CN112887148A - Method and device for simulating and predicting network flow - Google Patents

Abstract

The present invention relates to the field of communications, and in particular, to a method and an apparatus for network traffic simulation and prediction. The method comprises the following steps: generating a fault signal sequence for at least one network element or link; acquiring fault signals in fault signal sequences of all network elements and links, and accumulating the acquired fault signals on network topology; acquiring fault service corresponding to each fault signal according to the time sequence of the occurrence of the faults; calculating the route change of the fault service after each fault occurs according to the change topological graph after each fault occurs, and generating a simulation transaction; recording the flow of each fault service after each simulation transaction is executed, and predicting the change of flow distribution in the network; and after the execution of the simulation transaction is finished, rolling back the suspended simulation transaction. The method realizes flow simulation and prediction simulating the real situation as much as possible under the conditions of not changing network hardware equipment and not influencing the normal service of the network through simulated fault signals and real service calculation.

Description

Method and device for simulating and predicting network flow

[ technical field ] A method for producing a semiconductor device

The present invention relates to the field of communications, and in particular, to a method and an apparatus for network traffic simulation and prediction.

[ background of the invention ]

In the current network planning method, a network is regarded as a static topology, the flow of each link and each node is planned in advance, and the flow is distributed on each link and node in a load balancing manner as much as possible according to the plan when service configuration is performed. However, with the increase of dynamic services, the states of the nodes of the network will change dynamically, and the traffic distribution during planning will not be the same, which may cause links or nodes that are not bottleneck points during planning to become bottleneck points after a period of time change, thereby continuously decreasing the overall performance of the network. In order to solve such problems, a network management and control system is required to predict a possible change of traffic distribution according to a preset routing policy and a preset failure point during network planning.

However, the current network management and control system has the following problems: firstly, when the network is planned, the existing management and control system cannot predict the flow distribution after the network fault, and the network fault does not influence the existing network; secondly, the generation of network faults has the property of time series, so that the flow distribution has the property of dynamic change along with time, and the existing management and control system is difficult to track the change; thirdly, network failures are generated in a multipoint concurrent mode, dependency relationships can exist among the failures, changes of flow distribution are also influenced, and the existing management and control system is difficult to recognize the relationships.

In view of this, how to overcome the defects in the prior art, and solve the problem that the existing network management and control system cannot simulate the real fault and the traffic change condition, are problems to be solved in the technical field.

[ summary of the invention ]

Aiming at the defects or the improvement requirements of the prior art, the invention solves the problem of truly simulating the service flow change after the network failure under the condition of not influencing the normal operation of the current network.

The embodiment of the invention adopts the following technical scheme:

in a first aspect, the present invention provides a method for simulating and predicting network traffic, specifically: generating a set of fault signal sequences for at least one network element or link; acquiring fault signals in fault signal sequences of all network elements and links according to a time sequence, accumulating each acquired fault signal on network topology, and generating a change topological graph after each fault occurs; acquiring fault service corresponding to each fault signal according to the time sequence of the occurrence of the faults; calculating the route change of the fault service after each fault occurs according to the change topological graph after each fault occurs, generating a simulation transaction, and suspending the simulation transaction without submitting the simulation transaction; recording the flow of each fault service after each simulation transaction is executed, acquiring the change of the service flow, and predicting the change of flow distribution in the network; and after the execution of the simulation transaction is finished, rolling back the suspended simulation transaction.

Preferably, generating a group of fault signal sequences for at least one network element or link specifically includes: generating a set of random fault time sequences for at least one network element or link using a time sequence generating function; generating at least one piece of fault data for each faulty network element or faulty link; corresponding each fault data to a time point in the fault time sequence of the network element or the link to generate a fault signal sequence of the network element or the link; and serially arranging the fault data in the fault signal sequences of all the network elements and the links according to the time sequence.

