WO2021114402A1 - 防止基于人工智能的传输质量预测失败的保护方法及系统 - Google Patents

防止基于人工智能的传输质量预测失败的保护方法及系统 Download PDF

Info

Publication number
WO2021114402A1
WO2021114402A1 PCT/CN2019/128300 CN2019128300W WO2021114402A1 WO 2021114402 A1 WO2021114402 A1 WO 2021114402A1 CN 2019128300 W CN2019128300 W CN 2019128300W WO 2021114402 A1 WO2021114402 A1 WO 2021114402A1
Authority
WO
WIPO (PCT)
Prior art keywords
margin
optical channel
noise ratio
optical
optical signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2019/128300
Other languages
English (en)
French (fr)
Inventor
沈纲祥
郭宁宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou University
Original Assignee
Suzhou University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to US17/058,869 priority Critical patent/US11469817B2/en
Application filed by Suzhou University filed Critical Suzhou University
Publication of WO2021114402A1 publication Critical patent/WO2021114402A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
    • H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
    • H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
    • H04B10/0795—Performance monitoring; Measurement of transmission parameters
    • H04B10/07953—Monitoring or measuring OSNR, BER or Q
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
    • H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
    • H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
    • H04B10/0791—Fault location on the transmission path
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
    • H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
    • H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
    • H04B10/0793—Network aspects, e.g. central monitoring of transmission parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
    • H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
    • H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
    • H04B10/0795—Performance monitoring; Measurement of transmission parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04J—MULTIPLEX COMMUNICATION
    • H04J14/00—Optical multiplex systems
    • H04J14/02—Wavelength-division multiplex systems
    • H04J14/0287—Protection in WDM systems
    • H04J14/0289—Optical multiplex section protection
    • H04J14/0291—Shared protection at the optical multiplex section (1:1, n:m)

