, but also crosstalk 318 between channels. 320 In the remainder of this section we first outline how two key linear 321 impairments (PMD and ASE) might be handled by a set of analytical 322 formulae as additional constraints on routing. We next discuss how 323 the remaining constraints might be approached. Finally we take a 324 broader perspective and discuss the implications of such constraints 325 on control plane architecture and also on broader constrained domain 326 of transparency architecture issues. 328 4.2 Polarization Mode Dispersion 330 For a transparent fiber segment, the general rule for the PMD 331 requirement is that the time-average differential group delay (DGD) 332 between two orthogonal state of polarizations should be less than 333 fraction a of the bit duration, T=1/B, where B is the bit rate. The 334 value of the parameter a depends on three major factors: 1) margin 335 allocated to PMD, e.g. 1dB; 2) targeted outage probability, e.g. 336 4x10-5, and 3) sensitivity of the receiver to DGD. A typical value 337 for a is 10% [ITU]. More aggressive designs to compensate for PMD 338 may allow values higher than 10%. (This would be a system parameter 339 dependent on the system design. It would need to be known to the 340 routing process.) 342 The PMD parameter (Dpmd) is measured in pico-seconds (ps) per 343 sqrt(km). The square of the PMD in a fiber span, denoted as span- 344 PMD-square is then given by the product of Dpmd**2 and the span 345 length. (A fiber span in a transparent network refers to a segment 346 between two optical amplifiers.) If Dpmd is constant, this results 347 in a upper bound on the maximum length of an M-fiber-span 348 transparent segment, which is inversely proportional to the square 349 of the product of bit rate and Dpmd.(The detailed equation is 350 omitted due to the format constraint - see [Strand01] for details). 352 Impairments And Other Constraints May 2001 353 On Optical Layer Routing 355 For older fibers with a typical PMD parameter of 0.5 picoseconds per 356 square root of km, based on the constraint, the maximum length of 357 the transparent segment should not exceed 400km and 25km for bit 358 rates of 10Gb/s and 40Gb/s, respectively. Due to recent advances in 359 fiber technology, the PMD-limited distance has increased 360 dramatically. For newer fibers with a PMD parameter of 0.1 361 picosecond per square root of km, the maximum length of the 362 transparent segment (without PMD compensation) is limited to 10000km 363 and 625km for bit rates of 10Gb/s and 40Gb/, respectively. Still 364 lower values of PMD are attainable in commercially available fiber 365 today, and the PMD limit can be further extended if a larger value 366 of the parameter a (ratio of DGD to the bit period) can be 367 tolerated. In general, the PMD requirement is not an issue for most 368 types of fibers at 10Gb/s or lower bit rate. But it will become an 369 issue at bit rates of 40Gb/s and higher. 371 If the PMD parameter varies between spans, a slightly more 372 complicated equations results (see [Strand01]), but in any event the 373 only link dependent information needed by the routing algorithm is 374 the square of the link PMD, denoted as link-PMD-square. It is the 375 sum of the span-PMD-square of all spans on the link. 377 4.3 Amplifier Spontaneous Emission 379 ASE degrades the optical signal to noise ratio (OSNR). An acceptable 380 optical SNR level (SNRmin) which depends on the bit rate, 381 transmitter-receiver technology (e.g., FEC), and margins allocated 382 for the impairments, needs to be maintained at the receiver. In 383 order to satisfy this requirement, vendors often provide some 384 general engineering rule in terms of maximum length of the 385 transparent segment and number of spans. For example, current 386 transmission systems are often limited to up to 6 spans each 80km 387 long. For larger transparent domains, more detailed OSNR 388 computations will be needed to determine whether the OSNR level 389 through a domain of transparency is acceptable. This would provide 390 flexibility in provisioning or restoring a lightpath through a 391 transparent subnetwork. 393 Assume that the average optical power launched at the transmitter is 394 P. The lightpath from the transmitter to the receiver goes through M 395 optical amplifiers, with each introducing some noise power. Unity 396 gain can be used at all amplifier sites to maintain constant signal 397 power at the input of each span to minimize noise power and 398 nonlinearity. A constraint on the maximum number of spans can be 399 obtained [Kaminow97] which is proportional to P and inversely 400 proportional to SNRmin, optical bandwidth B, amplifier gain G-1 and 401 spontaneous emission factor n of the optical amplifier, assuming all 402 spans have identical gain and noise figure. (Again, the detailed 403 equation is omitted due to the format constraint - see [Strand01] 404 for details.) Let�s take a typical example. Assuming P=4dBm, 406 Impairments And Other Constraints May 2001 407 On Optical Layer Routing 409 SNRmin=20dB with FEC, B=12.5GHz, n=2.5, G=25dB, based on the 410 constraint, the maximum number of spans is at most 10. However, if 411 FEC is not used and the requirement on SNRmin becomes 25dB, the 412 maximum number of spans drops down to 3. 