idnits 2.17.00 (12 Aug 2021) /tmp/idnits14577/draft-fu-coinrg-joint-optimization-req-00.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 : ---------------------------------------------------------------------------- No issues found here. 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 has text resembling RFC 2119 boilerplate text. -- The document date (November 2, 2020) is 558 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'I-D.li-apn6-problem-statement-usecases' is defined on line 277, but no explicit reference was found in the text Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Computing in Network Research Group Y. Fu 3 Internet-Draft P. Liu 4 Intended status: Informational L. Geng 5 Expires: May 6, 2021 China Mobile 6 November 2, 2020 8 Requirements of computing and network joint optimization and scheduling 9 draft-fu-coinrg-joint-optimization-req-00 11 Abstract 13 With the development of edge computing, there is a trend that 14 computing is widely deployed in network rather than at other end of 15 network, and provides services at nearer location. With the deep 16 integration of network, traditional optimization and scheduling 17 within network domain is not enough, the endpoint of the path matters 18 a lot. So the relationship between computing and network are new and 19 important topics to be studied. This document focus on the 20 requirements of computing and network joint optimization and 21 scheduling based on the newly arising service requirements. 23 Requirements Language 25 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 26 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 27 document are to be interpreted as described in . 29 Status of This Memo 31 This Internet-Draft is submitted in full conformance with the 32 provisions of BCP 78 and BCP 79. 34 Internet-Drafts are working documents of the Internet Engineering 35 Task Force (IETF). Note that other groups may also distribute 36 working documents as Internet-Drafts. The list of current Internet- 37 Drafts is at https://datatracker.ietf.org/drafts/current/. 39 Internet-Drafts are draft documents valid for a maximum of six months 40 and may be updated, replaced, or obsoleted by other documents at any 41 time. It is inappropriate to use Internet-Drafts as reference 42 material or to cite them other than as "work in progress." 44 This Internet-Draft will expire on May 6, 2021. 46 Copyright Notice 48 Copyright (c) 2020 IETF Trust and the persons identified as the 49 document authors. All rights reserved. 51 This document is subject to BCP 78 and the IETF Trust's Legal 52 Provisions Relating to IETF Documents 53 (https://trustee.ietf.org/license-info) in effect on the date of 54 publication of this document. Please review these documents 55 carefully, as they describe your rights and restrictions with respect 56 to this document. Code Components extracted from this document must 57 include Simplified BSD License text as described in Section 4.e of 58 the Trust Legal Provisions and are provided without warranty as 59 described in the Simplified BSD License. 61 Table of Contents 63 1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 2 64 2. Requirements of awareness of computing . . . . . . . . . . . 3 65 3. Requirements of computing and network joint optimization . . 4 66 4. Requirements of computing and network joint resource 67 reservation . . . . . . . . . . . . . . . . . . . . . . . . . 5 68 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 6 69 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6 70 7. Security Considerations . . . . . . . . . . . . . . . . . . . 6 71 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 6 72 9. Informative References . . . . . . . . . . . . . . . . . . . 6 73 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 7 75 1. Overview 77 For traditional services without strict service requirements, the 78 best-effort network can meet the requirements with traditional path 79 optimization, which only consider the network condition. With new 80 services arising, such as cloud AR/VR, cloud gaming, V2X, new and 81 strict requirements towards network, also towards the service 82 endpoint are proposed to meet the service requirements. So the 83 computing and network joint optimization and scheduling are proposed 84 to guarantee the service performance. 86 The computing and network joint optimization means that there is not 87 only the path optimization in network, but also the endpoint joint 88 optimization; also the two will affect each other. Based on the 89 joint optimization, the service scheduling can be performed 90 considering the network condition and also the endpoint condition, 91 with the "optimal path+ optimal endpoint" policy. What's more, the 92 computing and network resources joint reservation is required for 93 services with strict performance requirements. 95 2. Requirements of awareness of computing 97 The service requirements arising include both network and computing 98 requirements, which further require future network should perform the 99 joint optimization according to service requirements. So the 100 requirements towards the joint optimization are: the awareness of 101 computing and network requirements, the awareness of computing 102 resources and services in network, the computing-aware path 103 optimization and the network-aware endpoint optimization. 105 2.1. Awareness of computing and network requirements 107 Awareness of computing and network requirements refers to consider 108 the computing requirements in addition to the network requirements, 109 including the awareness of computing requirements and the measurement 110 of computing services. 112 Since network requirements can be measured with bandwidth, delay etc, 113 it is also required to measure computing requirements in a unified 114 way. On the one hand, there are many different computing services 115 which are the "consumers" of computing resources, such as video 116 processing, image classification etc, and they propose various 117 requirements towards computing. It is required to firstly obtain the 118 computing requirements and then model the requirements in a unified 119 way, which then can be used as the constraint of joint optimization. 121 What's more, the computing service modes are abundant compared to 122 network services, which are the computing "producers", including 123 there are heterogeneous hardware such as GPU, CPU, FPGA etc, and also 124 various algorithms deployed in network, so it is also required to 125 model the computing producers in a unified way, which is another 126 important factor for joint optimization. As for the awareness of 127 computing requirements, some technologies such as application-aware 128 networking have proposed corresponding technical solutions to 129 delivery computing requirements in the packet head, however, it needs 130 further study on the security of application and also the efficiency 131 of the information delivery. As for the measurement of computing 132 services, there is no mature solution to model the computing 133 requirements and the computing resources in a unified way, which is a 134 challenge for the computing and network joint optimization. 