Definitions

  • the present invention relates to the technical field of communication, in particular to a protection method and system for preventing the failure of transmission quality prediction based on artificial intelligence.
  • AI artificial intelligence
  • OSNR optical signal-to-noise ratio
  • the technical problem to be solved by the present invention is to overcome the problem of high QoT prediction failure based on AI in the prior art, leading to instability problems in actual network applications, thereby providing a low prediction failure and stable prevention of actual network applications.
  • Artificial intelligence-based protection method and system for failure of transmission quality prediction is to overcome the problem of high QoT prediction failure based on AI in the prior art, leading to instability problems in actual network applications, thereby providing a low prediction failure and stable prevention of actual network applications.
  • a protection method for preventing the failure of transmission quality prediction based on artificial intelligence of the present invention includes: a method for allocating an optical signal-to-noise ratio margin for a working optical channel and a method for protecting an optical signal-to-noise ratio margin for an optical channel.
  • Method in which the number of first frequency slots meeting the traffic demand and the consumable margin of the working optical channel are calculated according to the method of allocating the optical signal-to-noise ratio margin of the working optical channel; the method of allocating the optical signal-to-noise ratio margin according to the protection optical channel is calculated The number of second frequency slots that meet the traffic demand and the margin that can be consumed by the protection optical channel; the utilization of spectrum resources is evaluated according to the number of first frequency slots and the number of second frequency slots, and the operating optical channel can be used The consumption margin evaluates the reliability of the optical channel.
  • the working optical channel consumable margin and the protective optical channel consumable margin jointly evaluate the service reliability of the optical channel.
  • the number of the first frequency slot and the number of the second frequency slot jointly evaluate the utilization rate of the optical channel.
  • the method for allocating the optical signal-to-noise ratio margin of the working optical channel includes: calculating the optical signal-to-noise ratio of the optical channel by using channel parameters; A random Gaussian error value is added to the link to simulate and generate real actual data; an optical signal-to-noise ratio model is established through artificial neural network training to predict the optical signal-to-noise ratio of the optical channel; and the optical signal-to-noise ratio is calculated to meet the flow rate
  • the required number of first frequency slots and working optical channels can consume margin.
  • the formula for calculating the number of first frequency slots and the working optical channel consumable margin is:
  • FEC limit is the forward error correction coding limit
  • U margin and S margin are unallocated margin and system margin, respectively
  • OSNR lightpath is the optical signal-to-noise ratio value of the optical channel.
  • the channel parameters include the number of hops of the optical channel, the total length of the optical channel, the length of the longest link in the optical channel, the sum of amplified spontaneous emission noise of all optical channels, and all optical channels The sum of the non-linear interference and the number of optical amplifiers on the optical channel.
  • the method for protecting an optical channel to allocate an optical signal-to-noise ratio margin includes: using channel parameters to calculate the optical signal-to-noise ratio of the optical channel; and calculating according to the optical signal-to-noise ratio that satisfies The number of second frequency slots required by the traffic and the margin that can be consumed by the protection optical channel.
  • the formula for calculating the number of second frequency slots and the margin that can be consumed by the protection optical channel is:
  • FEC limit is the forward error correction coding limit
  • U margin , D margin, and S margin are unallocated margin, design margin, and system margin, respectively
  • OSNR lightpath is the optical signal-to-noise ratio value of the optical channel.
  • the present invention also provides a protection system for preventing the failure of transmission quality prediction based on artificial intelligence, including a working optical channel allocation optical signal-to-noise ratio margin system and a protective optical channel allocation optical signal-to-noise ratio margin system, wherein the working optical channel
  • the optical signal-to-noise ratio margin system for channel allocation is used to calculate the first frequency slot number that meets the traffic demand and the working optical channel consumable margin
  • the protective optical channel allocation optical signal-to-noise ratio margin system is used to calculate the flow rate The number of second frequency slots required and the margin that can be consumed by the protection optical channel; the utilization rate of spectrum resources is evaluated according to the number of the first frequency slots and the number of the second frequency slot, and the margin that can be consumed according to the working optical channel Evaluate the reliability of the optical channel.
  • the protection method and system for preventing the failure of transmission quality prediction based on artificial intelligence of the present invention include a method for allocating an optical signal-to-noise ratio margin for a working optical channel and a method for allocating an optical signal-to-noise ratio margin for a protective optical channel, wherein even if it is based on The method of allocating the optical signal-to-noise ratio margin according to the working optical channel predicts failure, then the method of allocating the optical signal-to-noise ratio margin according to the protective optical channel can also ensure that network services are restored in time, wherein the optical signal-to-noise ratio margin is allocated according to the working optical channel.
  • the method of signal-to-noise ratio margin calculates the number of the first frequency slot that meets the traffic demand and the working optical channel consumable margin, and calculates the second frequency slot that meets the traffic demand according to the method of assigning optical signal-to-noise ratio margin to the protection optical channel
  • the number of optical channels and the margin that can be consumed by the protection optical channels are used to evaluate the utilization of spectrum resources according to the number of the first frequency slots and the number of the second frequency slots, and the margins that can be consumed by the working optical channels are used to evaluate optical
  • the reliability of the channel can effectively prevent the prediction failure of the method of assigning the optical signal-to-noise ratio margin of the working optical channel, and make it more stable in actual network applications; in addition, the working optical channel is used to allocate the optical signal-to-noise ratio
  • the margin method has the advantage of reducing the margin of the optical signal-to-noise ratio, ensuring efficient use of spectrum resources.
  • Fig. 1 is a flow chart of the protection method for preventing the failure of transmission quality prediction based on artificial intelligence according to the present invention
  • Fig. 2 is a schematic diagram of setting the margin of the optical signal-to-noise ratio of the distributed optical channel of the present invention
  • Fig. 3 is a schematic diagram of setting the margin of the optical signal-to-noise ratio of the working optical channel distribution of the present invention
  • Figure 4 is the relationship between the consumable margin and the failure probability of the present invention.
  • Figure 5 is a protection example of the present invention based on the AI-based OSNR prediction failure
  • Figure 6 is a diagram of the relationship between frequency slot capacity and FEC limit of the present invention.
  • Figure 7a is the maximum number of frequency slots used in NSFNET of the present invention.
  • Figure 7b is the maximum number of frequency slots used by the present invention under USNET
  • Figure 8a is the reliability analysis of the present invention under NSFNET
  • Figure 8b is the reliability analysis of the present invention under USNET.
  • the present embodiment provides a protection method for preventing the failure of transmission quality prediction based on artificial intelligence, including: a method for assigning an optical signal-to-noise ratio margin for a working optical channel and an optical signal-to-noise ratio margin for protecting an optical channel
  • the method for assigning optical signal-to-noise ratio margin according to the method of working optical channel calculates the number of first frequency slots that meets the traffic demand and the working optical channel consumable margin
  • the method of assigning optical signal-to-noise ratio margin according to the protection optical channel Calculate the number of second frequency slots that meet the traffic demand and the margin that can be consumed by the protection optical channel, evaluate the utilization of spectrum resources according to the number of first frequency slots and the number of second frequency slots, and according to the working optical channel
  • the margin can be consumed to evaluate the reliability of the optical channel.
  • the protection method for preventing the failure of transmission quality prediction based on artificial intelligence in this embodiment includes: a method for allocating an optical signal-to-noise ratio margin for a working optical channel and a method for allocating an optical signal-to-noise ratio margin for a protective optical channel, wherein the working optical channel
  • the method for allocating the optical signal-to-noise ratio margin for the optical channel is a prediction method based on AI. Even if the prediction fails based on the method for allocating the optical signal-to-noise ratio margin according to the working optical channel, then the optical signal-to-noise ratio margin is allocated according to the protection optical channel.
  • the method can also ensure that network services can be restored in time.
  • the number of first frequency slots that meet the traffic demand and the working optical channel consumable margin are calculated according to the protection optical channel.
  • the method of channel allocation optical signal-to-noise ratio margin calculates the number of second frequency slots that meet the traffic demand and the margin that can be consumed by the protection optical channel, and evaluates the spectrum resource according to the number of the first frequency slot and the number of the second frequency slot
  • the reliability of the optical channel can be evaluated according to the consumption margin of the working optical channel, which can effectively prevent the failure of the method of assigning the optical signal-to-noise ratio margin of the working optical channel to the actual network. It is more stable in application; in addition, the method of using the working optical channel to allocate the optical signal-to-noise ratio margin has the advantage of reducing the optical signal-to-noise ratio margin, and ensures efficient use of spectrum resources.