414 For ASE the only link-dependent information needed by the routing 415 algorithm is the noise of the link, denoted as link-noise, which is 416 the sum of the noise of all spans on the link. Hence the constraint 417 on ASE becomes that the aggregate noise of the transparent segment 418 which is the sum of the link-noise of all links can not exceed 419 P/SNRmin. 421 4.4 Approximating The Effects Of Some Other Impairment Constraints 423 There are a number of other impairment constraints that we propose 424 to approximate with a domain-wide margin on the OSNR, plus some 425 constraint on the total number of networking elements (OXC or OADM) 426 in some cases. Most impairments generated at OXCs or OADMs, 427 including polarization dependent loss, coherent crosstalk, and 428 effective passband width, are dealt with using this approach. In 429 principle, impairments generated at the nodes can be bounded by 430 system engineering rules because the node elements can be designed 431 and specified in a uniform manner. This approach is not feasible 432 with PMD and noise because neither can be uniformly specified. 433 Instead, they depend on node spacing and the characteristics of the 434 installed fiber plant, neither of which are likely to be under the 435 system designer�s control. 437 Examples of the constraints we propose to approximate with a domain- 438 wide margin are given in the remaining paragraphs in this section. 439 It should be kept in mind that as optical transport technology 440 evolves it may become necessary to include some of these impairments 441 explicitly in the routing process. Other impairments not mentioned 442 here at all may also become sufficiently important to require 443 incorporation either explicitly or via a domain-wide margin. 445 Other Polarization Dependent Impairments Other polarization- 446 dependent effects besides PMD influence system performance. For 447 example, many components have polarization-dependent loss (PDL) 448 [Ramaswami98] which accumulates in a system with many components on 449 the transmission path. The state of polarization fluctuates with 450 time and its distribution is very important also. It is generally 451 required to maintain the total PDL on the path to be within some 452 acceptable limit, potentially by using some compensation technology 453 for relatively long transmission systems, plus a small built-in 454 margin in OSNR. Since the total PDL increases with the number of 455 components in the data path, it must be taken into account by the 456 system vendor when determining the maximum allowable number of 457 spans. 459 Impairments And Other Constraints May 2001 460 On Optical Layer Routing 462 Chromatic Dispersion In general this impairment can be adequately 463 (but not optimally) compensated for on a per-link basis, and/or at 464 system initial setup time. A low margin in OSNR can be put in to 465 account for any mismatch in dispersion compensation. For high bit 466 rates, dynamic dispersion compensation may be required at the 467 receiver to clean up any residual dispersion. 469 Crosstalk Optical crosstalk refers to the effect of other signals on 470 the desired signal. It includes both coherent (i.e. intrachannel) 471 crosstalk and incoherent (i.e. interchannel) crosstalk. Main 472 contributors of crosstalk are the OADM and OXC sites that use a DWDM 473 multiplexer/demultiplexer (MUX/DEMUX) pair. For a relatively sparse 474 network where the number of OADM/OXC nodes on a path is low, 475 crosstalk can be treated with a low margin in OSNR without being a 476 binding constraint. But for some relatively dense networks where 477 crosstalk might become a binding constraint, one needs to propagate 478 the per-link crosstalk information to make sure that the end-to-end 479 crosstalk which is the sum of the crosstalks on all the 480 corresponding links to be within some limit, e.g. �25dB threshold 481 with 1dB penalty ([Goldstein94]). Another way to treat it without 482 having to propagate per-link crosstalk information is to have the 483 system evaluate what the maximum number of OADM/OXC nodes that has a 484 MUX/DEMUX pair for the worst route in the transparent domain for a 485 low built-in margin. The latter one should work well where all the 486 OXC/OADM nodes have similar level of crosstalk. 