Preferably, the generating a set of random failure time sequences for at least one network element or link further comprises: and adjusting the distribution density of the fault time in the fault time sequence of the network element or the link according to the fault probability curve of each network element or link, so that the fault occurrence probability of the network element or the link in a unit time period is in accordance with the fault probability curve of the network element or the link.

Preferably, the step of corresponding each piece of failure data to a time point in the failure time sequence of the network element or the link further includes: the types of the fault signals comprise fault occurrence and fault recovery, and the two types of the fault occurrence and the fault recovery alternately appear according to a time sequence.

Preferably, accumulating each acquired fault on the network topology specifically includes: acquiring a topology when no fault occurs as a baseline topology; sequentially acquiring each fault signal, and converting the fault signals into state changes of network elements or links; starting from the baseline topology, according to the state change of the network element or link corresponding to each fault signal, corresponding state change is carried out on the basis of the previous change topological graph, and the changed change topological graph corresponding to each fault signal is obtained.

Preferably, the acquiring the fault service corresponding to each fault signal specifically includes: acquiring a network element or a link corresponding to each fault signal; and searching all services passing through the network element or the link corresponding to the fault signal as fault services.

Preferably, calculating the route change of the failed service after each failure occurs specifically includes: if the working route of the fault service has a network element or link fault for the first time, generating a first standby route for the fault service; if the first standby route of the fault service has a network element or link fault for the first time, generating a second standby route for the fault service; if the working route of the fault service has the condition that all network elements and links have no fault for the first time, the standby route of the fault service is switched back to the working route.

Preferably, the acquiring the change of the service flow specifically includes: and according to the route change of the fault service, reducing the traffic of the service on the route of the fault service, and increasing the traffic of the service on the new route to which the service is transferred.

Preferably, predicting the change of the flow distribution in the network specifically includes: recalculating the bandwidth occupancy rate of the service according to the change of the service flow corresponding to the fault information; and recalculating the traffic distribution of the service in the changed network topology according to the network topology change corresponding to the fault information.

Preferably, the traffic distribution comprises bottleneck nodes of the traffic flow and idle points in the network topology.

On the other hand, the invention provides a device for simulating and predicting network traffic, which specifically comprises the following steps: the network traffic simulation system comprises at least one processor and a memory, wherein the at least one processor and the memory are connected through a data bus, and the memory stores instructions capable of being executed by the at least one processor, and the instructions are used for completing the network traffic simulation and prediction method in the first aspect after being executed by the processor.

Compared with the prior art, the embodiment of the invention has the beneficial effects that: when the network planning simulation is carried out, firstly, a simulated dynamic fault signal sequence is generated, the dynamic change of the network topology under the real condition due to faults is simulated, then, the real fault service routing change is calculated according to the dynamic change of the network topology, and the real change condition of the network flow is obtained. The method realizes flow simulation and prediction simulating the real situation as much as possible under the conditions of not changing network hardware equipment and not influencing the normal service of the network through simulated fault signals and real service calculation.

[ description of the drawings ]

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments of the present invention will be briefly described below. It is obvious that the drawings described below are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.

Fig. 1 is a flowchart of a method for simulating and predicting network traffic according to an embodiment of the present invention;

FIG. 2 is a flow chart of another method for network traffic simulation and prediction according to an embodiment of the present invention;

FIG. 3 is a flow chart of another method for network traffic simulation and prediction according to an embodiment of the present invention;

FIG. 4 is a flow chart of another method for network traffic simulation and prediction according to an embodiment of the present invention;

fig. 5 is a schematic diagram of a network topology used in a method for network traffic simulation and prediction according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a device for simulating and predicting network traffic according to an embodiment of the present invention.

[ detailed description ] embodiments

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The present invention is a system structure of a specific function system, so the functional logic relationship of each structural module is mainly explained in the specific embodiment, and the specific software and hardware implementation is not limited.

In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other. The invention will be described in detail below with reference to the figures and examples.

