idnits 2.17.1 draft-multiple-layer-resource-optimization-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** 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.) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document doesn't use any RFC 2119 keywords, yet seems to have RFC 2119 boilerplate text. -- The document date (October 28, 2019) is 1641 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 152 -- Looks like a reference, but probably isn't: '2' on line 90 -- Looks like a reference, but probably isn't: '3' on line 104 -- Looks like a reference, but probably isn't: '0' on line 171 == Missing Reference: 'Step' is mentioned on line 171, but not defined -- Looks like a reference, but probably isn't: '4' on line 282 == Unused Reference: 'Ref1' is defined on line 417, but no explicit reference was found in the text == Unused Reference: 'Ref2' is defined on line 421, but no explicit reference was found in the text == Unused Reference: 'Ref3' is defined on line 427, but no explicit reference was found in the text == Unused Reference: 'Ref4' is defined on line 431, but no explicit reference was found in the text Summary: 2 errors (**), 0 flaws (~~), 7 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Cross Stratum Optimization Research Group H. Yang 3 Internet-Draft K. Zhan 4 Intended status: Informational A. Yu 5 Expires: April 30, 2020 Q. Yao 6 J. Zhang 7 Beijing University of Posts and Telecommunications 8 October 28, 2019 10 Multiple Layer Resource Optimization for Optical as a Service 11 draft-multiple-layer-resource-optimization-02 13 Abstract 15 We have established a neural network model optimized by adaptive 16 artificial fish swarm algorithm. Then we propose a novel multi-path 17 pre-reserved resource allocation strategy to increase resource 18 utilization. The results prove the effectiveness of our method. 20 Status of This Memo 22 This Internet-Draft is submitted in full conformance with the 23 provisions of BCP 78 and BCP 79. 25 Internet-Drafts are working documents of the Internet Engineering 26 Task Force (IETF). Note that other groups may also distribute 27 working documents as Internet-Drafts. The list of current Internet- 28 Drafts is at https://datatracker.ietf.org/drafts/current/. 30 Internet-Drafts are draft documents valid for a maximum of six months 31 and may be updated, replaced, or obsoleted by other documents at any 32 time. It is inappropriate to use Internet-Drafts as reference 33 material or to cite them other than as "work in progress." 35 This Internet-Draft will expire on April 30, 2020. 37 Copyright Notice 39 Copyright (c) 2019 IETF Trust and the persons identified as the 40 document authors. All rights reserved. 42 This document is subject to BCP 78 and the IETF Trust's Legal 43 Provisions Relating to IETF Documents 44 (https://trustee.ietf.org/license-info) in effect on the date of 45 publication of this document. Please review these documents 46 carefully, as they describe your rights and restrictions with respect 47 to this document. Code Components extracted from this document must 48 include Simplified BSD License text as described in Section 4.e of 50 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 58 1.1. Conventions Used in This Document . . . . . . . . . . . . 3 59 2. PREDICTION STRATEGY . . . . . . . . . . . . . . . . . . . . . 3 60 2.1. Artificial neural network model . . . . . . . . . . . . . 3 61 2.2. Adaptive artificial fish swarm artificial neural networks 62 (AAFS-ANN ) . . . . . . . . . . . . . . . . . . . . . . . 4 63 3. MULTI-PATH PRE-RESERVED RESOURCE ALLOCATION . . . . . . . . . 4 64 3.1. Reconfiguration time calculation . . . . . . . . . . . . 6 65 3.2. Multi-path pre-reserved resource allocation(MP-RA) . . . 6 66 4. Experimental evaluation and results analysis . . . . . . . . 7 67 5. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . 9 68 6. ACKNOWLEDGMENT . . . . . . . . . . . . . . . . . . . . . . . 9 69 7. References . . . . . . . . . . . . . . . . . . . . . . . . . 9 70 7.1. Normative References . . . . . . . . . . . . . . . . . . 9 71 7.2. Informative References . . . . . . . . . . . . . . . . . 9 72 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 10 74 1. Introduction 76 With the rapid growth of cloud computing, 5G services, and the 77 periodicity of people's activities, traffic load has exhibited 78 periodicity in both time and space domains, namely tidal traffic [1]. 