WO2018202297A1 - Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement - Google Patents

Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement Download PDF

Info

Publication number
WO2018202297A1
WO2018202297A1 PCT/EP2017/060603 EP2017060603W WO2018202297A1 WO 2018202297 A1 WO2018202297 A1 WO 2018202297A1 EP 2017060603 W EP2017060603 W EP 2017060603W WO 2018202297 A1 WO2018202297 A1 WO 2018202297A1
Authority
WO
WIPO (PCT)
Prior art keywords
computing device
network computing
metric values
network
time
Prior art date
Application number
PCT/EP2017/060603
Other languages
English (en)
Inventor
Dror MIZRACHI
Aviv GRUBER
Original Assignee
Huawei Technologies Co., Ltd.
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
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Priority to CN201780090166.9A priority Critical patent/CN110574337A/zh
Priority to PCT/EP2017/060603 priority patent/WO2018202297A1/fr
Publication of WO2018202297A1 publication Critical patent/WO2018202297A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/34—Signalling channels for network management communication
    • H04L41/342—Signalling channels for network management communication between virtual entities, e.g. orchestrators, SDN or NFV entities
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0813—Configuration setting characterised by the conditions triggering a change of settings
    • H04L41/0816—Configuration setting characterised by the conditions triggering a change of settings the condition being an adaptation, e.g. in response to network events
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0876—Aspects of the degree of configuration automation
    • H04L41/0886—Fully automatic configuration