Example 1:

in the network planning process of the telecommunication transmission network, the flow distribution is an important index for judging the network planning quality. Firstly, a bottleneck point in a network can be identified through flow distribution, and the network with balanced load can be planned; secondly, the change of flow distribution when a fault occurs in the network needs to be considered during network planning, so that the instantaneous network flow overrun is avoided; finally, multiple changes of flow distribution caused by multiple network faults in a short time are considered in network planning, and the impact of frequent changes of the flow distribution on the existing network is avoided. When network planning is performed, in order to simulate and predict service traffic, a management and control system needs to simulate a real network topology and service routing according to a time sequence generated by a network fault, so that traffic distribution change caused by the fault is consistent with a real change condition of an existing network, and the change of the traffic distribution of the whole network after a plurality of faults continuously occur is tracked. However, in order to avoid the influence of fault simulation on the running service, the network fault generated when the management and control system performs flow simulation cannot influence the normal running of the existing network. Therefore, in the method provided by this embodiment, the management and control system identifies the dependency relationship between the multiple point faults, generates the simulation transaction to simulate and predict the traffic change, and performs rollback after the simulation transaction is executed, so as to reduce the traffic distribution change in the real network and avoid the influence of the simulation and simulation processes on the existing network.

As shown in fig. 1, the method for simulating and predicting network traffic provided by the embodiment of the present invention includes the following specific steps:

step 101: generating a set of fault signal sequences for at least one network element or link;

in real network usage, network elements and links in the network may be in dynamic changes due to failures, traffic scheduling, or device on/off-line, etc., and traffic connections and flow in the network may also change accordingly. During simulation, a part of network elements and links can be selected to generate fault signal sequences according to a fault scene to be simulated, and fault signal sequences can be generated for all the network elements and links. Because each network element and each link in the real network environment may have a fault, fault repair, on-line, off-line, and the like at any time, at least one possible fault needs to be generated for each faulty network element and each link, and a fault signal sequence is generated in a time sequence for simulation.

Specifically, as shown in fig. 2, the following steps may be used to generate the fault signal sequence.

Step 201: a set of random time series of failures is generated for at least one network element or link using a time series generation function.

In a real network environment, each network element or each link has one or more faults at random time, and in order to simulate the faults, a time point of the fault of each network element and each link is firstly obtained. Specifically, a time series generating function may be used to generate the fault time series, and the optional time series generating function includes: random generators, periodic function generators, poisson function generators, markov chain generators, and the like. For a certain network element or link, a failure time sequence is generated, which indicates that the network element or link has failed, and is referred to as a failed network element or a failed link in the embodiment of the present invention.

Further, since the failure probability of different network elements and different links is different, the same network element device may also have a failure rate curve related to the service life. In this embodiment, in a scenario with a higher prediction accuracy requirement, in order to make the simulation of the fault signal more consistent with the real situation, the uniformly distributed random fault time sequence may not be directly used, but the distribution density of the fault time sequence of each network element or link may be adjusted according to different fault probabilities of different network elements and links or the current fault rate of the network element device, so that the fault time sequence more conforms to the fault occurrence probability in the real scenario.

Step 202: at least one piece of fault data is generated for each faulty network element or faulty link.

In order to simulate the situation of each network element and link failure in the simulation and prediction time periods, one or more pieces of failure data corresponding to one or more failures of each network element and link need to be generated for each network element and link. Generally, each time point in the fault time sequence corresponds to a fault datum. In specific use, the content of the fault data can be determined according to specific needs of simulation, and generally comprises positioning information and fault codes, wherein the positioning information comprises a network element ID, a link ID, a single disk ID and a port number; the fault code is a standard fault code specified in a network communication protocol and a defined fault code in a network management and control system. Further, different sets of fault codes can be set for different network elements and links to simulate different network fault scenarios.

Step 203: and corresponding each fault data to a time point in the fault time sequence of the network element or the link, and generating a fault signal sequence of the network element or the link.