79 The number of people using optical metropolitan networks is enormous 80 and unevenly distributed. In addiction, the separation of work areas 81 and residential areas is an important cause of tidal traffic. 82 Generally, tidal traffic will reduce the performance of networks 83 during to following two reasons: firstly, the network traffic will be 84 blocked due to the sharp increase in traffic in the high-traffic 85 area; secondly, network nodes may be idle and waste resources in the 86 low-traffic areas. The static configuration resources will intensify 87 both network and service congestion during traffic peak hours, as 88 well as low resource utilization during low-traffic times and 89 regions. In the future, global mobile Internet traffic will increase 90 by 10 times [2], urbanization is rapidly advancing, the scope and 91 severity of space and time domains affected by tidal traffic are 92 increasing as communication need and network technologies developing. 93 Tidal traffic will further affect the optical access network and the 94 optical core network, making it essential issue for network 95 operators. Therefore, a more reasonable and efficient resource 96 allocation scheme is urgently needed to solve the congestion and 97 resource waste caused by the tidal traffic. 99 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 101 Known from the above, tidal traffic prediction becomes the core 102 process of network optimization decision-making. Currently, there 103 are several prediction methods, like support vector machine (SVM) and 104 multi-layer perceptron (MLP). Literature [3] proposes a deep- 105 learning-based prediction strategy to implement traffic assessment of 106 data center optical networks. At the same time, a deep-learning- 107 based global evaluation factor resource allocation algorithm is 108 suggested to achieve lower blocking rate of the network. Compared 109 with the traditional algorithm, deep learning can improve the 110 accuracy of prediction, but it cannot identify the tidal traffic in 111 specific festivals. In addition, the lower priority service will be 112 discarded to reduce the network blocking rate. This method does not 113 make good use of idle resources of other nodes, and some traffic 114 requests cannot be executed normally. So we propose multi-path pre- 115 reserved resource allocation based on traffic prediction. 117 In this paper, we establish an adaptive artificial fish-group neural 118 network model to predict traffic, then use the predicted traffic 119 demand to optimize the network at different times. Meanwhile, we 120 propose multi-path pre-reserved resource allocation to adapt to the 121 resource requirements of different nodes. Simulation results 122 demonstrate that our strategy achieves a lower network blocking rate 123 and higher resource utilization. 125 1.1. Conventions Used in This Document 127 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 128 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 129 document are to be interpreted as described in [RFC2119]. 131 2. PREDICTION STRATEGY 133 Before presenting the resource allocation algorithm, we provide an 134 introduction to traffic prediction model. We establish a neural 135 network of adaptive artificial fish algorithm to predict traffic 136 request. The key resides in the construction of the artificial fish 137 individual model. The optimal variables of the neural network are 138 two weight matrices and two threshold variables _io,v_o . 140 2.1. Artificial neural network model 142 We build the neural network structure as shown in figure 1. The 143 input is composed of six entries, i_(s,1) is the hour of the day, 144 i_(s,2) is the day of the week, i_(s,3) is a flag for holiday/ 145 weekend, i_(s,4) is the previous days average load, i_(s,5) is the 146 load from the same hour of the previous day, and i_(s,6) is the load 147 from the same hour and same day from the previous week. The result 148 of the output node Y_(s,1) represents the traffic request that we 150 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 152 want to predict[1]. Training sample setA={(X^i,Y^i )|i=1,2,,n}X^i is 153 the i_th group training data input, and Y^i is the i_th group input 154 corresponding expected output. We define the error function as 155 follows: 157 where O^i is the actual output of the i_th. 159 2.2. Adaptive artificial fish swarm artificial neural networks (AAFS- 160 ANN ) 162 In the artificial fish swarm algorithm, we introduce adaptive step 163 size and visible range to improve convergence accuracy and speed. 