488 Effective Passband As more and more DWDM components are cascaded, 489 the effective passband narrows. The number of filters along the 490 link, their passband width and their shape will determine the end- 491 to-end effective passband. In general, this is a system design 492 issue, i.e., the system is designed with certain maximum bit rate 493 using the proper modulation format. Then the filter effect can be 494 turned into a constraint on the maximum number of OADM/OXC with the 495 condition that filters in the systems are at least as wide as the 496 one in the receiver. Because traffic at lower bit rates can 497 tolerate a narrower passband, the maximum allowable number of 498 OADMs/OXCs will increase as the bit rate decreases. 500 Nonlinear Impairments It seems unlikely that these can be dealt with 501 explicitly in a routing algorithm because they lead to constraints 502 that can couple routes together and lead to complex dependencies, 503 e.g. on the order in which specific fiber types are traversed. Note 504 that different fiber types (standard single mode fiber, dispersion 505 shifted fiber, dispersion compensated fiber, etc.) have very 506 different effects from nonlinear impairments. A full treatment of 507 the nonlinear constraints would likely require very detailed 508 knowledge of the physical infrastructure, including measured 509 dispersion values for each span, fiber core area and composition, as 510 well as knowledge of subsystem details such as dispersion 511 compensation technology. This information would need to be combined 513 Impairments And Other Constraints May 2001 514 On Optical Layer Routing 516 with knowledge of the current loading of optical signals on the 517 links of interest to determine the level of nonlinear impairment. 518 Alternatively, one could assume that nonlinear impairments are 519 bounded and result in X dB margin in the required OSNR level for a 520 given bit rate, where X for performance reasons would be limited to 521 1 or 2 dB, consequently setting a limit on the maximum number of 522 spans. For the approach described here to be useful, it is desirable 523 for this span length limit to be longer than that imposed by the 524 constraints which can be treated explicitly. When designing a DWDM 525 transport system, there are tradeoffs between signal power launched 526 at the transmitter, span length, and nonlinear effects on BER that 527 need to be considered jointly. Here, we assume that an X dB margin 528 is obtained after the transport system has been designed with a 529 fixed signal power and maximum span length for a given bit rate. 530 Further work is required to determine the validity of this approach. 531 However, it is likely that there is an advantage in designing 532 systems which are less aggressive with respect to nonlinearities, 533 and therefore somewhat sub-optimal, in exchange for improved 534 scalability, simplicity and flexibility in routing and control plane 535 design. 537 4.5 Other Impairment Considerations 539 There are many other types of impairments which can degrade 540 performance. In this section we briefly mention one other type of 541 impairment, which we propose be dealt with by either by the system 542 designer or by the transmission engineers at the time the system is 543 installed. If dealt with successfully in this manner they should not 544 need to be considered in the dynamic routing process. 546 Gain Nonuniformity and Gain Transients For simple noise estimates to 547 be of use, the amplifiers must be gain-flattened and must have 548 automatic gain control (AGC). Furthermore, each link should have 549 dynamic gain equalization (DGE) to optimize power levels each time 550 wavelengths are added or dropped. Variable optical attenuators on 551 the output ports of an OXC or OADM can be used for this purpose, and 552 in-line devices are starting to become commercially available. 553 Optical channel monitors are also required to provide feedback to 554 the DGEs. AGC must be done rapidly if signal degradation after a 555 protection switch or link failure is to be avoided. 557 4.6 Other Considerations 559 Routing in an all-optical network without wavelength conversion 560 raises several additional issues: 562 Impairments And Other Constraints May 2001 563 On Optical Layer Routing 565 - Since the route selected must have the chosen wavelength 566 available on all links, this information needs to be considered 567 in the routing process. This is discussed in [Chaudhuri00], where 568 it is concluded that advertising detailed wavelength 569 availabilities on each link is not likely to scale. Instead they 570 propose an alternative method which probes along a chosen path to 571 determine which wavelengths (if any) are available. This would 572 require a significant addition to the routing logic normally used 573 in OSPF. Others have proposed simultaneously probing along 574 multiple paths. 