For each network element or link, the fault signal is sent out serially, and the network management equipment only receives one piece of fault data at the same time point. After each fault data is matched with one time point in the fault time sequence, a plurality of groups of data pairs (time and fault data) are generated, and the data pairs are arranged according to the time sequence, so that a fault signal sequence which may occur to the network element or the link in the time period of the simulation can be obtained.

Further, for each network element or each link, the situations of equipment failure and failure recovery may occur alternately. Therefore, in the present embodiment, the types of the fault signal can be divided into the fault occurrence and the fault recovery, and the two types of the fault occurrence and the fault recovery alternately appear according to the time sequence, so that the present embodiment is more suitable for the actual situation. Furthermore, according to different fault codes in the fault data, the fault occurrence can be further divided into specific types such as equipment offline, equipment complete unavailability, equipment communication congestion, equipment time delay lengthening and the like according to the specific and actual possible fault conditions, so as to provide corresponding different service simulations for different requirements of services.

Step 204: and serially arranging the fault data in the fault signal sequences of all the network elements and the links according to the time sequence.

In a real network, faults often occur in parallel on a plurality of network elements and a plurality of links, but fault signals reach a network management and control system in a time sequence. In the same network system, faults of all network elements and links are uniformly managed by one management and control system, and the management and control system receives fault signals of all the network elements and the links according to a time sequence. Therefore, in performing the simulation, it is also necessary to arrange fault signal sequences of all network elements and links on a uniform time axis and simulate the sequential receiving operation of the pipe control system. For example, the fault signal sequence of the network element 1 is { (time 1, fault 1), (time 4, fault 2) }, the fault signal sequence of the network element 2 is { (time 3, fault 3), (time 5, fault 4) }, the fault signal sequence of the network element 3 is { (time 2, fault 5) }, and the fault signal sequence after the serial arrangement is { (time 1, fault 1), (time 2, fault 5), (time 3, fault 3), (time 4, fault 2), (time 5, fault 4) }.

Through the steps 201 to 204, a group of simulated fault signals under different scenes are simulated by utilizing a random fault time sequence generated by software without using real hardware equipment for fault simulation, so that the influence on a real network environment during flow simulation and prediction is reduced.

Step 102: and acquiring fault signals in the fault signal sequences of all network elements and links according to the time sequence, accumulating each acquired fault signal on the network topology, and generating a change topological graph after each fault occurs.

In a real network environment, topology and traffic in a network dynamically change with the fault change of different network elements and links, and when the state of any one network element or link in the network changes, the current network topology changes on the basis of the network topology at the previous moment. In order to simulate the change of the network topology in the real network environment along with the time change, when each fault signal in the fault signal sequence is received, the fault signals are sequentially accumulated on the state of the current network topology according to the time sequence, and a change topological graph after the fault occurs is generated.

As shown in fig. 3, the following steps may be specifically used to generate the change topology map at the current time point.

Step 301: and acquiring the topology when no fault occurs as a baseline topology.

When fault simulation is carried out, firstly, the initial state of the network, namely the network topology state when no fault normally runs, needs to be obtained, and when simulation is carried out, the topology state is taken as the state change start, and changes caused by faults are accumulated on the topology state, so that the dynamic change conditions of the network topology at different moments are simulated.

Step 302: and sequentially acquiring each fault signal, and converting the fault signals into state changes of the network elements or the links.

When the simulation fault data are generated, different types of fault signals can be generated according to the requirements, and the different types of fault signals correspond to different change forms of the network topology. For example, when the type of the fault signal is divided into two types, namely fault occurrence and fault recovery, the topology change corresponding to the fault occurrence of the network element is to delete the node where the network element is located and the link to which the node is connected in the topology map, the topology change corresponding to the fault occurrence of the link is to delete the link in the topology map, the topology change corresponding to the fault recovery of the network element is to add the node corresponding to the network element in the topology map and recover the link to which the node is connected, and the topology change corresponding to the fault recovery of the link is to add the link in the topology map.