  • the consumable margin of the working optical channel and the margin consumable of the protection optical channel jointly evaluate the service reliability of the optical channel; the number of the first frequency slot and the number of the second frequency slot are jointly evaluated.
  • the utilization of optical channels the following describes in detail how to evaluate the reliability and utilization of optical channel services:
  • OSNR optical signal-to-noise ratio
  • ASE the amplified spontaneous emission noise
  • NLI non-linear interference
  • the present invention uses a gain-noise figure (NF) table specially established in advance for the amplifier type to find the NF corresponding to the amplifier gain; for NLI, the present invention uses the closed form of the incoherent Gaussian noise (IGN) model Approximate solution.
  • IGN incoherent Gaussian noise
  • three OSNR margins are usually set for the optical channel: system margin, unallocated margin and design margin.
  • the system margin describes the network operating conditions, including penalties for rapid signal changes, slow aging and nonlinear effects; the unallocated margin describes the capacity difference between signal requirements and network equipment; the design margin describes the evaluation The accuracy of QoT's design tools.
  • the method for allocating the margin of the optical signal-to-noise ratio for the protection optical channel (referred to as the traditional method) is described in detail below:
  • the optical signal-to-noise ratio of the optical channel is calculated by using channel parameters; the number of second frequency slots meeting the traffic demand and the margin that can be consumed by the protection optical channel are calculated according to the optical signal-to-noise ratio.
  • both system margin and design margin need to be considered.
  • the non-linear OSNR margin is 2dB
  • the margin for the slow aging of the fiber optic equipment is 2.3dB
  • the influence of the fast time-varying signal is 0.4dB.
  • the design margin at this time is 2dB.
  • the required OSNR of each optical channel can be calculated, and then a modulation format can be selected according to the formula. Specifically, the calculation of the second frequency slot number and the margin that can be consumed by the protection optical channel can be calculated.
  • the formula is:
  • FEC limit is the forward error correction coding limit
  • U margin , D margin, and S margin are unallocated margin, design margin, and system margin, respectively
  • OSNR lightpath is the optical signal-to-noise ratio value of the optical channel.
  • the optical signal-to-noise ratio of the optical channel is calculated by using channel parameters; a random Gaussian error value is added for each link in the signal transmission process, and real actual data is simulated and generated; the optical signal is established through artificial neural network training
  • the noise ratio model predicts the optical signal-to-noise ratio of the optical channel; according to the optical signal-to-noise ratio, the number of first frequency slots meeting the traffic demand and the working optical channel consumption margin are calculated.
  • the system margin at the initial stage of the network life is still set to 4.7dB, as shown in Figure 3. Because the AI-based optical signal-to-noise ratio prediction can more accurately evaluate the QoT of the optical channel, the design margin can be significantly reduced. It is not even necessary to set a design margin. In this study, this application sets the design margin to 0. In order to accurately predict the optical signal-to-noise ratio, sufficient AI training data is required. Since it is difficult to collect enough real network data, this application simulates the data by calculating the optical signal-to-noise ratio of all routes in the Elastic Optical Network (EON).
  • EON Elastic Optical Network
  • this application adds a Gaussian error with a maximum value of ⁇ 0.3dB [3] to the non-linear interference (NLI).
  • ANN artificial neural network
  • the simulation of the above training data will not affect the effectiveness of the method proposed in this application, because if enough real data can be collected to replace these simulated data, no other steps need to be modified.
  • the artificial neural network (ANN) used for training has multiple input neurons, that is, channel parameters.
  • the channel parameters include the number of hops of the optical channel, the total length of the optical channel, and the length of the longest link in the optical channel. , The sum of amplified spontaneous emission noise of all optical channels, the sum of nonlinear interference of all optical channels, and the number of optical amplifiers on the optical channel.
  • an AI-based OSNR prediction model After training, an AI-based OSNR prediction model can be obtained. According to the OSNR predicted by the model, while ignoring the value of the design margin (2dB), the following formula is used to select the most efficient modulation format. Specifically, the formula for calculating the number of first frequency slots and the margin of consumption of the working optical channel is:
  • FEC limit is the forward error correction coding limit
  • U margin and S margin are unallocated margin and system margin, respectively
  • OSNR lightpath is the optical signal-to-noise ratio value of the optical channel.
  • the actual optical signal-to-noise ratio is the sum of the forward error correction coding (FEC) limit and the total margin setting including the three margins.
  • FEC forward error correction coding
  • the optical channel has the best transmission quality and the lowest failure rate, which is close to zero.
  • Ether end of life stage of the optical channel, when the OSNR value is lower than the sum of the FEC limit and the system margin of the fast time-varying influence, the optical channel service will fail, and the failure rate is 1.
  • the difference between the margins of the BoL phase (the initial phase of network life) and the EoL phase (the end phase of network life) is defined as the margin (CM) that the optical channel can consume, that is, the total optical path margin T margin minus
  • CM the consumable margin
  • the transmission failure rate of the optical channel is estimated according to the Gaussian distribution shown in Fig. 4, where the shaded area is the integral of the probability density function of the failure rate.
  • the frequency slot capacity corresponds to 75GB/s, and at this time the node AC
  • the AI-based method predicts the OSNR value of the optical channel, and calculates the optical signal-to-noise ratio of the working optical channel to be 21.03Db according to the channel parameters.
  • this application proposes to set up a protection optical channel based on 1+1 protection in addition to the working optical channel.
  • the traditional method is used to set the OSNR margin for protecting the optical channel. This method makes full use of the high frequency spectrum utilization brought by the more accurate OSNR prediction on the working optical channel. At the same time, even if the prediction model fails, business services will not be interrupted due to a reliable protection optical channel.
  • the protection optical channel is set to the routing AEDC, and the OSNR value of this optical channel is 18.2dB.
  • the traditional method is used to set the margin. Therefore, FEC limit + U margin ⁇ 11.5 dB can be obtained. Since the range of FEC limit is [8.38, 12.43] dB, QPSK is selected as the modulation format. At this time, 4 frequency slots need to be reserved on link AE, link ED, and link DC. Under this setting condition, the unallocated margin of the protection optical channel is 3.12dB, which corresponds to a CM value of 9.42dB. It can be seen from Figure 4 that the failure rate of the protection optical channel is close to zero.
  • this application considers two test networks, one is the NSFNET with 14 nodes and 21 links, and the other is the USNET with 24 nodes and 43 links.
  • the service requirements between each node pair are evenly distributed in the range of [100, X] Gb/s, where X is the maximum required service capacity, and the NSFNET and USNET networks are 600 and 200, respectively.
  • the bandwidth granularity of each frequency slot is 12.5 GHz, as shown in FIG. 6, there are 6 modulation formats (namely: BPSK, QPSK, 8-QAM, 16-QAM, 32-QAM, and 64-QAM).
  • the shortest path algorithm is used to set the working optical channel;
  • the protection optical channel is adaptively selected based on the most efficient spectrum window plane (SWP) algorithm.
  • SWP spectrum window plane
  • the frequency slot used in this case The least number.
  • the number of frequency slots used by "W:AI_P:traditional" is in the middle of the results of the other two cases.
  • the difference in the number of frequency slots used by them is very small, that is, in the most conservative case of NSFNET and USNET, "W:AI_P:traditional", “W:AI_P
  • the reduction of "AI” is 10.2%, 14.6% and 10.2%, 14.6% respectively.
  • this application evaluates the reliability of the optical channel in the above three situations.
  • the reliability of all optical channel services is averaged as the final result as shown in Figure 8a and Figure 8b. Because “W:traditional_P:traditional” adopts the most conservative method, it has the highest reliability. On the contrary, "W:AI_P:AI” adopts the most advanced prediction method and sets the lowest margin for the working optical channel and the protection optical channel, so the reliability is the lowest.
  • this embodiment provides a protection system for preventing the failure of transmission quality prediction based on artificial intelligence.
  • the principle of solving the problem is similar to the protection method for preventing the failure of transmission quality prediction based on artificial intelligence. No longer.
  • the protection system for preventing the failure of transmission quality prediction based on artificial intelligence described in this embodiment includes a working optical channel allocation optical signal-to-noise ratio margin system and a protective optical channel allocation optical signal-to-noise ratio margin system, wherein
  • the working optical channel allocation optical signal-to-noise ratio margin system is used to calculate the number of first frequency slots that meet the traffic demand and the working optical channel consumable margin;
  • the protection optical channel allocation optical signal-to-noise ratio margin system is used to calculate the number of second frequency slots that meet the traffic demand and the margin that the protection optical channel can consume,
  • the utilization rate of spectrum resources is evaluated according to the number of the first frequency slots and the number of the second frequency slots, and the reliability of the optical channel is evaluated according to the working optical channel consumption margin.
  • These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device.
  • the device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment.
  • the instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Optical Communication System (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明涉及一种防止基于人工智能的传输质量预测失败的保护方法,包括：工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度；根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。本发明在实际的网络应用时更加稳定。