576 - Choosing a path first and then a wavelength along the path is 577 known to give adequate results in simple topologies such as rings 578 and trees ([Yates99]). This does not appear to be true in large 579 mesh networks under realistic provisioning scenarios, however 580 ([Strand01b]). Instead significantly better results are achieved 581 if wavelength and route are chosen simultaneously. This approach 582 would however also have a significant affect on OSPF. 584 4.7 Implications For Routing and Control Plane Design 586 Additional state information will be required by the routing to 587 explicitly deal with the impairments described in Sections 4.2 - 588 4.4: 590 - Additional link-specific state information will be required by the 591 routing algorithm for each type of impairment that has the 592 potential of being limiting for some routes. The link-dependent 593 information identified above for PMD is link-PMD-square which is 594 the square of the total PMD on a link. For ASE the link-dependent 595 information identified is link-noise which is the total noise on a 596 link. Other link-dependent information includes link-span-length 597 which is the total number of spans on a link, and OADM-OXC-number 598 which is the number of OADM/OXC on a link. 600 . In addition to the link-specific information, bounds on each of 601 the impairments need to be quantified. Since these bounds are 602 determined by the system designer's impairment allocations, these 603 will be system dependent. For PMD, the constraint is that the sum 604 of the link-PMD-square of all links on the transparent segment is 605 less than the square of (a/B) where B is the bit rate. Hence, the 606 required information is the parameter "a". For ASE, the constraint 607 is that the sum of the link-noise of all links is no larger than 608 P/SNRmin. Thus, the information needed include the launch power P 609 and OSNR requirement SNRmin. The minimum acceptable OSNR, in 610 turn, depends on the strength of the FEC being used and the 611 margins reserved for other types of impairments. Other bounds 612 include the maximum span length of the transmission system and the 613 maximum number of OADM/OXC nodes, both of which are bit rate 615 Impairments And Other Constraints May 2001 616 On Optical Layer Routing 618 dependent. In single-vendor �islands� some of these parameters may 619 be available in a local or EMS database and would not need to be 620 advertised 622 . It is likely that the physical layer parameters do not change 623 value rapidly and could be stored in some database; however these 624 are physical layer parameters that today are frequently not known 625 at the granularity required. If the ingress node of a lightpath 626 does path selection these parameters would need to be available at 627 this node. 629 . The specific constraints required in a given situation will depend 630 on the design and engineering of the domain of transparency; for 631 example it will be essential to know whether chromatic dispersion 632 has been dealt with on per-link basis, and whether the domain is 633 operating in a linear or nonlinear regime. 635 . As optical transport technology evolves, the set of constraints 636 that will need to be considered either explicitly or via a domain- 637 wide margin may change. The routing and control plane design 638 should therefore be as open as possible, allowing parameters to be 639 included as necessary. 641 . In the absence of wavelength conversion, the necessity of finding 642 a single wavelength that is available on all links introduces the 643 need to either advertise detailed information on wavelength 644 availability, which probably doesn't scale, or have some mechanism 645 for probing potential routes with or without crankback to 646 determine wavelength availability. Choosing the route first, and 647 then the wavelength, may not yield acceptable utilization levels 648 in mesh-type networks. 650 5. More Complex Networks 652 Mixing optical equipment in a single domain of transparency that has 653 not been explicitly designed to interwork is beyond the scope of 654 this document. This includes most multi-vendor all-optical networks. 656 An all-optical network composed of multiple domains of transparency 657 optically isolated from each other by OEO devices (transponders) is 658 more plausible. A network composed of both "opaque" (optically 659 isolated) OLXC's and one or more all-optical "islands" isolated by 660 transponders is of particular interest because this is most likely 661 how all-optical technologies (such as that described in Sec. 2) are 662 going to be introduced. (We use the term "island" in this discussion 663 rather than a term like "domain" or "area" because these terms are 664 associated with specific approaches like BGP or OSPF.) 666 Impairments And Other Constraints May 2001 667 On Optical Layer Routing 669 We consider the complexities raised by these alternatives now. 671 The first requirement for routing in a multi-island network is that 672 the routing process needs to know the extent of each island. There 673 are several reasons for this: 674 . When entering or leaving an all-optical island, the regeneration 675 process cleans up the optical impairments discussed in Sec. 3. 