Definitions

  • the present invention relates to the field of computer networking. More specifically, the present invention relates to a device, method and system for time-series forecasting based on a moving- median algorithm, in particular to predict real-time network performance.
  • the forecasting is based on metric values, which are obtained by the network computing device, and is preferably performed by a software-defined network (SDN) application.
  • SDN software-defined network
  • a conventional network computing device such as a router or a load balancer, is configured to predict a future network state or a future network performance of a network in which the conventional network computing device is located, in order to optimize and adjust a future configuration of the conventional network computing device, or a conventional network node which is connected to the conventional network computing device.
  • the prediction of the future network state and the future network performance can be implemented by means of an SDN application.
  • metric values are received by the conventional network computing device and are processed according to conventional models in order to obtain predicted metric values according to the processed metric values.
  • the conventional models are e.g.
  • PCT/EP2014/075452 can be based on a moving average calculation, an exponential smoothing calculation, an auto regression calculation and/or an autoregressive integrated moving average calculation.
  • the future state or configuration of a computer network can in turn be determined based on the predicted metric values.
  • the metric values can either be provided to the conventional network computing device, for example by an external system such as a monitoring system, or can be generated in the network computing device.
  • the information on which the metric values are based can be subject to noise, or suffer from sudden outliers or extreme values, which e.g. result from changing transmission characteristics.
  • the above-mentioned models used in the prior art are sufficient when analyzing metric values that are determined based on regular network traffic (which is not subject to noise, outliers or extreme values), the predicted metric values which are provided based on those models are incorrect when the underlying network traffic suffers from noise, outliers or extreme values. Noise, outliers or extreme values may e.g.
  • Another prior art approach is to compare the network traffic to predetermined network traffic patterns, wherein the predetermined network traffic patterns include patterns which indicate that a presently compared portion of network traffic includes noise or an unintentionally caused extreme value.
  • This prior art approach however has the drawback that a large amount of predetermined network traffic patterns has to be stored in the conventional network computing device, and a lot of comparison actions have to be performed, which results in poor performance of prediction of results.
  • the present invention aims to improve the conventional devices, systems and methods.
  • the present invention thereby has the object to provide a network computing device, method and system that allow for predicting network performance while avoiding false classification of analyzed network traffic due to transmission noise and sudden unintentional outliers or extreme values in the network traffic.
  • the present invention proposes a solution that uses a time-series forecasting model which comprises a model based on a moving-median algorithm in order to determine predicted metric values (which represent a predicted network state, network configuration or network performance) based on metric values which are provided to the network computing device, method and system.
  • Time-series forecasting is a machine- learning (ML) technique, which enables predicting future values according to observed historical data by following a forecasting model.
  • time-series forecasting model which comprises a model based on a moving- median algorithm facilitates avoiding erroneous predicted metric values, since a processing result of a moving-median algorithm is not as dramatically affected by noise, outliers or extreme values in monitored network traffic and the processed metric values, as for example a moving average algorithm.
  • a first aspect of the present invention provides a network computing device, including a prediction module configured to process metric values, which are obtained by the network computing device, based on at least one time-series forecasting model, and provide predicted metric values according to the processed metric values, wherein the at least one time-series forecasting model comprises a model based on a moving-median algorithm.
  • the network computing device of the first aspect uses at least one time-series forecasting model comprising a model based on a moving-median algorithm in order to provide predicted metric values according to metric values which are processed in a prediction module of the network computing device.
  • processing of the metric values may include separating a higher half of the processed metric values from a lower half of the processed metric values to determine a middle value of the processed metric values, on which a prediction of the predicted metric values is based.
  • a basic advantage of the moving-median algorithm (for example compared to a moving average algorithm) is that processing results are not skewed so much by extremely large or extremely small values in the processed metric values.
  • the network computing device advantageously avoids that erroneous predicted metric values are provided due to a false classification of metric values which are subject to occasional noise, outliers or extreme values in monitored network traffic.
  • the network computing device when applied in the field of SDN, provides a time-series forecasting model which is more robust to extreme values or noise in the analyzed data compared to the most common time-series forecasting models, such as moving average, or reactive response.
  • a further advantage of the moving-median algorithm based time-series forecasting model as used by the network computing device according to the present invention is that upon consistent change of properties of the processed metric values, the predicted metric values can be more rapidly adapted to the changing attributes of the processed metric values.
  • the network computing device With its robustness against noise and extreme values in the processed metric values, and with its ability to rapidly respond to a consistent change of the processed metric values, the network computing device according to the present invention is ideal for application in the field of SDN, for example in an SDN-based load balancer, router or switch.
