idnits 2.17.1 draft-multiple-layer-resource-optimization-01.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 (May 5, 2019) is 1789 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 144 -- Looks like a reference, but probably isn't: '2' on line 87 -- Looks like a reference, but probably isn't: '3' on line 99 -- Looks like a reference, but probably isn't: '0' on line 163 == Missing Reference: 'Step' is mentioned on line 163, but not defined -- Looks like a reference, but probably isn't: '4' on line 269 == Unused Reference: 'Ref1' is defined on line 395, but no explicit reference was found in the text == Unused Reference: 'Ref2' is defined on line 399, but no explicit reference was found in the text == Unused Reference: 'Ref3' is defined on line 403, but no explicit reference was found in the text == Unused Reference: 'Ref4' is defined on line 407, 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: November 6, 2019 Q. Yao 6 J. Zhang 7 Beijing University of Posts and Telecommunications 8 May 5, 2019 10 Multiple Layer Resource Optimization for Optical as a Service 11 draft-multiple-layer-resource-optimization-01 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 November 6, 2019. 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 49 the Trust Legal Provisions and are provided without warranty as 50 described in the Simplified BSD License. 52 Table of Contents 54 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 55 1.1. Conventions Used in This Document . . . . . . . . . . . . 3 56 2. PREDICTION STRATEGY . . . . . . . . . . . . . . . . . . . . . 3 57 2.1. Artificial neural network model . . . . . . . . . . . . . 3 58 2.2. Adaptive artificial fish swarm artificial neural networks 59 (AAFS-ANN ) . . . . . . . . . . . . . . . . . . . . . . . 4 60 3. MULTI-PATH PRE-RESERVED RESOURCE ALLOCATION . . . . . . . . . 4 61 3.1. Reconfiguration time calculation . . . . . . . . . . . . 6 62 3.2. Multi-path pre-reserved resource allocation(MP-RA) . . . 6 63 4. Experimental evaluation and results analysis . . . . . . . . 7 64 5. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . 9 65 6. ACKNOWLEDGMENT . . . . . . . . . . . . . . . . . . . . . . . 9 66 7. References . . . . . . . . . . . . . . . . . . . . . . . . . 9 67 7.1. Normative References . . . . . . . . . . . . . . . . . . 9 68 7.2. Informative References . . . . . . . . . . . . . . . . . 9 69 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 10 71 1. Introduction 73 With the rapid growth of cloud computing, 5G services, and the 74 periodicity of people's activities, traffic load has exhibited 75 periodicity in both time and space domains, namely tidal traffic [1]. 76 The number of people using optical metropolitan networks is enormous 77 and unevenly distributed. In addiction, the separation of work areas 78 and residential areas is an important cause of tidal traffic. 79 Generally, tidal traffic will reduce the performance of networks 80 during to following two reasons: firstly, the network traffic will be 81 blocked due to the sharp increase in traffic in the high-traffic 82 area; secondly, network nodes may be idle and waste resources in the 83 low-traffic areas. The static configuration resources will intensify 84 both network and service congestion during traffic peak hours, as 85 well as low resource utilization during low-traffic times and 86 regions. In the future, global mobile Internet traffic will increase 87 by 10 times [2], urbanization is rapidly advancing, the scope and 88 severity of space and time domains affected by tidal traffic are 89 increasing as communication need and network technologies developing. 90 Tidal traffic will further affect the optical access network and the 91 optical core network, making it essential issue for network 92 operators. Therefore, a more reasonable and efficient resource 93 allocation scheme is urgently needed to solve the congestion and 94 resource waste caused by the tidal traffic. 96 Known from the above, tidal traffic prediction becomes the core 97 process of network optimization decision-making. Currently, there 98 are several prediction methods, like support vector machine (SVM) and 99 multi-layer perceptron (MLP). Literature [3] proposes a deep- 100 learning-based prediction strategy to implement traffic assessment of 101 data center optical networks. At the same time, a deep-learning- 102 based global evaluation factor resource allocation algorithm is 103 suggested to achieve lower blocking rate of the network. Compared 104 with the traditional algorithm, deep learning can improve the 105 accuracy of prediction, but it cannot identify the tidal traffic in 106 specific festivals. In addition, the lower priority service will be 107 discarded to reduce the network blocking rate. This method does not 108 make good use of idle resources of other nodes, and some traffic 109 requests cannot be executed normally. So we propose multi-path pre- 110 reserved resource allocation based on traffic prediction. 