676 . Each all-optical island may have its own bounds on each 677 impairment. 678 . The routing process needs to be sensitive to the costs associated 679 with "island-hopping". 681 This last point needs elaboration. It is extremely important to 682 realize that, at least in the short to intermediate term, the 683 resources committed by a single routing decision can be very 684 significant: The equipment tied up by a single coast-to-coast OC-192 685 can easily have a first cost of $10**6, and the holding times on a 686 circuit once established is likely to be measured in months. 687 Carriers will expect the routing algorithms used to be sensitive to 688 these costs. Simplistic measures of cost such as the number of 689 "hops" are not likely to be acceptable. 691 Taking the case of an all-optical island consisting of an "ultra 692 long-haul" system like that in Fig. 2-1 embedded in an OEO network 693 of electrical fabric OLXC's as an example: It is likely that the ULH 694 system will be relatively expensive for short hops but relatively 695 economical for longer distances. It is therefore likely to be 696 deployed as a sort of "express backbone". In this scenario a carrier 697 is likely to expect the routing algorithm to balance OEO costs 698 against the additional costs associated with ULH technology and 699 route circuitously to make maximum use of the backbone where 700 appropriate. Note that the metrics used to do this must be 701 consistent throughout the routing domain if this expectation is to 702 be met. 704 The first-order implications for GMPLS seem to be: 705 . Information about island boundaries needs to be advertised. 706 . The routing algorithm needs to be sensitive to island transitions 707 and to the connectivity limitations and impairment constraints 708 particular to each island. 709 . The cost function used in routing must allow the balancing of 710 transponder costs, OXC and OADM costs, and line haul costs across 711 the entire routing domain. 713 Several distributed approaches to multi-island routing seem worth 714 investigating: 715 . Advertise the internal topology and constraints of each island 716 globally; let the ingress node compute an end-to-end strict 717 explicit route sensitive to all constraints and wavelength 718 availabilities. In this approach the routing algorithm used by the 720 Impairments And Other Constraints May 2001 721 On Optical Layer Routing 723 ingress node must be able to deal with the details of routing 724 within each island. 725 . Have the EMS or control plane of each island determine and 726 advertise the connectivity between its boundary nodes together 727 with additional information such as costs and the bit rates and 728 formats supported. As the spare capacity situation changes, 729 updates would be advertised. In this approach impairment 730 constraints are handled within each island and impairment-related 731 parameters need not be advertised outside of the island. The 732 ingress node would then do a loose explicit route and leave the 733 routing and wavelength selection within each island to the island. 734 . Have the ingress node send out probes or queries to nearby gateway 735 nodes or to an NMS to get routing guidance. 737 6. Diversity 739 6.1 Background On Diversity 741 "Diversity" is a relationship between lightpaths. Two lightpaths are 742 said to be diverse if they have no single point of failure. In 743 traditional telephony the dominant transport failure mode is a 744 failure in the interoffice plant, such as a fiber cut inflicted by a 745 backhoe. 747 Why is diversity a unique problem that needs to be considered for 748 optical networks? So far, data network operators have relied on 749 their private line providers to ensure diversity and so have not had 750 to deal directly with the problem. GMPLS makes the complexities 751 handled by the private line provisioning process, including 752 diversity, part of the common control plane and so visible to all. 754 To determine whether two lightpath routings are diverse it is 755 necessary to identify single points of failure in the interoffice 756 plant. To do so we will use the following terms: A fiber cable is a 757 uniform group of fibers contained in a sheath. An Optical Transport 758 System will occupy fibers in a sequence of fiber cables. Each fiber 759 cable will be placed in a sequence of conduits - buried honeycomb 760 structures through which fiber cables may be pulled - or buried in a 761 right of way (ROW). A ROW is land in which the network operator has 762 the right to install his conduit or fiber cable. It is worth noting 763 that for economic reasons, ROW�s are frequently obtained from 764 railroads, pipeline companies, or thruways. It is frequently the 765 case that several carriers may lease ROW from the same source; this 766 makes it common to have a number of carriers� fiber cables in close 767 proximity to each other. Similarly, in a metropolitan network, 768 several carriers might be leasing duct space in the same RBOC 770 Impairments And Other Constraints May 2001 771 On Optical Layer Routing 773 conduit. There are also "carrier's carriers" - optical networks 774 which provide fibers to multiple carriers, all of whom could be 775 affected by a single failure in the "carrier's carrier" network. 