164 Generate initial artificial fish population N, namely N group 165 {omega_ij,nu_io,omega_io,nu_o}. Every artificial fish is a neural 166 network. The food concentration is defined as FC=1/E. X_i is the 167 state of current location state,X_j is random state of the 168 search,d_ij is the distance between X_i and X_j, omega_ij 169 (i),omega_ij (j) and omega_ij (i+1) respectively are X_i,X_jnext 170 state X_(i+1) matrix omega_ij} element of i_th row j_th column, 171 "Rand(Step") represents a random number between [0, Step]. 173 Let X_0 be the current artificial fish, its position is C, X_1 is the 174 current optimal fish, X_2 is the nearest fish, Then we set two 175 visible fields viusual_1=d_01,viusual_2=d_02. Two target positions 176 A, B are randomly determined in the range of viusual_1 and viusual_2 177 respectively, then compare FC_A,FC_B,FC_C, 179 If FC_A,FC_B 181 If FC_A,FC_B 183 omega(i+1)=omega(i)+Rand(step) 185 If one or both of them are better than C, Then advance to the best 186 point, and execute formula (3) 188 omega(i+1)=omega(i)+Rand(step)(omega(j)-omega(i))/d_ij 190 Go for A with viusual_1Rand() as the step size, to B with 191 viusual_2alphaRand(), where a ,which equal to 1 or slightly less than 192 1,is the visual factor.The other three optimization variables are 193 similarly. 195 3. MULTI-PATH PRE-RESERVED RESOURCE ALLOCATION 197 The resource allocation method bases on the AAFS-ANN described above, 198 and we propose a multi-path pre-reserved resource allocation way to 199 optimize optical network. We uses the predicted result to perform 201 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 203 configuration time calculation and estimate the future network 204 resource demand to pre-reserve resource for traffic request. 206 ------------------------------------- 207 | --- --- | 208 | | A |--------------| B | | 209 | --- \ --- \ | 210 | \ \ | 211 | \ \ | 212 | \ \ | 213 | --- --- | 214 | | C |---------------| D | | 215 | --- --- | 216 ------------------------------------- 218 Fig.1(a) Sample network 220 |------------------------------------ 221 T4 | | | | | | | | | | 222 |------------------------------------ 223 T3 | | * | * | | | | | | | 224 |------------------------------------ 225 T2 | | * | * | | | | | | | 226 |------------------------------------ 227 T1 | | | | | | | | | | 228 |------------------------------------ 229 T0 | | | | | | | | | | 230 ------------------------------------- 231 S0 S1 S2 S3 S4 233 Fig.1(b) Requested resources 235 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 237 |------------------------------------ 238 T4 | | | | | | | | | | 239 |------------------------------------ 240 T3 | # | * | # | # | # | | | | | 241 |------------------------------------ 242 T2 | # | * | # | # | # | | | | | 243 |------------------------------------ 244 T1 | # | | | # | # | | | | | 245 |------------------------------------ 246 T0 | # | | | # | # | | | | | 247 ------------------------------------ 248 S0 S1 S2 S3 S4 250 Fig.1(c) Requested resources 252 |------------------------------------ 253 T4 | | | | | | | | | | 254 |------------------------------------ 255 T3 | # | # | * | | # | # | | | | 256 |------------------------------------ 257 T2 | # | # | * | # | # | # | | | | 258 |------------------------------------ 259 T1 | # | | | # | | | | | | 260 |------------------------------------ 261 T0 | # | | | | | | | | | 262 ------------------------------------ 263 S0 S1 S2 S3 S4 265 Fig.1(d) Requested resources 267 3.1. Reconfiguration time calculation 269 Frequent reconfiguration can result in service interruption and 270 unstable of distributed routing algorithm, so we need to predict the 271 next 24-hour traffic demand D^24 for the next configuration time 272 point calculation. Algorithm 1 is the calculation process of the 273 reconfiguration time point. 275 3.2. Multi-path pre-reserved resource allocation(MP-RA) 277 We reserve network resources for the predicted traffic. This type of 278 service request is called Advance Reservation Service (AR). As the 279 optical link is continuously established and removed, fragments are 280 easily generated in both the time domain and the spectrum domain. 