Description

防止基于人工智能的传输质量预测失败的保护方法及系统 技术领域
本发明涉及通信的技术领域，尤其是指一种防止基于人工智能的传输质量预测失败的保护方法及系统。
背景技术
在预测光通道传输质量(QoT)的过程中，采用人工智能(AI)技术可以获得相对于传统方法更精确的预测，因此可以降低光网络中光通道的光信噪比(OSNR)裕度的设置，从而有效提高网络资源利用率。但是，许多基于AI预测的方法是仅基于实验室测试的数据，是处于稳定的已知环境中，因此存在基于AI的QoT预测失败高的风险，导致在实际的网络应用中存在不稳定的问题。
发明内容
为此，本发明所要解决的技术问题在于克服现有技术中基于AI的QoT预测失败高，导致实际的网络应用中存在不稳定的问题，从而提供一种预测失败低，实际网络应用稳定的防止基于人工智能的传输质量预测失败的保护方法及系统。
为解决上述技术问题，本发明的一种防止基于人工智能的传输质量预测失败的保护方法，包括：工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度；根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
在本发明的一个实施例中，所述工作光通道可消耗裕度和所述保护光通道可消耗的裕度共同评估出光通道业务可靠性。
在本发明的一个实施例中，所述第一频隙数目和所述第二频隙数目共同评估光通道的利用率。
在本发明的一个实施例中，所述光通道业务可靠性的计算公式：R lp＝1-F w·F p，其中F w和F p分别是工作光通道的失败率和保护光通道的失败率。
在本发明的一个实施例中，所述工作光通道分配光信噪比裕度的方法包括：利用信道参数计算出所述光通道的光信噪比；在信号传输的过程中，每经过一个链路增加一个随机高斯分别的误差值，模拟产生真实的实际数据；通过人工神经网络训练建立光信噪比模型，预测出光通道的光信噪比；根据所述光信噪比计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度。
在本发明的一个实施例中，所述计算第一频隙数目以及工作光通道可消耗裕度的公式为：
FEC limit+U margin≤OSNR lightpath-S margin＝OSNR lightpath-4.7dB
U margin≤OSNR lightpath-4.7dB-FEC limit
其中，FEC limit是前向纠错编码限制，U margin、S margin分别是未分配裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
在本发明的一个实施例中，所述信道参数包括光通道的跳数、光通道的总长度、光通道中最长链路的长度、所有光通道的放大自发辐射噪声的和、所有光通道的非线性干扰的和、光通道上的光放大器的数量。
在本发明的一个实施例中，所述保护光通道分配光信噪比裕度的方法包括：利用信道参数计算出所述光通道的光信噪比；根据所述光信噪比计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度。
在本发明的一个实施例中，所述计算第二频隙数目以及保护光通道可消 耗的裕度的公式为：
FEC limit+U margin≤OSNR lightpath-D margin-S margin＝OSNR lightpath-6.7dB
U margin≤OSNR lightpath-6.7dB-FEC limit
其中，FEC limit是前向纠错编码限制，U margin、D margin和S margin分别是未分配裕度、设计裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
本发明还提供了一种防止基于人工智能的传输质量预测失败的保护系统，包括工作光通道分配光信噪比裕度系统以及保护光通道分配光信噪比裕度系统，其中所述工作光通道分配光信噪比裕度系统用于计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；所述保护光通道分配光信噪比裕度系统用于计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度；根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
本发明的上述技术方案相比现有技术具有以下优点：
本发明所述的防止基于人工智能的传输质量预测失败的保护方法及系统，包括工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中即使基于根据工作光通道分配光信噪比裕度的方法预测失败，那么根据所述保护光通道分配光信噪比裕度的方法，也可以确保网络业务得到及时地恢复，其中根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度，根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度，根据所述第一频隙数目以及所述第二频隙数目用于评估频谱资源的利用率，根据所述工作光通道可消耗裕度用于评估光通道的可靠性，从而可以有效的防止所述工作光通道分配光信噪比裕度的方法预测失败，并使其在实际的网络应用时更加稳定；另外，采用工作光通道分配光信噪比裕度的方法具有光信噪比裕度减少的优势，保证在频谱资源使用上是高效的。
附图说明
为了使本发明的内容更容易被清楚的理解，下面根据本发明的具体实施例并结合附图，对本发明作进一步详细的说明，其中
图1是本发明防止基于人工智能的传输质量预测失败的保护方法流程图；
图2是本发明保护光通道分配光信噪比裕度的设置示意图；
图3是本发明工作光通道分配光信噪比裕度的设置示意图；
图4是本发明可消耗的裕度与失败概率的关系；
图5是本发明基于AI的OSNR预测失效的一个保护实例；
图6是本发明频隙容量与FEC限制的关系图；
图7a是本发明在NSFNET下最大使用的频隙数目；
图7b是本发明在USNET下最大使用的频隙数目；
图8a是本发明在NSFNET下的可靠性分析；
图8b是本发明在USNET下的可靠性分析。
具体实施方式
实施例一
如图1所示，本实施例提供一种防止基于人工智能的传输质量预测失败的保护方法，包括：工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度，根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度，根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