777 In a typical intercity facility network there might be on the order 778 of 100 offices that are candidates for OLXC�s. To represent the 779 inter-office fiber network accurately a network with an order of 780 magnitude more nodes is required. In addition to Optical Amplifier 781 (OA) sites, these additional nodes include: 782 - Places where fiber cables enter/leave a conduit or right of way; 783 - Locations where fiber cables cross; 784 - Locations where fiber splices are used to interchange fibers 785 between fiber cables. 787 An example of the first might be: 789 A B 790 A-------------B \ / 791 \ / 792 X-----Y 793 / \ 794 C-------------D / \ 795 C D 797 (a) Fiber Cable Topology (b) Right-Of-Way/Conduit Topology 799 Figure 6-1: Fiber Cable vs. ROW Topologies 801 Here the A-B fiber cable would be physically routed A-X-Y-B and the 802 C-D cable would be physically routed C-X-Y-D. This topology might 803 arise because of some physical bottleneck: X-Y might be the Lincoln 804 Tunnel, for example, or the Bay Bridge. 806 Fiber route crossing (the second case) is really a special case of 807 this, where X and Y coincide. In this case the crossing point may 808 not even be a manhole; the fiber routes might just be buried at 809 different depths. 811 Fiber splicing (the third case) often occurs when a major fiber 812 route passes near to a small office. To avoid the expense and 813 additional transmission loss only a small number of fibers are 814 spliced out of the major route into a smaller route going to the 815 small office. This might well occur in a manhole or hut. An 817 Impairments And Other Constraints May 2001 818 On Optical Layer Routing 820 example is shown in Fig. 6-2(a), where A-X-B is the major route, X 821 the manhole, and C the smaller office. The actual fiber topology 822 would then look like Fig. 6-2(b), where there would typically be 823 many more A-B fibers than A-C or C-B fibers, and where A-C and C-B 824 might have different numbers of fibers. (One of the latter might 825 even be missing.) 827 C C 828 | / \ 829 | / \ 830 | / \ 831 A------X------B A---------------B 833 (a) Fiber Cable Topology (b) Fiber Topology 835 Figure 6-2. Fiber Cable vs Fiber Topologies 837 The imminent deployment of ultra-long (>1000 km) Optical Transport 838 Systems introduces a further complexity: Two OTS's could interact a 839 number of times. To make up a hypothetical example: A New York - 840 Atlanta OTS and a Philadelphia - Orlando OTS might ride on the same 841 right of way for x miles in Maryland and then again for y miles in 842 Georgia. They might also cross at Raleigh or some other intermediate 843 node without sharing right of way. 845 Diversity is often equated to routing two lightpaths between a 846 single pair of points, or different pairs of points so that no 847 single route failure will disrupt them both. This is too simplistic, 848 for a number of reasons: 850 - A sophisticated client of an optical network will want to derive 851 diversity needs from his/her end customers' availability 852 requirements. These often lead to more complex diversity 853 requirements than simply providing diversity between two 854 lightpaths. For example, a common requirement is that no single 855 failure should isolate a node or nodes. If a node A has single 856 lightpaths to nodes B and C, this requires A-B and A-C to be 857 diverse. In real applications, a large data network with N 858 lightpaths between its routers might describe their needs in an 859 NxN matrix, where (i,j) defines whether lightpaths i and j must 860 be diverse. 862 - Two circuits that might be considered diverse for one 863 application might not be considered diverse for in another 865 Impairments And Other Constraints May 2001 866 On Optical Layer Routing 868 situation. Diversity is usually thought of as a reaction to 869 interoffice route failures. High reliability applications may 870 require other types of failures to be taken into account. Some 871 examples: 872 o Office Outages: Although less frequent than route failures, 873 fires, power outages, and floods do occur. Many network 874 managers require that diverse routes have no (intermediate) 875 nodes in common. In other cases an intermediate node might 876 be acceptable as long as there is power diversity within 877 the office. 878 o Shared Rings: Many applications are willing to allow 879 "diverse" circuits to share a SONET ring-protected link; 880 presumably they would allow the same for optical layer 881 rings. 