136 2.2. Awareness of computing resources and services 138 With the development of edge computing, the computing resources and 139 computing services will be distributed in network, since the limited 140 physical conditions, each computing site will be small scale and with 141 limited computing resources, so different from the cloud computing, 142 which can finish the computing task within one site, the edge 143 computing sites need the collaboration among many sites, and this 144 collaboration can be done in network. To coordinate the computing 145 sites, it is required for network to be aware of the computing status 146 of edge sites, including the real-time status of computing resources 147 and computing services. So how to generate the required information 148 and then broadcast it to network brings new challenges. 150 3. Requirements of computing and network joint optimization 152 3.1 Computing-aware path optimization 154 With new services requiring computing and network resources, 155 traditional network-based path optimization can not accurately 156 guarantee the service requirements. The network-based path 157 optimization only according to network conditions can only make sure 158 the performance of network services, it can only find a best path 159 towards a given endpoint, however, the given endpoint may be not 160 optimal, causing the service requirements cannot be met. 162 So It is required to do the computing-aware path optimization to 163 consider the status of endpoint. For example, before the path 164 optimization, according to the awareness of computing resources and 165 services in network, including the location and status, the network 166 could firstly find a list of optimal computing nodes, then the 167 network could do path optimization with different computing 168 endpoints, which changes the traditional way to only do the path 169 optimization with one destination. 171 To better optimize the computing-aware path, we need to consider 172 different weights of computing and network metrics when calculating 173 the optimal path. For traditional path optimization, there are only 174 network metrics as the parameters of algorithm; it is required to add 175 computing metrics also as the calculation metrics of the algorithm 176 and to combine the computing and network metrics. 178 What's more, based on the awareness of service requirements, for 179 different services, there will be different requirements towards 180 computing and network. For some computing-intensive services, 181 computing counts more on the whole process of services, so they will 182 require more on computing than network; and for communication- 183 intensive services, the computing is less during the service process, 184 while there will be frequent communication, which will propose higher 185 requirement towards network than computing. So it can be inferred 186 that computing and network matters differently during the service 187 process for various services. 189 Based on what discussed above, it is required to adaptively define 190 different weights of computing and network metrics for different 191 services, adapting to various service requirements. For example, for 192 the computing-intensive services, it is required to put more weights 193 on computing metrics than network metrics, which could be based on 194 the percentage of predicted computing time in whole time; as for 195 communication-intensive services, more weights could be put to adapt 196 to the service requirements. 198 3.2 Network-aware endpoint optimization 200 Based on the computing-aware path optimization, there will be the 201 optimal "path + endpoint" pair, combing the computing and network 202 status. But there will also be inner scheduling in computing node, 203 which may also influence the computing time. With proper task 204 assignment, the computing time could be less to make sure that 205 endpoint provides the promised services. So it is also required for 206 endpoint to know the service requirements precisely, otherwise the 207 endpoint will just do the usual scheduling without considering the 208 service requirements. 210 With the network-awareness, the endpoint will know the performance of 211 network, such as the endpoint will know the transmission time in 212 network and then calculate the rest of required time, and then it 213 will do the inner scheduling accordingly. 215 4. Requirements of computing and network joint resource reservation 217 For services with strict computing requirements, the resource 218 reservation should include network reservation and computing 219 reservation, also, the two will affect each other. 221 There is network resource reservation in traditional QoS guarantee 222 mechanism based on the network resources reservation calculation to 223 reserve specific resources for specific services. With new services 224 arising, the network resources reservation is not enough, since the 225 completion of services include not only network transmission but also 226 endpoint calculation, only reserving the network resources cannot 227 make sure the required computing resources are available during the 228 required time for specific service. 230 So facing the trend of computing and network convergence, it is also 231 required to reserve the computing resources together with the network 232 resources. Based on the awareness of service requirements and the 233 joint path optimization, it is required to map the computing 234 requirements into the corresponding computing resources reservation, 235 for example, to map the services type into the computing resources 236 type, and translate the computing latency requirements towards the 237 required amount computing resources. 239 On the other hand, the reservation of network and computing resources 240 are closely linked, there will be different network resource 241 reservation policy considering the computing resources reservation. 242 For example, the order of the two resources reservation requires to 243 be considered since they are relative independent usually. 245 What's more, it is also required to dynamically adjust the resources 246 reservation according to real-time status. One scenario is that the 247 computing resource reservation could be adjusted based on the 248 information from network domain, including the reservation time and 249 also the reservation amount. Another scenario is the co-adjust of 250 the two resources reservation, in network domain, the path and the 251 relative reservation could be adjusted, and then the computing domain 252 is required to adjust on-demand. 254 5. Conclusion 256 Based on the new services' requirements on computing and network, 257 this document puts forward requirements of computing and network 258 joint optimization, and also proposes requirements of computing and 259 network joint resource reservation. Computing in network is a new 260 direction, how to collaborate computing and network need further 261 study. 263 6. IANA Considerations 265 TBD. 267 7. Security Considerations 269 TBD. 271 8. Acknowledgements 273 TBD. 275 9. Informative References 277 [I-D.li-apn6-problem-statement-usecases] 278 Li, Z., Peng, S., Voyer, D., Xie, C., Liu, P., Liu, C., 279 Ebisawa, K., Previdi, S., and J. Guichard, "Problem 280 Statement and Use Cases of Application-aware IPv6 281 Networking (APN6)", draft-li-apn6-problem-statement- 282 usecases-01 (work in progress), November 2019. 284 Authors' Addresses 286 Yuexia Fu 287 China Mobile 288 No.32 XuanWuMen West Street 289 Beijing 100053 290 China 292 Email: fuyuexia@chinamobile.com 294 Peng Liu 295 China Mobile 296 No.32 XuanWuMen West Street 297 Beijing 100053 298 China 300 Email: liupengyjy@chinamobile.com 302 Liang Geng 303 China Mobile 304 No.32 XuanWuMen West Street 305 Beijing 100053 306 China 308 Email: gengliang@chinamobile.com