本实施例所述防止基于人工智能的传输质量预测失败的保护方法，包括：工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中所述工作光通道分配光信噪比裕度的方法是基于AI的预测方法， 即使基于根据工作光通道分配光信噪比裕度的方法预测失败，那么根据所述保护光通道分配光信噪比裕度的方法，也可以确保网络业务得到及时地恢复，其中根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度，根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度，根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性，从而可以有效的防止所述工作光通道分配光信噪比裕度的方法预测失败，并使其在实际的网络应用时更加稳定；另外，采用工作光通道分配光信噪比裕度的方法具有光信噪比裕度减少的优势，保证在频谱资源使用上是高效的。
在本发明中，所述工作光通道可消耗裕度和所述保护光通道可消耗的裕度共同评估光通道业务可靠性；所述第一频隙数目和所述第二频隙数目共同评估光通道的利用率，下面详细说明如何评估光通道业务可靠性和利用率：
在计算计算光通道的光信噪比(OSNR)时，通常考虑两个方面：光放大器的放大自发辐射噪声(ASE)和非线性干扰(NLI)。对于ASE噪声，本发明通过一个针对放大器类型专门提前建立的增益-噪声系数(NF)表来找到放大器增益对应的NF；对于NLI，本发明采用了非相干高斯噪声(IGN)模型的封闭形式的近似解。根据计算出的光信噪比，通常为光通道设置三种OSNR裕度：系统裕度，未分配裕度和设计裕度。其中系统裕度说明了网络运行条件，包括信号快速变化的惩罚，缓慢老化和非线性的影响；未分配裕度说明了信号需求与网络设备之间的容量差异；设计裕度说明了用于评估QoT的设计工具的准确性。
下面具体说明所述保护光通道分配光信噪比裕度的方法(简称传统方法)：
利用信道参数计算出所述光通道的光信噪比；根据所述光信噪比计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度。
上述方法中，需要同时考虑系统裕度和设计裕度。如图2所示，假设在 网络生命的初始阶段(BoL)，非线性的OSNR裕度为2dB，光纤设备缓慢老化的裕度为2.3dB，信号快速时变的影响为0.4dB，此时表明系统裕度为2+2.3+0.4＝4.7dB，同时假设此时的设计裕度为2dB。通过上述对裕度的假设，可以计算出每个光通道所需的OSNR，再根据公式来选择一个调制格式，具体地，所述计算第二频隙数目以及保护光通道可消耗的裕度的公式为：
FEC limit+U margin≤OSNR lightpath-D margin-S margin＝OSNR lightpath-6.7dB   (1)
U margin≤OSNR lightpath-6.7dB-FEC limit      (2)
其中，FEC limit是前向纠错编码限制，U margin、D margin和S margin分别是未分配裕度、设计裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
下面具体说明所述工作光通道分配光信噪比裕度的方法：(简称基于AI的预测方法)
利用信道参数计算出所述光通道的光信噪比；在信号传输的过程中每经过一个链路增加一个随机高斯分别的误差值，模拟产生真实的实际数据；通过人工神经网络训练建立光信噪比模型，预测出光通道的光信噪比；根据所述光信噪比计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度。
在上述方法中，网络生命初始阶段的系统裕度仍然设置为4.7dB，如图3所示，由于基于AI的光信噪比预测可以更准确地评估光通道的QoT，所以可以显著降低设计裕度，甚至是不需要设置设计裕度。在本研究中，本申请将设计裕度设置为0。为了准确地预测光信噪比，需要足够的AI训练数据。由于很难收集到足够的真实网络数据，所以本申请通过计算弹性光网络(EON)中所有路由的光信噪比来模拟数据。为了模拟网络的统计特性，本申请在非线性干扰(NLI)中加入了一个最大值是±0.3dB[3]的高斯误差。通过这种方法，计算网络中每个节点对间光通道的OSNR，并使用这些值作为人工神经网络(ANN)的训练数据。上述训练数据的模拟并不会影响本申请所提出方法的有效性，因为如果能够收集到足够的真实数据来代替这些模拟数据，那么不需要修改任何其它步骤。另外，用于训练的人工神经网络 (ANN)有多个输入神经元，也就是信道参数，所述信道参数包括光通道的跳数、光通道的总长度、光通道中最长链路的长度、所有光通道的放大自发辐射噪声的和、所有光通道的非线性干扰的和、光通道上的光放大器的数量。
经过训练，可以得到了一个基于AI的OSNR预测模型，根据所述模型预测的OSNR，同时忽略设计裕度的值(2dB)，使用下面的公式选择最高效的调制格式。具体地，所述计算第一频隙数目以及工作光通道可消耗裕度的公式为：
FEC limit+U margin≤OSNR lightpath-S margin＝OSNR lightpath-4.7dB      (3)
U margin≤OSNR lightpath-4.7dB-FEC limit       (4)
其中，FEC limit是前向纠错编码限制，U margin、S margin分别是未分配裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
在光通道的生命初始阶段，实际的光信噪比是前向纠错编码(FEC)限制和包括三种裕度的总的裕度设置的总和。此时光通道传输质量最优，失败率最低，接近于0。在光通道的生命结束(EoL)阶段，当OSNR值低于FEC限制和快速时变影响的系统裕度的总和时，光通道服务将会失效，此时失败率为1。本申请中，定义BoL阶段(网络生命的初始阶段)和EoL阶段(网络生命的结束阶段)裕度的差值为光通道可消耗的裕度(CM)，也就是光路总裕度T margin减去快速时变的影响F margin，也就是CM＝T margin-F margin。根据上述可以计算出的光通道可消耗的裕度(CM)值，然后根据图4所示的高斯分布来估计光通道的传输失败率，其中阴影区域是失败率的概率密度函数的积分。
如图5所示，假设节点对A-C之间的流量需求为180Gb/s，先找到最短路径A-B-C作为工作光通道。首先根据信道参数计算得到保护光通道的光信噪比值为21.1dB，在传统的裕度设置方法下，系统裕度和设计裕度都需要被考虑在内，共为4.7+2＝6.7dB。根据公式(1)，可以计算出FEC limit+U margin≤14.4dB(其中21.1-6.7＝14.4dB)；再根据图6所示，由于FEC被限制范围在[12.43，15.13]dB，因此对应选择8-QAM作为调制格式。在这 种调制格式下，链路A-B和B-C需要预留3个频隙来满足流量需求(其中，FEC限制为8-QAM时，所述频隙容量对应为75GB/s，此时节点A-C之间的流量需求除以频隙容量对应的数值就是需要预留的频隙值，即180/75＝2.4，近似为3个频隙)，再根据公式(2)未分配裕度可以进一步计算为14.4-12.43＝1.97dB，由于信号快速时变的影响为：0.4dB，再根据图2对应计算得到可消耗的裕度CM值为2+1.97+4.7-0.4＝8.27dB。因此，根据图4的失效曲线，在网络生命初始阶段阶段(BoL)，此时的工作光通道的失败率接近为0。