882 o Disasters: Earthquakes and floods can cause failures over 883 an extended area. Defense Department circuits might need 884 to be routed with nuclear damage radii taken into account. 885 - Conversely, some networks may be willing to take somewhat larger 886 risks. Taking route failures as an example: Such a network 887 might be willing to consider two fiber cables in heavy duty 888 concrete conduit as having a low enough chance of simultaneous 889 failure to be considered "diverse". They might also be willing 890 to view two fiber cables buried on opposite sides of a railroad 891 track as being diverse because there is minimal danger of a 892 single backhoe disrupting them both even though a bad train 893 wreck might jeopardize them both. A network seeking N mutually 894 diverse paths from an office with less than N diverse ROW�s will 895 need to live with some level of compromise in the immediate 896 vicinity of the office. 898 These considerations strongly suggest that the routing algorithm 899 should be sensitive to the types of threat considered unacceptable 900 by the requester. Note that the impairment constraints described in 901 the previous section may eliminate some of the long circuitous 902 routes sometimes needed to provide diversity. This would make it 903 harder to find many diverse paths through an all-optical network 904 than an opaque one. 906 [Chaudhuri00] introduced the term "Shared Risk Link Group" (SRLG) to 907 describe the relationship between two non-diverse links. The above 908 discussion suggests that an SRLG should be characterized by 2 909 parameters: 911 Impairments And Other Constraints May 2001 912 On Optical Layer Routing 914 - Type of Compromise: Examples would be shared fiber cable, shared 915 conduit, shared ROW, shared optical ring, shared office without 916 power sharing, etc.) 917 - Extent of Compromise: For compromised outside plant, this would 918 be the length of the sharing. 919 A CSPF algorithm could then penalize a diversity compromise by an 920 amount dependent on these two parameters. 922 Two links could be related by many SRLG's (AT&T's experience 923 indicates that a link may belong to over 100 SRLG's, each 924 corresponding to a separate fiber group. Each SRLG might relate a 925 single link to many other links. For the optical layer, similar 926 situations can be expected where a link is an ultra-long OTS). 928 The mapping between links and different types of SRLG�s is in 929 general defined by network operators based on the definition of each 930 SRLG type. Since SRLG information is not yet ready to be 931 discoverable by a network element and does not change dynamically, 932 it need not be advertised with other resource availability 933 information by network elements. It could be configured in some 934 central database and be distributed to or retrieved by the nodes, or 935 advertised by network elements at the topology discovery stage. 937 6.2 Implications For Routing 938 Dealing with diversity is an unavoidable requirement for routing in 939 the optical layer. It requires dealing with constraints in the 940 routing process but most importantly requires additional state 941 information � the SRLG relationships and also the routings of any 942 existing circuits from the new circuit is to be diverse � to be 943 available to the routing process. 945 At present SRLG information cannot be self-discovered. Indeed, in a 946 large network it is very difficult to maintain accurate SRLG 947 information. The problem becomes particularly daunting whenever 948 multiple administrative domains are involved, for instance after the 949 acquisition of one network by another, because there normally is a 950 likelihood that there are diversity violations between the domains. 951 It is very unlikely that diversity relationships between carriers 952 will be known any time in the near future. 954 Considerable variation in what different customers will mean by 955 acceptable diversity should be anticipated. Consequently we suggest 956 that an SRLG should be defined as follows: (i) It is a relationship 957 between two or more links, and (ii) it is characterized by two 958 parameters, the type of compromise (shared conduit, shared ROW, 959 shared optical ring, etc.) and the extent of the compromise (e.g., 961 Impairments And Other Constraints May 2001 962 On Optical Layer Routing 964 the number of miles over which the compromise persisted). This will 965 allow the SRLG�s appropriate to a particular routing request to be 966 easily identified. 968 7. Security Considerations 970 The solution developed to address the requirements defined in this 971 document must address security aspects. 973 8. Acknowledgments 975 This document has benefited from discussions with Michael Eiselt, 976 Jonathan Lang, Mark Shtaif, Jennifer Yates, Dongmei Wang, Guangzhi 977 Li, Robert Doverspike, Albert Greenberg, Jim Maloney, John Jacob, 978 Katie Hall and with our OIF colleagues. 