Step 303: starting from the baseline topology, according to the state change of the network element or link corresponding to each fault signal, corresponding state change is carried out on the basis of the previous change topological graph, and the changed change topological graph corresponding to each fault signal is obtained.

After the state change corresponding to the fault signal is obtained, according to the time sequence of the fault occurrence, corresponding state change is sequentially carried out on the basis of the previous network topological graph from the baseline topology, and the changed topological graph after each fault occurrence is obtained.

Further, in order to facilitate calculation and save storage consumption, in the embodiment of the present invention, all of the baseline topology and subsequent changed topology maps are not saved, and only the topology changed map at the current time, that is, the topology snapshot at the current time is retained. And after each fault signal is received, modifying the topology change graph at the previous moment to generate a topology snapshot at the current moment, wherein only one topology snapshot is maintained at each moment.

After steps 301 to 303, a network change topology map after each fault occurs is obtained according to the fault signal sequence, and in subsequent steps, the service traffic of the network can be simulated and predicted according to the network topology change maps at different times.

Step 103: and acquiring the fault service corresponding to each fault signal according to the time sequence of the occurrence of the fault.

After acquiring the simulation fault data in a time period and the simulation network topology after each fault occurs, acquiring the service influenced by the network topology change, namely the fault service generating the fault due to the network element or link fault.

As shown in fig. 4, the following steps may be used to obtain the fault traffic corresponding to each fault signal.

Step 401: and acquiring the network element or link corresponding to each fault signal.

In an embodiment of the invention, each fault signal indicates a fault of one network element or one link. In this embodiment, after each failure occurs, the change caused by the current failure is superimposed on the network change topology map at the previous time. Each change topology map contains network topology changes caused by all faults before the current fault time point, and the network changes can cause service faults relative to a baseline topology when the network does not have faults. Therefore, when acquiring a fault service corresponding to a fault signal, not only the service currently passing through the network element or link corresponding to the current fault signal needs to be acquired, but also the network elements or links corresponding to all the generated fault signals on the graph are acquired according to the current network change topology graph.

Step 402: and searching all services passing through the network element or the link corresponding to the fault signal as fault services.

Changes in the network element or link caused by each fault signal, whether the fault type is a fault occurrence or a fault recovery, may result in changes in the traffic path. After acquiring network elements or links corresponding to all fault signals at the current time point, searching all services which need to use the network elements or links corresponding to the fault signals before the fault occurs by comparing a network topology change diagram before the fault occurs with a network topology change diagram after the fault occurs, or checking paths in a service route, wherein the services are fault services influenced by the fault signals.

Through steps 401 to 402, all the fault services in the current network topology state can be acquired.

Step 104: and calculating the route change of the fault service after each fault occurs according to the change topological graph after each fault occurs, generating a simulation transaction, and suspending the simulation transaction without submitting the simulation transaction.

In order to simulate the service flow change condition after the fault occurs, the network management and control system directly calculates the route change of the fault service according to the network topology change diagram after the fault occurs, and produces the simulation affair of rerouting according to the calculation result.

During normal service routing planning, an optimal working path is allocated to a service according to factors such as service requirements, network element and link bearing conditions, path length and the like. After each fault occurs, according to the difference of fault types, fault services related to the fault need to be rerouted to a non-fault standby route for transmission, and the specific simulated transaction corresponding to the route change comprises the following types.

(1) If the fault service working route has a network element or link fault for the first time, the fault service working route cannot be used, the working path is recalculated for the fault service according to the current network topology change diagram, a first standby route is generated, and the first standby route is used as the fault service route. According to the service needs and the specific attributes of the fault, the first backup route can be set as the optimal working path of the fault service except the working route, and can also be set as the route which is easiest to switch.