  • the network computing device in particular can be provided by means of an SDN application.
  • the network computing device further can comprise a control module configured to determine control parameters according to the predicted metric values.
  • the predicted metric values can thus be used by the control module to obtain control parameters, according to which a future configuration or behavior in the network in which the network computing device is applied can be adapted.
  • control module can be configured to control the network computing device and/or a network node connected to the network computing device according to the control parameters.
  • control parameters can relate to a forwarding configuration and/or a routing configuration of the computing device and/or the network node.
  • control parameters in a computer network can be provided by the network computing device and in particular be based on a moving-median algorithm. This ensures that the provided control parameters do no longer suffer from noise or extreme values.
  • the network computing device further can comprise an obtaining module configured to obtain the metric values by receiving metric values provided to the network computing device, and/or by analyzing network traffic processed in the network computing device.
  • the metric values can thereby be received from an external system, such as a monitoring system, which facilitates efficient obtaining of the metric values by the network computing device.
  • an external system such as a monitoring system
  • analyzing the network traffic by means of the obtaining module in the network computing device allows for a more versatile application of the network computing device, for example in computer networks in which no external monitoring system is present.
  • the metric values and/or the predicted metric values can include information on at least one of a frame delay, a frame latency, a jitter, a packet loss rate, a mean opinion score, transmitted packets, received packets, received bytes, transmitted drops, received drops, transmitted errors, flow count transmitted packets, transmitted bytes and received errors.
  • the metric values and/or the predicted metric values thereby can include various types of information which can be used to characterize network traffic or network behavior.
  • the at least one time-series forecasting model can be a model for predicting network performance metrics.
  • the prediction module can be implemented by means of an SDN application.
  • the functionality of the network computing device according to the first aspect can be provided by means of a networking device which is capable of executing an SDN application, thereby ensuring a broad field of applicability for the network computing device according to the present invention.
  • the network computing device can be an SDN controller, and/or can preferably reside in a control plane.
  • the network computing device can be implemented by means of typical SDN hardware, such as an SDN controller, and/or preferably a network computing device which resides in a control plane, thereby improving overall efficiency of a computer network in which the network computing device according to the first aspect is located.
  • a second aspect of the present invention provides a method for operating a network computing device, the method comprising the steps of: obtaining, by the network computing device, metric values, processing, by a prediction module in the network computing device, the metric values based on at least one time-series forecasting model, and providing, by the prediction module, predicted metric values according to the processed metric values, wherein the at least one time- series forecasting model comprises a model based on a moving-median algorithm.
  • the method further can comprise the step of determining, by a control module, control parameters according to the predicted metric values.
  • the method further can comprise the step of controlling, by the control module, the network computing device and/or a network node connected to the network computing device according to the control parameters.
  • control parameters can relate to a forwarding configuration and/or a routing configuration of the network computing device and/or the network node.
  • the method further can comprise the step of obtaining, by an obtaining module, the metric values by receiving metric values provided to the network computing device, and/or by analyzing network traffic processed in the network computing device.
  • the metric values and/or the predicted metric values can include information on at least one of a frame delay, a frame latency, a jitter, a packet loss rate, a mean opinion score, transmitted packets, received packets, received bytes, transmitted drops, received drops, transmitted errors, flow count transmitted packets, transmitted bytes and received errors.
  • the at least one time-series forecasting model can be a model for predicting network performance metrics.
  • the prediction module can be implemented by means of a software-defined networking (SDN) application.
  • the network computing device can be an SDN controller, and/or preferably resides in a control plane.
  • a third aspect of the present invention provides a computer program product, which, when executed on a computing device, is adapted to perform the method of the second aspect or anyone of its implementation forms.
  • a fourth aspect of the present invention provides a network computing system comprising: a monitoring system, a network computing device, and a network node connected to the network computing device, wherein the monitoring system is configured to provide metric values to the network computing device, wherein the network computing device comprises a prediction module and a control module, wherein the prediction module is configured to process the metric values obtained from the monitoring system based on at least one time-series forecasting model, and provide predicted metric values according to the processed metric values, wherein the at least one time-series forecasting model comprises a model based on a moving-medial algorithm, wherein the control module is configured to determine control parameters according to the predicted metric values, and control the network computing device and/or the network node according to the control parameters.