281 The application of the Sliceable Bandwidth Variable Transceiver 282 (S-BVT) [4] further enhances the flexibility of EON. The S-BVT has a 283 slicing capability, i.e. it can provide multiple optical carriers for 285 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 287 carrying optical links to different destinations. In order to reduce 288 time and spectral fragmentation (referred to as two-dimensional 289 fragmentation) and to solve the problem of insufficient resources, we 290 propose cutting the request into multiple parts, and transfer on 291 multiple paths. 293 The underlying optical network can be modeled as 294 G_s=(L_s,N_s,R_st,D_s){ L_s: link set, N_s: optical node set, R_st: 295 resource status of optical nodes and optical links at time t, D_s: 296 distance of each pair of nodes in the set of nodes N in the network 297 topology}. R_A=(s,d,w,b,h) denotes a predicted service request, where 298 s and d represent the source and destination nodes of the service, b 299 is the time of service starts, h is the duration of the AR service, 300 and w is the service start time b, and the duration h period required 301 link rate. P_((s,d)) represents the path set of the source node to 302 the destination node. 304 If there are not enough spectrum resources available in the link for 305 the incoming request, we will attempt to cut the request into 306 different parts and assign those parts to different frequency bands. 307 For a simple example, as shown in Firgue 3(a), in order to reflect 308 the state of the spectral resources in the time domain, we use a two- 309 dimensional time spectrum resource model and assume that each time 310 slot has the same time period. The network diagram is illustrated in 311 the figure 3(a). Now there was an AR request, from node A to node D. 312 The request requires two spectrum slots, lasting from T2 to T3, as 313 shown in figure 2(b). Figure 2(c) and Figure 2(d) show the spectrum 314 states of path A-C-D and path A-B-D, respectively. The black slot 315 represents the occupied spectrum slot, the white slot represents the 316 spectrum slot available for the spectrum resource, and the blue slot 317 represents the spectrum slot occupied by the AR request. Before 318 splitting the AR request, the two paths do not have enough resources 319 to allocate. However, after we split the request into two parts, we 320 can distribute them to two spectrum segments to implement AR- 321 requested service provision. The MP-RA is as shown in Algorithm 2. 323 4. Experimental evaluation and results analysis 325 In this paper, we present the results of the AAFS-ANN prediction. 326 Our goal is to demonstrate the accuracy and network performance of 327 AAFS-ANN in different network environments. To fully reflect the 328 changes in the network environment, we use WIDE data from 96h traffic 329 data from April 6th to 9th, 2017, to train and verify. Figure 3(a) 330 is the comparison between the actual traffic and the prediction 331 results, which verify the effectiveness of our method. As shown in 332 Figure 3(a), the prediction results of AAFS-ANN are significantly 333 better than the traditional predictions. This is because the 334 introduction of the adaptive step size and the visible field, making 336 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 338 the artificial fish compares the FC in the large field of view. Our 339 method enhances the global convergence and the optimization 340 precision. The prediction error occurs because the traffic is 341 directly affected by many non-linear sudden factors such as hot 342 events, user movement patterns. Therefore, many traffic cannot be 343 accurately predicted. 345 We also compare MP-RA with several state-of-the-art resource 346 allocation techniques including evolutionary algorithms(EA) and 347 artificial neural networks (ANN). From firgue 3(b), we can see that 348 MP-RA performs well among the three optimization resource allocation 349 method, MP-RA greatly improves resource utilization. 351 According to the prediction results, the MP-RA can allocate resources 352 to traffic more reasonably. This is because the algorithm considers 353 the traffic that will be reached at each point in time and the 354 resources it needs. Then re-plans the resources at the configuration 355 time. As can be observed in the results shown in Figure 3(c), MP-RA 356 can greatly reduce the probability of traffic blocking. 