112 In this paper, we establish an adaptive artificial fish-group neural 113 network model to predict traffic, then use the predicted traffic 114 demand to optimize the network at different times. Meanwhile, we 115 propose multi-path pre-reserved resource allocation to adapt to the 116 resource requirements of different nodes. Simulation results 117 demonstrate that our strategy achieves a lower network blocking rate 118 and higher resource utilization. 120 1.1. Conventions Used in This Document 122 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 123 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 124 document are to be interpreted as described in [RFC2119]. 126 2. PREDICTION STRATEGY 128 Before presenting the resource allocation algorithm, we provide an 129 introduction to traffic prediction model. We establish a neural 130 network of adaptive artificial fish algorithm to predict traffic 131 request. The key resides in the construction of the artificial fish 132 individual model. The optimal variables of the neural network are 133 two weight matrices and two threshold variables _io,v_o . 135 2.1. Artificial neural network model 137 We build the neural network structure as shown in figure 1. The 138 input is composed of six entries, i_(s,1) is the hour of the day, 139 i_(s,2) is the day of the week, i_(s,3) is a flag for holiday/ 140 weekend, i_(s,4) is the previous days average load, i_(s,5) is the 141 load from the same hour of the previous day, and i_(s,6) is the load 142 from the same hour and same day from the previous week. The result 143 of the output node Y_(s,1) represents the traffic request that we 144 want to predict[1]. Training sample setA={(X^i,Y^i )|i=1,2,,n}X^i is 145 the i_th group training data input, and Y^i is the i_th group input 146 corresponding expected output. We define the error function as 147 follows: 149 where O^i is the actual output of the i_th. 151 2.2. Adaptive artificial fish swarm artificial neural networks (AAFS- 152 ANN ) 154 In the artificial fish swarm algorithm, we introduce adaptive step 155 size and visible range to improve convergence accuracy and speed. 156 Generate initial artificial fish population N, namely N group 157 {omega_ij,nu_io,omega_io,nu_o}. Every artificial fish is a neural 158 network. The food concentration is defined as FC=1/E. X_i is the 159 state of current location state,X_j is random state of the 160 search,d_ij is the distance between X_i and X_j, omega_ij 161 (i),omega_ij (j) and omega_ij (i+1) respectively are X_i,X_jnext 162 state X_(i+1) matrix omega_ij} element of i_th row j_th column, 163 "Rand(Step") represents a random number between [0, Step]. 165 Let X_0 be the current artificial fish, its position is C, X_1 is the 166 current optimal fish, X_2 is the nearest fish, Then we set two 167 visible fields viusual_1=d_01,viusual_2=d_02. Two target positions 168 A, B are randomly determined in the range of viusual_1 and viusual_2 169 respectively, then compare FC_A,FC_B,FC_C, 171 If FC_A,FC_B 173 If FC_A,FC_B 175 omega(i+1)=omega(i)+Rand(step) 177 If one or both of them are better than C, Then advance to the best 178 point, and execute formula (3) 180 omega(i+1)=omega(i)+Rand(step)(omega(j)-omega(i))/d_ij 182 Go for A with viusual_1Rand() as the step size, to B with 183 viusual_2alphaRand(), where a ,which equal to 1 or slightly less than 184 1,is the visual factor.The other three optimization variables are 185 similarly. 187 3. MULTI-PATH PRE-RESERVED RESOURCE ALLOCATION 189 The resource allocation method bases on the AAFS-ANN described above, 190 and we propose a multi-path pre-reserved resource allocation way to 191 optimize optical network. We uses the predicted result to perform 192 configuration time calculation and estimate the future network 193 resource demand to pre-reserve resource for traffic request. 195 ------------------------------------- 196 | --- --- | 197 | | A |--------------| B | | 198 | --- \ --- \ | 199 | \ \ | 200 | \ \ | 201 | \ \ | 202 | --- --- | 203 | | C |---------------| D | | 204 | --- --- | 205 ------------------------------------- 207 Fig.1(a) Sample network 209 |------------------------------------ 210 T4 | | | | | | | | | | 211 |------------------------------------ 212 T3 | | * | * | | | | | | | 213 |------------------------------------ 214 T2 | | * | * | | | | | | | 215 |------------------------------------ 216 T1 | | | | | | | | | | 217 |------------------------------------ 218 T0 | | | | | | | | | | 219 ------------------------------------- 220 S0 S1 S2 S3 S4 222 Fig.1(b) Requested resources 224 |------------------------------------ 225 T4 | | | | | | | | | | 226 |------------------------------------ 227 T3 | # | * | # | # | # | | | | | 228 |------------------------------------ 229 T2 | # | * | # | # | # | | | | | 230 |------------------------------------ 231 T1 | # | | | # | # | | | | | 232 |------------------------------------ 233 T0 | # | | | # | # | | | | | 234 ------------------------------------ 235 S0 S1 S2 S3 S4 237 Fig.1(c) Requested resources 239 |------------------------------------ 240 T4 | | | | | | | | | | 241 |------------------------------------ 242 T3 | # | # | * | | # | # | | | | 243 |------------------------------------ 244 T2 | # | # | * | # | # | # | | | | 245 |------------------------------------ 246 T1 | # | | | # | | | | | | 247 |------------------------------------ 248 T0 | # | | | | | | | | | 249 ------------------------------------ 250 S0 S1 S2 S3 S4 252 Fig.1(d) Requested resources 254 3.1. Reconfiguration time calculation 256 Frequent reconfiguration can result in service interruption and 257 unstable of distributed routing algorithm, so we need to predict the 258 next 24-hour traffic demand D^24 for the next configuration time 259 point calculation. Algorithm 1 is the calculation process of the 260 reconfiguration time point. 262 3.2. Multi-path pre-reserved resource allocation(MP-RA) 264 We reserve network resources for the predicted traffic. This type of 265 service request is called Advance Reservation Service (AR). As the 266 optical link is continuously established and removed, fragments are 267 easily generated in both the time domain and the spectrum domain. 268 The application of the Sliceable Bandwidth Variable Transceiver 269 (S-BVT) [4] further enhances the flexibility of EON. The S-BVT has a 270 slicing capability, i.e. it can provide multiple optical carriers for 271 carrying optical links to different destinations. In order to reduce 272 time and spectral fragmentation (referred to as two-dimensional 273 fragmentation) and to solve the problem of insufficient resources, we 274 propose cutting the request into multiple parts, and transfer on 275 multiple paths. 277 The underlying optical network can be modeled as 278 G_s=(L_s,N_s,R_st,D_s){ L_s: link set, N_s: optical node set, R_st: 279 resource status of optical nodes and optical links at time t, D_s: 280 distance of each pair of nodes in the set of nodes N in the network 281 topology}. R_A=(s,d,w,b,h) denotes a predicted service request, where 282 s and d represent the source and destination nodes of the service, b 283 is the time of service starts, h is the duration of the AR service, 284 and w is the service start time b, and the duration h period required 285 link rate. P_((s,d)) represents the path set of the source node to 286 the destination node. 288 If there are not enough spectrum resources available in the link for 289 the incoming request, we will attempt to cut the request into 290 different parts and assign those parts to different frequency bands. 291 For a simple example, as shown in Firgue 3(a), in order to reflect 292 the state of the spectral resources in the time domain, we use a two- 293 dimensional time spectrum resource model and assume that each time 294 slot has the same time period. The network diagram is illustrated in 295 the figure 3(a). Now there was an AR request, from node A to node D. 296 The request requires two spectrum slots, lasting from T2 to T3, as 297 shown in figure 2(b). Figure 2(c) and Figure 2(d) show the spectrum 298 states of path A-C-D and path A-B-D, respectively. The black slot 299 represents the occupied spectrum slot, the white slot represents the 300 spectrum slot available for the spectrum resource, and the blue slot 301 represents the spectrum slot occupied by the AR request. Before 302 splitting the AR request, the two paths do not have enough resources 303 to allocate. However, after we split the request into two parts, we 304 can distribute them to two spectrum segments to implement AR- 305 requested service provision. The MP-RA is as shown in Algorithm 2. 307 4. Experimental evaluation and results analysis 309 In this paper, we present the results of the AAFS-ANN prediction. 310 Our goal is to demonstrate the accuracy and network performance of 311 AAFS-ANN in different network environments. To fully reflect the 312 changes in the network environment, we use WIDE data from 96h traffic 313 data from April 6th to 9th, 2017, to train and verify. Figure 3(a) 314 is the comparison between the actual traffic and the prediction 315 results, which verify the effectiveness of our method. As shown in 316 Figure 3(a), the prediction results of AAFS-ANN are significantly 317 better than the traditional predictions. This is because the 318 introduction of the adaptive step size and the visible field, making 319 the artificial fish compares the FC in the large field of view. Our 320 method enhances the global convergence and the optimization 321 precision. The prediction error occurs because the traffic is 322 directly affected by many non-linear sudden factors such as hot 323 events, user movement patterns. Therefore, many traffic cannot be 324 accurately predicted. 