  • the network computing system of the fourth aspect and in particular the network computing device which is comprised by the network computing system uses at least one time-series forecasting model comprising a model based on a moving-median algorithm in order to provide predicted metric values according to metric values which are processed in a prediction module of the network computing device.
  • processing of the metric values may include separating a higher half of the processed metric values from a lower half of the processed metric values to determine a middle value of the processed metric values, on which a prediction of the predicted metric values is based.
  • a basic advantage of the moving-median algorithm (for example compared to a moving average algorithm) is that processing results are not skewed so much by extremely large or extremely small values in the processed metric values.
  • the network computing system of the fourth aspect and in particular the network computing device which is comprised by the network computing system advantageously avoids that erroneous predicted metric values are provided due to a false classification of metric values which are subject to occasional noise, outliers or extreme values in monitored network traffic.
  • the network computing system and device provide a time-series forecasting model which is more robust to extreme values or noise in the analyzed data compared to the most common time-series forecasting models, such as moving average, or reactive response.
  • a further advantage of the moving-median algorithm based time-series forecasting model as used by the network computing system according to the fourth aspect is that upon consistent change of properties of the processed metric values, the predicted metric values can be more rapidly adapted to the changing attributes of the processed metric values.
  • the network computing system according to the fourth aspect is ideal for application in the field of SDN, for example in an SDN-based load balancer, router or switch.
  • the network computing device comprised by the network computing system according to the fourth aspect in particular can be provided by means of an SDN application.
  • Another advantage of the network computing system according to the fourth aspect is that the network computing device and/or a network node connected to the network computing device can be controlled according to the control parameters which are determined by the control module according to the predicted metric values.
  • This is in particular beneficial, when the functionality of the network computing system according to the fourth aspect and in particular of the network computing device which is comprised by the network computing system according to the fourth aspect is provided by a means of an SDN application, since the SDN application thereby has the possibility to control the network computing system or device which executes the SDN application, or any network node which is connected to the network computing device.
  • a further advantage of the network computing system according to the fourth aspect is that the metric values can be received from an external system, such as a monitoring system, which facilitates efficient obtaining of the metric values by the network computing system.
  • the network node can be at least one of a switch, a router, a bridge, a firewall, or a load balancer, or is configured to run an SDN application.
  • the network computing device can control a network node according to anyone of the above-listed types, and in particular can control a network node which is configured to run an SDN application.
  • the network node can be a network element, and/or preferably reside in a data plane.
  • the network computing system can in particular control a network node which works according to SDN operating principles (i.e. which are typically implemented by means of a network element, and/or by network nodes which reside in a data plane).
  • SDN operating principles i.e. which are typically implemented by means of a network element, and/or by network nodes which reside in a data plane.
  • a fifth aspect of the present invention provides a method for operating a network computing system, the method comprising the steps of providing, by a monitoring system, metric values to a network computing device, processing, by a prediction module in the network computing device, the metric values obtained from the monitoring system based on at least one time-series forecasting model, providing, by the prediction module, predicted metric values according to the processed metric values according to the processed metric values, determining, by a control module in the network computing device, control parameters according to the predicted metric values, and controlling, by the control module, the network computing device and/or a network node connected to the network computing device according to the control parameters, wherein the at least one time-series forecasting model comprises a model based on a moving-median algorithm.
  • the network node can be at least one of a switch, a router, a bridge, a firewall, or a load balancer, or is configured to run an SDN application.
  • the network node can be a network element, and/or preferably reside in a data plane.
  • the method of the fifth aspect and its implementation forms achieve the same advantages as the system of the fourth aspect and its respective implementation forms.
  • FIG. 1 shows a network computing device according to an embodiment of the present invention.
  • FIG. 2 shows a network computing device according to an embodiment of the present invention in more detail.
  • FIG. 3 shows example results of processing metric values according to various algorithms.
  • FIG. 4 shows a method according to an embodiment of the present invention.
  • FIG. 5 shows a computer program product according to an embodiment of the present invention.
  • FIG. 6 shows a network computing system according to an embodiment of the present invention.
  • FIG. 7 shows a method according to an embodiment of the present invention.
  • FIG. 1 shows a network computing device 100 according to an embodiment of the present invention.
  • the network computing device 100 is in particular suitable for performing processing based on at least one time-series forecasting model 103 which comprises a model 105 based on a moving-median algorithm.
  • the network computing device 100 comprises a prediction module 101.
  • the network computing device 100 is configured to obtain metric values 102.
  • the metric values 102 can be a numeric measure for quality and/or state of a network connection.