(2) If the first backup route of the fault service has a network element or link fault for the first time, the original working route and the first backup route cannot be used, the working path is recalculated for the fault service according to the current network topology change diagram, a second backup route is generated, the second backup route is used as the service route of the fault service, and the second backup route is a secondary route after the first backup route. And after the fault service is migrated to the second standby route, the first standby route is regarded as the working route, and the second standby route is regarded as the first standby route.

(3) If the working route of the fault service has the condition that all network elements and links have no fault for the first time, the working route of the service can be normally used, and the fault service is switched back to the working route from the standby route so as to achieve the best communication effect. After switching back to the working route, the standby route can be deleted to reduce the storage consumption, and the standby route can also be reserved for reuse.

(4) If the working route has a fault and the reserved standby route exists, the route is not recalculated, the reserved standby route is used, and the repeated calculation caused by the topology change is reduced.

In this embodiment, in order to enable simulation of a fault and a service flow not to affect actual operation of a network, only a route change of a service is generated into a simulation transaction, the simulation transaction is used for subsequent flow calculation and prediction, the simulation transaction is suspended and not submitted, and change of actual network service is not caused.

Step 105: and recording the flow of each fault service after each simulation transaction is executed, acquiring the change of the service flow, and predicting the change of flow distribution in the network.

After the simulation transaction is generated, the network management and control system may be used to execute the simulation transaction, and record simulation data of each service flow change after the simulation transaction is executed. The change of the service flow can be directly obtained according to the route change of the fault service, the flow of the service is reduced on the service route with the fault, and the flow of the service is increased on a new route to which the service is switched. If the service is switched from the working route to the standby route, reducing the flow of the service on the working route and increasing the flow of the service on the standby route; if the service is switched from the standby route to the working route, the traffic of the service is reduced on the standby route, and the traffic of the service is increased on the working route. If the service route is not changed, the service flow distribution is not changed.

After the traffic flow change is obtained, the change of the traffic distribution in the network can be predicted according to the traffic changes of all the services in the network. Specifically, the bandwidth occupancy rate of the service can be recalculated according to the change of the service traffic corresponding to the fault information; and recalculating the traffic distribution of the service in the changed network topology according to the network topology change corresponding to the fault information. According to the calculated service bandwidth occupancy rate and the calculated traffic distribution of the services, the service traffic condition when the same fault as the simulation fault is generated in the real network scene can be predicted. When predicting traffic distribution, main prediction objects include a bottleneck node of traffic flow with high correlation with traffic path planning and an idle point in a network topology.

Step 106: and after the execution of the simulation transaction is finished, rolling back the suspended simulation transaction.

After acquiring the service traffic simulation data and the prediction data in the network, in order to avoid the influence of the simulation transaction on the real network service and the network topology, the suspended simulation service needs to be rolled back, so that the network service operation is restored to the state without performing fault simulation and simulation.

After the steps 101 to 106 provided in this embodiment, the fault simulation and the service traffic distribution prediction under the condition of dynamic change can be completed by using a real network management and control system without affecting the existing network hardware, topology and service operation.

Further, steps 101 to 106 may be executed multiple times, and different fault conditions are set in step 101 to simulate service traffic distribution conditions in different fault scenarios for reference in service traffic planning.

In the method for simulating and predicting the network traffic, the simulated fault signal sequence is generated by software, so that the condition that the network topology dynamically changes along with the fault in a real scene is simulated, and real hardware does not need to participate; by accumulating the network change topology after each fault occurs on the baseline topology, the network management and control system can identify the dependency relationship among the multiple points of faults; receiving a fault signal consistent with a real fault through a network management and control system, and calculating real service routing change to keep flow distribution change caused by simulation consistent with the change of a real network; the network management and control system executes the simulation affair and withdraws after execution, can simulate the real network topology and the service route under the condition of not influencing the network topology and the real service, tracks the change of the flow distribution of the whole network, and outputs the change condition of the flow distribution after each fault occurs according to the time sequence.