980 References: 982 [ABB01] Ashwood-Smith, P., et. al., "Generalized MPLS Signaling 983 Functional Description,_ Internet draft, draft-ietf- generalized- 984 mpls-signaling-00.txt, work in progress, March 2001. 986 [Ashwood00] Ashwood-Smith, P. et al., "MPLS Optical/Switching 987 Signaling Functional Description", Work in Progress, draft-ashwood- 988 generalized-mpls-signaling-00.txt. 990 [Awduche99] Awduche, D. O., Rekhter, Y., Drake, J., and Coltun, R., 991 "Multi-Protocol Lambda Switching: Combining MPLS Traffic Engineering 992 Control With Optical Crossconnects", Work in Progress, draft- 993 awduche-mpls-te-optical-01.txt. 995 [Bra96] Bradner, S., "The Internet Standards Process -- Revision 3," 996 BCP 9, RFC 2026, October 1996. 998 [CBD00] Ceuppens, L., Blumenthal, D., Drake, J., Chrostowski, J., 999 Edwards, W., "Performance Monitoring in Photonic Networks in 1000 Support of MPL(ambda)S", Internet draft, work in progress, March 1001 2000. 1003 [Chaudhuri00] Chaudhuri, S., Hjalmtysson, G., and Yates, J., 1004 "Control of Lightpaths in an Optical Network", Work in Progress, 1005 draft-chaudhuri-ip-olxc-control-00.txt. 1007 Impairments And Other Constraints May 2001 1008 On Optical Layer Routing 1010 [Doverspike00] Doverspike, R. and Yates, J., "Challenges For MPLS in 1011 Optical Network Restoration", IEEE Communication Magazine, February, 1012 2001. 1014 [Gerstel 2000] O. Gorstel, "Optical Layer Signaling: How Much Is 1015 Really Needed?", IEEE Communications Magazine, vol. 38 no. 10, Oct. 1016 2000, pp. 154-160 1018 [Goldstein94] Goldstein, E. L., Eskildsen, L., and Elrefaie, A. F., 1019 Performance Implications of Component Crosstalk in Transparent 1020 Lightwave Networks", IEEE Photonics Technology Letters, Vol.6, No.5, 1021 May 1994. 1023 [ITU] ITU-T Doc. G.663, Optical Fibers and Amplifiers, Section 1024 II.4.1.2. 1026 [Kaminow97] Kaminow, I. P. and Koch, T. L., editors, Optical Fiber 1027 Telecommunications IIIA, Academic Press, 1997. 1029 KRB01a] Kompella, K., et.al., "IS-IS extensions in support of 1030 Generalized MPLS," Internet Draft, draft-ietf-gmpls- extensions- 1031 01.txt, work in progress, 2001. 1033 [KRB01b] Kompella, K., et. al., "OSPF extensions in support of 1034 Generalized MPLS," Internet draft, draft-ospf-generalized- mpls- 1035 00.txt, work in progress, March 2001. 1037 [Moy98] Moy, John T., OSPF: Anatomy of an Internet Routing Protocol, 1038 Addison-Wesley, 1998. 1040 [Ramaswami98] Ramaswami, R. and Sivarajan, K. N., Optical Networks: 1041 A Practical Perspective, Morgan Kaufmann Publishers, 1998. 1043 [Strand01] J. Strand, A. Chiu, and R. Tkach, �Issues for Routing in 1044 the Optical Layer�, IEEE Communications Magazine, Feb. 2001, vol. 39 1045 No. 2, pp. 81-88; also see "Unique Features and Requirements for The 1046 Optical Layer Control Plane", Internet Draft, draft-chiu-strand- 1047 unique-olcp-01.txt, work in progress, November 2000. 1049 [Strand01b] J. Strand, R. Doverspike, and G. Li, �Importance of 1050 Wavelength Conversion In An Optical Network�, Optical Networks 1051 Magazine, accepted for publication. Available from the author. 1053 Impairments And Other Constraints May 2001 1054 On Optical Layer Routing 1056 [Tkach98] Tkach, R., Goldstein, E., Nagel, J., and Strand, J., 1057 "Fundamental Limits of Optical Transparency", Optical Fiber 1058 Communication Conf., Feb. 1998, pp. 161-162. 1060 [Yates99] Yates, J. M., Rumsewicz, M. P. and Lacey, J. P. R., 1061 "Wavelength Converters in Dynamically-Reconfigurable WDM Networks", 1062 IEEE Communications Surveys, 2Q1999 (online at 1063 www.comsoc.org/pubs/surveys/2q99issue/yates.html). 1065 Authors' Addresses: 1067 Ayan Banerjee 1068 Calient Networks 1069 5853 Rue Ferrari 1070 San Jose, CA 95138 1071 Email: abanerjee@calient.net 1073 Dan Blumenthal 1074 Calient Networks 1075 5853 Rue Ferrari 1076 San Jose, CA 95138 1077 Email: dblumenthal@calient.net 1079 Angela Chiu 1080 Celion Networks 1081 1 Shiela Dr., Suite 2 1082 Tinton Falls, NJ 07724 1083 Phone:(732) 747-9987 1084 Email: angela.chiu@celion.com 1086 John Drake 1087 Calient Networks 1088 5853 Rue Ferrari 1089 San Jose, CA 95138 1090 Email: jdrake@calient.net 1092 Andre Fredette 1093 PhotonEx Corporation 1094 200 Metrowest Technology Dr. 1095 Maynard, MA 01754 1096 Email: fredette@photonex.com 1098 Nan Froberg 1099 PhotonEx Corporation 1100 200 Metrowest Technology Dr. 1101 Maynard, MA 01754 1102 Email: nfroberg@photonex.com 1104 Impairments And Other Constraints May 2001 1105 On Optical Layer Routing 1107 James V. Luciani 1108 900 Chelmsford St. 1109 Lowell, MA 01851 1110 +1 978 275 3182 1111 james_luciani@mindspring.com 1113 John Strand 1114 AT&T Labs 1115 200 Laurel Ave., Rm A5-1D06 1116 Middletown, NJ 07748 1117 Phone:(732) 420-9036 1118 Email: jls@research.att.com 1120 Robert Tkach 1121 Celion Networks 1122 1 Shiela Dr., Suite 2 1123 Tinton Falls, NJ 07724 1124 Phone:(732) 747-9909 1125 Email: bob.tkach@celion.com