  • the metric values 102 can be obtained based on analyzing network traffic, or a state of a network to which the network computing device 100 is connected. More preferably, the metric values 102 can be obtained by analyzing IPv4 or IPv6 network traffic, or network state.
  • the metric values 102 can also be provided to the network computing device 100 by means of an external system.
  • the prediction module 101 is configured to process the metric values 102 based on at least one time-series forecasting model 103. After processing the metric values 102, the prediction module 101 provides predicted metric values 104 according to the processed metric values 102.
  • the predicted metric values 104 can be a numeric measure for future quality and/or future state of a network connection.
  • the prediction module 101 comprises at least one time-series forecasting model 103 which comprises a model 105 based on a moving-median algorithm.
  • the network computing device 100 can also comprise more time-series forecasting models 103 , which is illustrated by the three dots below the time-series forecasting model 103 in FIG. 1.
  • n is a parameter that can be used to configure a sliding window size of the moving-median algorithm, n in particular can be in the range of X to Y.
  • the sliding window size (which also can be called sample window size) specifies the amount of metric values 102, which is considered in one calculating step of the moving-median algorithm.
  • PVt+k represents the predicted metric value at time t+k.
  • t represents the time at which a metric value 102 is obtained, and k represents the time ahead of t, for which a predicted metric value 104 is provided, while l ⁇ k ⁇ .
  • k can also be regarded as a value, specifying how long the term of the prediction is.
  • s x represents the metric value at position x of the array S.
  • FIG. 2 shows a network computing device 100 according to the present invention in more detail.
  • Features of the network computing device 100 which were already described in view of FIG. 1, but are also disclosed or further specified in view of FIG. 2, are labelled with identical reference signs.
  • the network computing device 100 can further comprise an optional control module 201.
  • the control module 201 can be implemented by means of an SDN application which can be executed by the network computing device 100.
  • the control module 201 is configured to receive the predicted metric values 104, which are provided by the prediction module 101.
  • the control module 201 is further configured to determine control parameters 202 according to the received predicted metric values 104.
  • the control module 201 allows for controlling, adjusting or updating the state, and/or attributes of all devices of a computer network, to which the network computing device 100 is connected (including the network computing device 100 itself).
  • control module 201 can be configured to control the network computing device 100, and/or a network node 203, which can be connected to the network computing device 100.
  • the control operations which are performed by the control module 201, are performed according to the control parameters 202, which are determined based on the predicted metric values 104.
  • Controlling the network computing device 100 and/or the network node 203 by the control module 201 is illustrated by the dashed arrow connecting the control module 201 with the network node 203, respectively by the dashed arrow starting from the control module 201 and ending in the network computing device 100 in FIG. 2.
  • the control parameters 202 are illustrated by means of the rectangles next to the dashed arrows in FIG. 2.
  • control parameters 202 can relate to a forwarding configuration and/or a routing configuration, either of the network computing device 100, the network node 203, or both.
  • the forwarding configuration and/or the routing configuration can for example be a lookup table, according to the entries of which forwarding, respectively routing of packets can be influenced.
  • a forwarding behavior and/or a routing behavior of the network computing device 100 itself, and/or the network node 203 can be adjusted.
  • the forwarding configuration can in particular relate to a forwarding table of the network computing device 100 and/or the network node 203
  • the routing configuration can in particular relate to a routing table of the network computing device 100 and/or the network node 203.
  • the network computing device 100 can further comprise an optional obtaining module 204.
  • the obtaining module 204 is in particular configured to obtain the metric values 102.
  • the obtaining module can receive metric values 205, which are provided to the network computing device 100 from an external system, e.g. a monitoring system.
  • the obtaining module 204 additionally or alternatively can be configured to analyze network traffic 206, which is processed in the network computing device 100.
  • the network computing device 100 can also comprise at least one network interface, which is connected to the obtaining module 204.
  • the obtaining module 204 preferably can be provided by means of an SDN application, which is executed by the network computing device 100.
  • the metric values 102 and/or the predicted values 104 as described in view of FIG. 1 and FIG. 2, as well as the metric values 205 as described in FIG. 2 can include information on at least one of a frame delay, a frame latency, a jitter, a packet loss rate, a mean opinion score, transmitted packets, received packets, received bytes, transmitted drops, received drops, transmitted errors, flow count transmitted packets, transmitted bytes, or received errors.
  • This ensures that the network computing device 100 can consider various numeric measures according to which the quality and properties of a communication path or a network connection, and in particular of transmitted network traffic, can be determined, when providing the predicted metric values 104 based on the metric values 102.
  • the at least one time-series forecasting model 105 can also be considered a model for predicting network performance metrics.
  • the network computing device 100 can be an SDN controller, and/or can more preferably reside in a control plane.
  • FIG. 3 shows results of processing metric values according to several algorithms.
  • section 301 of FIG. 3 two graphs are shown, in which metric values and a corresponding processing result, which is based on a reactive-response algorithm, are illustrated.
  • the processing result based on the reactive-response algorithm is almost identical to the portions of the graph in which the metric values include the extreme values. This can particularly be seen in portion 301a of section 301.