326 We also compare MP-RA with several state-of-the-art resource 327 allocation techniques including evolutionary algorithms(EA) and 328 artificial neural networks (ANN). From firgue 3(b), we can see that 329 MP-RA performs well among the three optimization resource allocation 330 method, MP-RA greatly improves resource utilization. 332 According to the prediction results, the MP-RA can allocate resources 333 to traffic more reasonably. This is because the algorithm considers 334 the traffic that will be reached at each point in time and the 335 resources it needs. Then re-plans the resources at the configuration 336 time. As can be observed in the results shown in Figure 3(c), MP-RA 337 can greatly reduce the probability of traffic blocking. 339 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 340 | | Resource utilization rate | 341 | Traffic load +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 342 | | MP-RA | ANN | EA | 343 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 344 | 40 | 0.254 | 0.242 | 0.251 | 345 | 70 | 0.263 | 0.253 | 0.272 | 346 | 95 | 0.273 | 0.275 | 0.300 | 347 | 120 | 0.332 | 0.29 | 0.420 | 348 | 145 | 0.389 | 0.325 | 0.504 | 349 | 170 | 0.457 | 0.356 | 0.583 | 350 | 200 | 0.52 | 0.403 | 0.723 | 351 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 353 Tab.1 Network blocking probability of four strategies 354 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 355 | | Network blocking probability | 356 | Traffic load +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 357 | | MP-RA | ANN | EA | 358 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 359 | 50 | 0.008 | 0.0075 | 0.0078 | 360 | 70 | 0.009 | 0.010 | 0.012 | 361 | 100 | 0.0095 | 0.025 | 0.029 | 362 | 125 | 0.01 | 0.06 | 0.074 | 363 | 150 | 0.0108 | 0.08 | 0.10 | 364 | 175 | 0.025 | 0.115 | 0.129 | 365 | 200 | 0.06 | 0.15 | 0.20 | 366 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 368 Tab.2 Average hop of four strategies 370 5. CONCLUSION 372 In the tidal traffic scenario, we propose AAFS-ANN model and MP-RA 373 strategy. We use AAFS-ANN model to predict traffic and MP-RA to 374 optimize metropolitan optical network. Results demonstrate that 375 AAFS-ANN and MP-RA successfully increase prediction accuracy and 376 resource utilization, as well as reduce the traffic blocking rate. 378 6. ACKNOWLEDGMENT 380 This work has been supported in part by NSFC project (61871056), 381 National Postdoctoral Program for Innovative Talents (BX201600021), 382 Fundamental Research Funds for the Central Universities (2018XKJC06) 383 and State Key Laboratory of Information Photonics and Optical 384 Communications (BUPT), P. R. China (No. IPOC2017ZT11). 386 7. References 388 7.1. Normative References 390 [RFC2119] Bradner, S., "Key words for use in RFC's to Indicate 391 Requirement Levels", RFC 2119, March 1997. 393 7.2. Informative References 395 [Ref1] Alvizu, R., Troia, S., and G. Maier, "Matheuristic with 396 machine-learning-based prediction for software-defined 397 mobile metro-core networks", May 2017. 399 [Ref2] Zhong, Z., Hua, N., and H. Liu, "Considerations of 400 effective tidal traffic dispatching in software-defined 401 metro IP over optical networks", July 2015. 403 [Ref3] Yu, A., Yang, H., and W. Bai, "Leveraging deep learning to 404 achieve efficient resource allocation with traffic 405 evaluation in datacenter optical networks", March 2018. 407 [Ref4] Zhong, Z., Hua, N., and M. Tornatore, "Energy efficiency 408 and blocking reduction for tidal traffic via stateful 409 grooming in IP-over-optical networks", September 2016. 411 Authors' Addresses 413 Hui Yang 414 Beijing University of Posts and Telecommunications 415 No.10,Xitucheng Road,Haidian District 416 Beijing 100876 417 P.R.China 419 Phone: +8613466774108 420 Email: yang.hui.y@126.com 421 URI: http://www.bupt.edu.cn/ 423 Kaixuan Zhan 424 Beijing University of Posts and Telecommunications 425 No.10,Xitucheng Road,Haidian District 426 Beijing 100876 427 P.R.China 429 Phone: +8618401695826 430 Email: zhankai@bupt.edu.cn 431 URI: http://www.bupt.edu.cn/ 433 Ao Yu 434 Beijing University of Posts and Telecommunications 435 No.10,Xitucheng Road,Haidian District 436 Beijing 100876 437 P.R.China 439 Email: yuao@bupt.edu.cn 440 URI: http://www.bupt.edu.cn/ 441 Qiuyan Yao 442 Beijing University of Posts and Telecommunications 443 No.10,Xitucheng Road,Haidian District 444 Beijing 100876 445 P.R.China 447 Email: yqy89716@bupt.edu.cn 448 URI: http://www.bupt.edu.cn/ 450 Jie Zhang 451 Beijing University of Posts and Telecommunications 452 No.10,Xitucheng Road,Haidian District 453 Beijing 100876 454 P.R.China 456 Phone: +8613911060930 457 Email: lgr@bupt.edu.cn 458 URI: http://www.bupt.edu.cn/