而基于AI的方法预测光通道的OSNR值，根据信道参数计算得到工作光通道的光信噪比值为21.03Db。此时，在不考虑设计裕度的情况下，根据公式(3)，可以计算出FEC limit+U margin≤16.33dB(此时21.03-4.7＝16.33dB)；接下根据图6所示，由于FEC限制范围在[15.13，18.11]dB，因此对应选择16-QAM作为调制格式。使用这种调制格式，应该在链路A-B和链路B-C上预留2个频隙(其中，FEC限制为16-QAM时，所述频隙容量对应为100GB/s，此时节点A-C之间的流量需求除以频隙容量对应的数值就是需要预留的频隙值，即180/100＝1.8，近似为2个频隙)。在这种设置条件下，未分配裕度进一步计算为16.33-15.13＝1.2dB，由于信号快速时变的影响为：0.4dB，再根据图3对应计算得到可消耗的裕度CM值为1.2+4.7-0.4＝5.5dB。因此，根据图4的失效曲线，在网络生命初始阶段阶段(BoL)，此时的工作光通道的失败率接近于7％。
由上述可知：虽然基于AI的预测方法可以使用更高效的调制格式，但是此方法的失败率要远远高于传统方法。因此，为了提高光通道业务服务的可靠性，本申请提出在工作光通道之外，基于1+1保护设置一条保护光通道。为了保证在AI预测失败时，光通道服务能够完全恢复，从而利用传统的方法来设置保护光通道的OSNR裕度。这种方法充分利用了工作光通道上更精确的OSNR预测带来地频谱高利用率，同时即使预测模型失效，由于有一条可靠的保护光通道，业务服务也不会中断。在图5中，所述保护光通道设置为路由A-E-D-C上，这条光通道的OSNR值为18.2dB，此时采用传统方法来 设置裕度。因此，可以得到FEC limit+U margin≤11.5dB，由于FEC限制的范围是[8.38，12.43]dB，因此选择QPSK作为调制格式。此时，链路A-E，链路E-D和链路D-C上需要预留4个频隙。在这种设置条件下，保护光通道的未分配裕度为3.12dB，对应于9.42dB的CM值，从图4可以看出，保护光通道的失败率接近为0。
为了评估提出的保护机制的性能，本申请考虑了两个测试网络，一个是14节点和21链路的NSFNET，另一个是24节点和43链路的USNET。假设每个节点对之间的业务需求均匀分布在[100，X]Gb/s范围内，其中X是最大所需业务容量，NSFNET和USNET网络分别是600和200。假设每个频隙的带宽粒度是12.5GHz，如图6所示有6种调制格式(即：BPSK，QPSK，8-QAM，16-QAM，32-QAM，和64-QAM)。
首先，使用最短路径算法来设置工作光通道；其次，保护光通道采用基于最高效频谱窗平面(SWP)的算法来自适应选择。本申请考虑了建立保护光通道的三种不同的情形，第一个情形假设工作光通道和保护光通道都采用传统方法设置OSNR裕度，第二个情形假设工作光通道是采用基于AI的方法设置裕度，保护光通道是采用传统的方法设置裕度，第三个情形假设工作光通道和保护光通道都采用基于AI的方法设置裕度。
在本申请中，考虑在满足所有流量需求后所使用的频隙数量，结果如图7a和图7b所示，其中“W”表示工作光通道，“P”表示保护光通道，因此“W:AI_P：traditional”表示工作光通道采用基于AI的OSNR裕度设置方法，保护光通道采用传统的裕度设置方法。由于“W：traditional_P:traditional”设置OSNR裕度的方法都是采用传统的方法，因此需要更多的频谱资源才能满足流量需求，与其他两个情形相比，这种情形使用的频隙数最多。相比之下，由于“W:AI_P:AI”工作光通道和保护光通道都是采用基于AI预测的裕度设置方法，所以可以实现更高效的频谱资源分配，因此这种情形使用的频隙数最少。作为中间情形，“W:AI_P:traditional”使用的频隙数量处于其他两种情形结果中间。但是，与最有效的“W:AI_P:AI”情形相比，它们使用的频隙数量差异很小，即在NSFNET和USNET最保守的 情况下，“W:AI_P:traditional”、“W:AI_P:AI”的减少量分别是10.2％、14.6％和10.2％、14.6％。
另外，本申请评估了以上三种情形下光通道的可靠性。光通道业务可靠性R lp的计算公式为R lp＝1-F w·F p，其中F w和F p分别是工作光通道和保护光通道的失败率。在申请中，对所有光通道服务的可靠性取平均作为最终结果如图8a和图8b所示。因为“W:traditional_P:traditional”采用了最保守的方法，所以可靠性最高。相反，“W:AI_P:AI”采用了最先进的预测方法，为工作光通道和保护光通道都设置了最低的裕度，所以可靠性最低。由于“W:AI_P:traditional”的工作光通道采用基于AI的方法，保护光通道采用传统的方法，可靠性处于中间结果。此外，有趣的是，虽然“W:AI_P:traditional”和“W:AI_P:AI”所使用的频隙数量非常接近，但前者的可靠性远远高于后者，在NSFNET中，相对于“W:traditional_P:traditional”，“W:AI_P:AI”的可靠性降低高达87％，而“W:AI_P:traditional”的可靠性降低仅为17％，在USNET中，“W:AI_P:AI”的可靠性降低高达82％，而“W:AI_P:traditional”的可靠性降低仅为24％。因此，本申请认为，在不会大幅度地增加频谱资源使用的情况下，所提出的防止基于AI的QoT预测失败的方法显著地提高了光路服务的可靠性。
实施例二
基于同一发明构思，本实施例提供了一种防止基于人工智能的传输质量预测失败的保护系统，其解决问题的原理与所述防止基于人工智能的传输质量预测失败的保护方法类似，重复之处不再赘述。
本实施例所述的防止基于人工智能的传输质量预测失败的保护系统，包括工作光通道分配光信噪比裕度系统以及保护光通道分配光信噪比裕度系统，其中
所述工作光通道分配光信噪比裕度系统用于计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；
所述保护光通道分配光信噪比裕度系统用于计算出满足流量需求的第 二频隙数目以及保护光通道可消耗的裕度，
根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然，上述实施例仅仅是为清楚地说明所作的举例，并非对实施方式的限定。对于所属领域的普通技术人员来说，在上述说明的基础上还可以做出其它不同形式变化或变动。这里无需也无法对所有的实施方式予以穷举。而由此所引伸出的显而易见的变化或变动仍处于本发明创造的保护范围之中。