  • the processing results based on the reactive-response algorithm quickly follow the underlying metric values. Nevertheless, the processing results based on the reactive response algorithm exceed a predetermined threshold in case of unintentional extreme values in the metric values as well as in case of a consistent change of the metric values. Since a predetermined action is performed upon exceedance of the predetermined threshold, the processing results based on the reactive- response algorithm as shown in section 301 of FIG. 3 are not reliable and effective for time- series forecasting, in particular for prediction of network performance.
  • FIG. 3 two graphs, which illustrate metric values and processing results based on a moving average algorithm, are shown.
  • the processing results which are based on the moving average algorithm, do not as dramatically suffer from extreme values in the metric values, as for example the processing results based on the reactive- response algorithm.
  • the processing results, which are based on the moving-average algorithm also more slowly react to a consistent change of the metric values. Both in the case of unintentional extreme values, as well as in the case of a consistent change of metric values, the processing results exceed a predetermined threshold, as can be seen in portions 302a and 302b of FIG. 3.
  • processing metric values based on a time-series forecasting model which comprises a model based on a moving-median algorithm is not reliable and effective for the same reasons as mentioned regarding section 301 above.
  • FIG. 3 two graphs are shown in which metric values and processing results based on a moving-median algorithm are illustrated.
  • the processing results based on the moving-median algorithm remain unaffected from peak values or noise in the metric values upon which the processing is based.
  • the processing results, which are based on the moving-median algorithm quickly adapt to the change of the processed metric values. That is, the predetermined threshold, upon exceedance of which further actions are based, is only exceeded by the processing results, which are based on the moving-median algorithm, if a consistent change of processed metric values occurs.
  • time-series forecasting and in particular the prediction of network performance, according to a model based on a moving-median algorithm is reliable and effective and solves the problem of the prior art solutions, as for example illustrated in sections 301 and 302 of FIG. 3.
  • FIG. 4 shows a method 400 according to an embodiment of the present invention.
  • the method 400 corresponds to the network computing device 100 of FIG. 1 and is in particular suitable for time-series forecasting based on a moving-median algorithm.
  • the method 400 comprises a first step of obtaining S401, by a network computing device 100, metric values 102. Further, the method 400 comprises a second step of processing S402, by a prediction module 101 in the network computing device 100, the metric values 102 based on at least one time-series forecasting model 103. Finally, the method 400 comprises a step of providing S403, by the prediction module 101, predicted metric values 104 according to the processed metric values 102, wherein the at least one time-series forecasting model 103 comprises a model 105 based on a moving-median algorithm.
  • FIG. 5 shows a computer program product 500.
  • the computer program product 500 is adapted to perform the method 400 when being executed on a computing device.
  • the computer program product 500 can in particular include program code which is adapted to perform the method 400.
  • the program code can be stored on or transmitted by a storage medium, which together with the program code forms the computer program product 500.
  • the storage medium may be a computer-readable storage medium, such as, but not limited to, any type of disk including floppy disks, optical disks, CD-ROMs, DVDs, magnetic-optical disks, read-only-memories (ROMs), random access memories (RAMs), EPROMs, EEPROMs, magnetic or optical cards, application specific integrated circuits (ASICs) or any type of media suitable for storing or transmitting electronic instructions (including electronic instructions being transmitted by means of network transmission technology) and each coupled to a computer system bus.
  • FIG. 6 shows a network computing system 600 according to an embodiment of the present invention.
  • the system includes a network computing device 100, a network node 203, and a monitoring system 601.
  • the network computing device 100 and the network node 203, as well as all modules and functionality which are comprised by them operate according to the principle as described in the above. Therefore, the network computing device 100 and its modules and features, as well as the network node 203 are labeled with reference signs identical to those as used in the figures which are previously described.
  • the monitoring system 601 of the network computing system 600 is configured to provide metric values 102 to the network computing device 100.
  • the monitoring system 601 can obtain the metric values 102 by analyzing a state or a configuration of a network to which the monitoring system 601 is connected, as well as network traffic which is processed in the network to which the monitoring system 601 is connected.
  • the metric values 102 are processed in the network computing device 100 as described in the above.
  • the network computing device 100 and/or the network node 203 can in turn be controlled according to the processed metric values 102.
  • FIG. 7 shows a method 700 according to an embodiment of the present invention.
  • the method 700 is suitable for operating the network computing system 600.
  • the method 700 comprises a first step of providing S701 by a monitoring system 601, metric values 102 to a network computing device 100.
  • the method 700 further comprises a second step of processing S702, by a prediction module 101 in the network computing device 100, the metric values 102 obtained from the monitoring system 601 based on at least one time-series forecasting model 103.
  • the method 700 further comprises a third step of providing S703, by the prediction module 101, predicted metric values 104 according to the processed metric values 102.
  • the method 700 further comprises a fourth step of determining S704, by a control module 201 in the network computing device 100, control parameters 202 according to the predicted metric values 104.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Algebra (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