358 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 359 | | Resource utilization rate | 360 | Traffic load +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 361 | | MP-RA | ANN | EA | 362 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 363 | 40 | 0.254 | 0.242 | 0.251 | 364 | 70 | 0.263 | 0.253 | 0.272 | 365 | 95 | 0.273 | 0.275 | 0.300 | 366 | 120 | 0.332 | 0.29 | 0.420 | 367 | 145 | 0.389 | 0.325 | 0.504 | 368 | 170 | 0.457 | 0.356 | 0.583 | 369 | 200 | 0.52 | 0.403 | 0.723 | 370 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 372 Tab.1 Network blocking probability of four strategies 374 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 376 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 377 | | Network blocking probability | 378 | Traffic load +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 379 | | MP-RA | ANN | EA | 380 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 381 | 50 | 0.008 | 0.0075 | 0.0078 | 382 | 70 | 0.009 | 0.010 | 0.012 | 383 | 100 | 0.0095 | 0.025 | 0.029 | 384 | 125 | 0.01 | 0.06 | 0.074 | 385 | 150 | 0.0108 | 0.08 | 0.10 | 386 | 175 | 0.025 | 0.115 | 0.129 | 387 | 200 | 0.06 | 0.15 | 0.20 | 388 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 390 Tab.2 Average hop of four strategies 392 5. CONCLUSION 394 In the tidal traffic scenario, we propose AAFS-ANN model and MP-RA 395 strategy. We use AAFS-ANN model to predict traffic and MP-RA to 396 optimize metropolitan optical network. Results demonstrate that 397 AAFS-ANN and MP-RA successfully increase prediction accuracy and 398 resource utilization, as well as reduce the traffic blocking rate. 400 6. ACKNOWLEDGMENT 402 This work has been supported in part by NSFC project (61871056), 403 National Postdoctoral Program for Innovative Talents (BX201600021), 404 Fundamental Research Funds for the Central Universities (2018XKJC06) 405 and State Key Laboratory of Information Photonics and Optical 406 Communications (BUPT), P. R. China (No. IPOC2017ZT11). 408 7. References 410 7.1. Normative References 412 [RFC2119] Bradner, S., "Key words for use in RFC's to Indicate 413 Requirement Levels", RFC 2119, March 1997. 415 7.2. Informative References 417 [Ref1] Alvizu, R., Troia, S., and G. Maier, "Matheuristic with 418 machine-learning-based prediction for software-defined 419 mobile metro-core networks", May 2017. 421 [Ref2] Zhong, Z., Hua, N., and H. Liu, "Considerations of 422 effective tidal traffic dispatching in software-defined 423 metro IP over optical networks", July 2015. 425 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 427 [Ref3] Yu, A., Yang, H., and W. Bai, "Leveraging deep learning to 428 achieve efficient resource allocation with traffic 429 evaluation in datacenter optical networks", March 2018. 431 [Ref4] Zhong, Z., Hua, N., and M. Tornatore, "Energy efficiency 432 and blocking reduction for tidal traffic via stateful 433 grooming in IP-over-optical networks", September 2016. 435 Authors' Addresses 437 Hui Yang 438 Beijing University of Posts and Telecommunications 439 No.10,Xitucheng Road,Haidian District 440 Beijing 100876 441 P.R.China 443 Phone: +8613466774108 444 Email: yang.hui.y@126.com 445 URI: http://www.bupt.edu.cn/ 447 Kaixuan Zhan 448 Beijing University of Posts and Telecommunications 449 No.10,Xitucheng Road,Haidian District 450 Beijing 100876 451 P.R.China 453 Phone: +8618401695826 454 Email: zhankai@bupt.edu.cn 455 URI: http://www.bupt.edu.cn/ 457 Ao Yu 458 Beijing University of Posts and Telecommunications 459 No.10,Xitucheng Road,Haidian District 460 Beijing 100876 461 P.R.China 463 Email: yuao@bupt.edu.cn 464 URI: http://www.bupt.edu.cn/ 466 Internet-DrafMultiple Layer Resource Optimization for Optic October 2019 468 Qiuyan Yao 469 Beijing University of Posts and Telecommunications 470 No.10,Xitucheng Road,Haidian District 471 Beijing 100876 472 P.R.China 474 Email: yqy89716@bupt.edu.cn 475 URI: http://www.bupt.edu.cn/ 477 Jie Zhang 478 Beijing University of Posts and Telecommunications 479 No.10,Xitucheng Road,Haidian District 480 Beijing 100876 481 P.R.China 483 Phone: +8613911060930 484 Email: lgr@bupt.edu.cn 485 URI: http://www.bupt.edu.cn/