Claims (10)

  1. 一种防止基于人工智能的传输质量预测失败的保护方法，其特征在于，包括：工作光通道分配光信噪比裕度的方法和保护光通道分配光信噪比裕度的方法，其中
    根据工作光通道分配光信噪比裕度的方法计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；
    根据保护光通道分配光信噪比裕度的方法计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度；
    根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
  2. 根据权利要求1所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述工作光通道可消耗裕度和所述保护光通道可消耗的裕度共同评估出光通道业务可靠性。
  3. 根据权利要求1所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述第一频隙数目和所述第二频隙数目共同评估光通道的利用率。
  4. 根据权利要求1所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述光通道业务可靠性的计算公式：R lp＝1-F w·F p，其中F w和F p分别是工作光通道的失败率和保护光通道的失败率。
  5. 根据权利要求1所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述工作光通道分配光信噪比裕度的方法包括：利用信道参数计算出所述光通道的光信噪比；在信号传输的过程中，每经过一个链路增加一个随机高斯分别的误差值，模拟产生真实的实际数据；通过人工神经网络训练建立光信噪比模型，预测出光通道的光信噪比；根据所述光信噪比计算 出满足流量需求的第一频隙数目以及工作光通道可消耗裕度。
  6. 根据权利要求1或5所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述计算第一频隙数目以及工作光通道可消耗裕度的公式为：
    FEC limit+U margin≤OSNR lightpath-S margin＝OSNR lightpath-4.7dB
    U margin≤OSNR lightpath-4.7dB-FEC limit
    其中，FEC limit是前向纠错编码限制，U margin、S margin分别是未分配裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
  7. 根据权利要求5所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述信道参数包括光通道的跳数、光通道的总长度、光通道中最长链路的长度、所有光通道的放大自发辐射噪声的和、所有光通道的非线性干扰的和、光通道上的光放大器的数量。
  8. 根据权利要求1所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述保护光通道分配光信噪比裕度的方法包括：利用信道参数计算出所述光通道的光信噪比；根据所述光信噪比计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度。
  9. 根据权利要求1或8所述的防止基于人工智能的传输质量预测失败的保护方法，其特征在于：所述计算第二频隙数目以及保护光通道可消耗的裕度的公式为：
    FEC limit+U margin≤OSNR lightpath-D margin-S margin＝OSNR lightpath-6.7dB
    U margin≤OSNR lightpath-6.7dB-FEC limit
    其中，FEC limit是前向纠错编码限制，U margin、D margin和S margin分别是未分配裕度、设计裕度和系统裕度，OSNR lightpath是光通道的光信噪比值。
  10. 一种防止基于人工智能的传输质量预测失败的保护系统，其特征在 于：包括工作光通道分配光信噪比裕度系统以及保护光通道分配光信噪比裕度系统，其中
    所述工作光通道分配光信噪比裕度系统用于计算出满足流量需求的第一频隙数目以及工作光通道可消耗裕度；
    所述保护光通道分配光信噪比裕度系统用于计算出满足流量需求的第二频隙数目以及保护光通道可消耗的裕度；
    根据所述第一频隙数目以及所述第二频隙数目评估频谱资源的利用率，根据所述工作光通道可消耗裕度评估光通道的可靠性。
PCT/CN2019/128300 2019-12-13 2019-12-25 防止基于人工智能的传输质量预测失败的保护方法及系统 Ceased WO2021114402A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/058,869 US11469817B2 (en) 2019-12-13 2019-12-15 Protection method and system against failure of AI-based QoT prediction