La présente invention concerne un dispositif informatique de réseau (100). Le dispositif informatique de réseau (100) comprend un module de prédiction (101) configuré pour traiter des valeurs métriques (102), qui sont obtenues par le dispositif informatique de réseau (100), sur la base d'au moins un modèle de prévision chronologique (103), et fournir des valeurs métriques prédites (104) en fonction des valeurs métriques (102) traitées, le ou les modèles de prévision chronologique (103) comprenant un modèle (105) basé sur un algorithme médian en mouvement.
PCT/EP2017/060603 2017-05-04 2017-05-04 Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement WO2018202297A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201780090166.9A CN110574337A (zh) 2017-05-04 2017-05-04 基于滑动中值算法进行时间序列预测的网络计算设备、方法及系统
PCT/EP2017/060603 WO2018202297A1 (fr) 2017-05-04 2017-05-04 Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/EP2017/060603 WO2018202297A1 (fr) 2017-05-04 2017-05-04 Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement

Publications (1)

Publication Number Publication Date
WO2018202297A1 true WO2018202297A1 (fr) 2018-11-08

Family

ID=58671656

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2017/060603 WO2018202297A1 (fr) 2017-05-04 2017-05-04 Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement

Country Status (2)

Country Link
CN (1) CN110574337A (fr)
WO (1) WO2018202297A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114205853A (zh) * 2020-09-02 2022-03-18 中兴通讯股份有限公司 流量预测方法、流量预测模型建立方法和存储介质
WO2023125880A1 (fr) * 2021-12-30 2023-07-06 维沃移动通信有限公司 Procédé et appareil de construction de modèle, et dispositif de communication

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8811172B1 (en) * 2014-04-10 2014-08-19 tw telecom holdings inc. Network path selection using bandwidth prediction
US20150163159A1 (en) * 2013-12-10 2015-06-11 International Business Machines Corporation Software-defined networking single-source enterprise workload manager
WO2015195159A1 (fr) * 2014-06-18 2015-12-23 Hewlett-Packard Development Company, L.P. Réacheminement automatique d'un trafic de réseau dans un réseau défini par logiciel
US20150372904A1 (en) * 2014-06-18 2015-12-24 Radware, Ltd. Predictive traffic steering over software defined networks
WO2016082859A1 (fr) * 2014-11-25 2016-06-02 Huawei Technologies Co., Ltd. Système et procédé pour une optimisation de flux de données

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101442807B (zh) * 2008-12-30 2012-09-05 北京邮电大学 通信系统中资源分配的方法及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150163159A1 (en) * 2013-12-10 2015-06-11 International Business Machines Corporation Software-defined networking single-source enterprise workload manager
US8811172B1 (en) * 2014-04-10 2014-08-19 tw telecom holdings inc. Network path selection using bandwidth prediction
WO2015195159A1 (fr) * 2014-06-18 2015-12-23 Hewlett-Packard Development Company, L.P. Réacheminement automatique d'un trafic de réseau dans un réseau défini par logiciel
US20150372904A1 (en) * 2014-06-18 2015-12-24 Radware, Ltd. Predictive traffic steering over software defined networks
WO2016082859A1 (fr) * 2014-11-25 2016-06-02 Huawei Technologies Co., Ltd. Système et procédé pour une optimisation de flux de données

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114205853A (zh) * 2020-09-02 2022-03-18 中兴通讯股份有限公司 流量预测方法、流量预测模型建立方法和存储介质
WO2023125880A1 (fr) * 2021-12-30 2023-07-06 维沃移动通信有限公司 Procédé et appareil de construction de modèle, et dispositif de communication

Also Published As

Publication number Publication date
CN110574337A (zh) 2019-12-13

Similar Documents

Publication Publication Date Title
US11580449B2 (en) Traffic-based inference of influence domains in a network by using learning machines
EP4195594A1 (fr) Procédé et appareil de gestion d'encombrement, dispositif de n'oeud de réseau, et support d'enregistrement lisible par ordinateur
US10958506B2 (en) In-situ OAM (IOAM) network risk flow-based “topo-gram” for predictive flow positioning
CN111526096A (zh) 智融标识网络状态预测与拥塞控制系统
US20150195212A1 (en) Dynamic network-driven application packet resizing
EP3092745A1 (fr) Commande de réacheminement utilisant une prédiction de pannes basée sur un apprentissage automatique précoce
WO2016060751A1 (fr) Gestion de flux de trafic de réseau utilisant l'apprentissage automatique
US9813438B2 (en) Anomaly prediction method and system for heterogeneous network architecture
US9652720B2 (en) Triggering on-the-fly requests for supervised learning of learning machines
CN116545936B (zh) 拥塞控制方法、系统、装置、通信设备及存储介质
CN109257252B (zh) 一种判断网络拥塞状态的方法
US20150186799A1 (en) Point-to-multipoint communication infrastructure for expert-based knowledge feed-back using learning machines
WO2018202297A1 (fr) Dispositif informatique de réseau, procédé et système de prévision en série chronologique sur la base d'un algorithme médian en mouvement
Seddiqi et al. A new protection-based approach for link failure management of software-defined networks
CN117527479B (zh) 软总线组网连接方法、装置、设备及存储介质
US8788735B2 (en) Interrupt control apparatus, interrupt control system, interrupt control method, and interrupt control program
Razmara et al. A Hybrid Neural Network Approach for Congestion Control in TCP/IP Networks
US20220368590A1 (en) Fault Detection Method, Apparatus, and System
CN117729602A (zh) 链路切换控制方法和装置
El Khayat et al. Improving TCP in wireless networks with an adaptive machine-learnt classifier of packet loss causes
CN113079127B (zh) 攻击识别数据模型的生成与应用方法
JP2018533327A (ja) イーサネット仮想接続に関するリアルタイム分散エンジンフレームワーク
El Khayat et al. Machine-learnt versus analytical models of TCP throughput
TWI724734B (zh) 攻擊辨識資料模型的生成與應用方法
CN117896776B (zh) 无线自组网的数据自动重传方法及相关装置

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: 17721649

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: 17721649

Country of ref document: EP

Kind code of ref document: A1