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911282911.7 2019-12-13
CN201911282911.7A CN111049579B (zh) 2019-12-13 2019-12-13 防止基于人工智能的传输质量预测失败的保护方法及系统

Publications (1)

Publication Number Publication Date
WO2021114402A1 true WO2021114402A1 (zh) 2021-06-17

Family

ID=70236136

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/128300 Ceased WO2021114402A1 (zh) 2019-12-13 2019-12-25 防止基于人工智能的传输质量预测失败的保护方法及系统

Country Status (3)

Country Link
US (1) US11469817B2 (zh)
CN (1) CN111049579B (zh)
WO (1) WO2021114402A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115706977B (zh) * 2021-08-03 2025-11-04 华为技术有限公司 一种数据传输方法及相关设备
CN116939713A (zh) * 2022-04-02 2023-10-24 维沃移动通信有限公司 信息反馈方法、装置及设备
CN114584871B (zh) * 2022-04-28 2022-08-05 华南师范大学 弹性光网络的频谱分配方法、装置、存储介质和设备
CN116248184B (zh) * 2022-12-28 2025-08-19 重庆邮电大学 一种基于空间覆盖的hpgs 16qam通信优化方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130044621A1 (en) * 2011-08-16 2013-02-21 Motorola Mobility Llc Self-interference handling in a wireless communication terminal supporting carrier aggregation
CN107623548A (zh) * 2017-10-16 2018-01-23 电子科技大学 一种认知光网络中主从信号传输方法
CN108966687A (zh) * 2017-03-17 2018-12-07 骁阳网络有限公司 在真实网络条件下确定通道osnr和通道osnr裕度
US10177843B2 (en) * 2016-04-19 2019-01-08 Fujitsu Limited Network control apparatus and transmission quality margin calculation method

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2745441A1 (en) * 2011-09-08 2014-06-25 Telefonaktiebolaget L M Ericsson (publ) Path computation in wavelength switched optical networks
EP2868013B1 (en) * 2012-07-02 2016-06-29 Telefonaktiebolaget LM Ericsson (publ) Method and apparatus for configuring an optical path
CN102946275B (zh) * 2012-10-29 2016-02-24 中兴通讯股份有限公司 一种实现高速dwdm系统中osnr监测的方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130044621A1 (en) * 2011-08-16 2013-02-21 Motorola Mobility Llc Self-interference handling in a wireless communication terminal supporting carrier aggregation
US10177843B2 (en) * 2016-04-19 2019-01-08 Fujitsu Limited Network control apparatus and transmission quality margin calculation method
CN108966687A (zh) * 2017-03-17 2018-12-07 骁阳网络有限公司 在真实网络条件下确定通道osnr和通道osnr裕度
CN107623548A (zh) * 2017-10-16 2018-01-23 电子科技大学 一种认知光网络中主从信号传输方法

Also Published As

Publication number Publication date
US20210367666A1 (en) 2021-11-25
CN111049579B (zh) 2021-05-04
CN111049579A (zh) 2020-04-21
US11469817B2 (en) 2022-10-11

Similar Documents

Publication Publication Date Title
WO2021114402A1 (zh) 防止基于人工智能的传输质量预测失败的保护方法及系统
CN113067661B (zh) 光通道性能保证下的osnr感知频谱分配方法及系统
Fallahpour et al. Energy efficient routing and spectrum assignment with regenerator placement in elastic optical networks
CN113114365B (zh) 一种光纤部署方法、存储介质、电子设备及系统
Ravipudi et al. Impairment-and fragmentation-aware dynamic routing, modulation and spectrum allocation in C+ L band elastic optical networks using Q-learning
Zhao et al. Nonlinear impairment aware resource allocation in elastic optical networks
Vale et al. Power, routing, Modulation Level and Spectrum Assignment in all-optical and elastic networks
Cavalcante et al. Optimizing the cost function of power series routing algorithm for transparent elastic optical networks
CN119676596A (zh) C+l光网络自适应信噪比的智能化频谱分配方法及系统
Arpanaei et al. A comparative study on routing selection algorithms for dynamic planning of EONs over C+ L bands
CN102523068A (zh) 弹性光网络中的频谱资源分配方法及系统
Yan et al. DRL-based fragmentation-and impairment-aware resource allocation algorithm in C+ L band elastic optical networks
Teng et al. DRL-assisted dynamic QoT-aware service provisioning in multi-band elastic optical networks
Guo et al. What if AI fails: protection against failure of AI-based QoT prediction
Uzunidis et al. Enhancing closed-form based physical layer performance estimations in EONs via machine learning techniques
Habibi et al. Reducing blocking probability and QoT violation in dynamic elastic optical networks via load-aware margin selection
Liu et al. Auxiliary graph-based energy-efficient routing resource allocation algorithm in space division multiplexing elastic optical networks
Wang et al. Load-aware nonlinearity estimation for efficient resource allocation in elastic optical networks
Mitra et al. Capacity Benefits of Operation Over $\mathrm {C}+\mathrm {L} $ Band Elastic Optical Network in the Indian Network Scenario
Raeesi et al. Low-complexity physical layer impairment aware spectrum assignment based on discretized Gaussian model for nonlinear noise in elastic optical networks
Wang et al. A novel traffic grooming scheme for nonlinear elastic optical network
Ravipudi et al. Does considering only crosstalk suffice for QoT-aware provisioning in multicore fiber networks?
Xie et al. Physical layer-aware route and spectrum allocation in optical networks by multi-objective deep reinforcement learning
CN112203168B (zh) 一种基于任播和多播的节能路由频谱分配方法
CN116320839A (zh) 混合光纤网络中频谱资源的优化分配方法和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19955480

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19955480

Country of ref document: EP

Kind code of ref document: A1