Example 2:

based on the method for simulating and predicting the network traffic provided in embodiment 1, the following network traffic simulation and prediction can be performed.

As shown in fig. 5, a schematic diagram of a baseline topology of a network used in this embodiment includes network elements: n1, N2, N3 and N4, including links L1, L2, L3, L4 and L5. Wherein, L1 connects N1 and N2, L2 connects N2 and N3, L3 connects N1 and N4, L4 connects N3 and N4, and L5 also connects N3 and N4. The network runs services B1 and B2, the working routes of which are distributed as follows, where W denotes the working route and P denotes the backup route, and the working route passes through each network element and link in the sequence in turn.

B1＝{W:(N1,L1,N2,L2,N3),P:()}

B2＝{W:(N1,L3,N4,L4,N3),P:()}

A fault signal sequence is generated for each network element and link in the baseline topology, and a time sequence is used to indicate the time at which a series of faults occur, denoted as Ti { Ti1, Ti 2.., Tin }, where T denotes a fault signal and i denotes the network element or link number at which the fault occurred, according to step 101. In this example, a fault on N2, L4 was simulated.

According to step 201, fault time sequences are generated for N2 and L4, assuming that the initial time of no fault is 0, the fault occurrence time is generated by a periodic function generator:

the failure occurrence time sequence of N2 is TN2 ═ 1s,3s,5 s.

The failure occurrence time sequence of L4 is TL4 ═ 2s,4s,6 s.

At least one piece of failure data is generated for each network element and each link, according to step 202, as D i { Di1, Di 2., Din }, where D denotes the failure data and i denotes the failed network element or link label. The fault data consists of positioning information and fault codes, wherein the positioning information comprises a network element ID, a link ID, a single disk ID and a port number. Multiple shares of DN2 and DL4 may be generated to correspond to each point in time of failure in the sequence of time of failure.

For N2, each DN2 { (network element ID ═ ID of N2, link ID ═ 0, single disk ID ═ 0, port number ═ 0), failure code };

for L4, each DL4 { (network element ID 0, link ID L4, single disk ID 0, port number 0), failure code }.

In this embodiment, the failure types are failure occurrence and failure recovery, a failure code corresponding to the failure occurrence is DOWN, and a failure code corresponding to the failure recovery is UP.

According to step 203, for each network element or each link, the fault data generated in step 202 is in one-to-one correspondence with each time point in the fault time sequence, and a fault signal sequence is generated, which is marked as Si { (Ti, Di) } { (Ti1, Di1), (Ti2, Di2),. ·, (Tin, Din) }, wherein S denotes the fault time sequence, i denotes the index of the network element or link, and (Ti, Di) denotes that the network element or link i has a Di fault at the Ti time point.

In an actual network environment, faults on N2 and L4 oscillate, and fault occurrence and fault recovery alternately occur, namely DOWN and UP alternately occur.

For N2, SN2 { (1s, DN21), (3s, DN22), (9s, DN25) }, where DN21 is DOWN and DN22 is UP, alternating;

for L4, SL4 { (2s, DL41), (4s, DL42), (8s, DL44) }, where DL41 is DOWN and DL42 is UP, alternating.

According to step 204, the entire fault signal sequence is serially output to the topology simulation module in the time order in which each network element and each link generate a fault signal.

In order to acquire the faults of all network elements and links according to the time sequence of the faults, signals of all the network elements and the links are sequenced according to time, and finally output in series so as to simulate the situations of fault occurrence and receiving processing in a real network.

Specifically, the Si of ALL the failed network elements and links are merged and sorted according to the time sequence to form a complete fault signal sequence, which is denoted as S _ ALL ═ F ({ S1, S2,.. Sn }), where the function F is a merged and sorted function, and the result is, for example, S _ ALL { (T11, D11), (T12, D12),.., (Tij, Dij),.., (Tkm, Dkm) }, where T11, T12, Tij, Tkm are arranged in the time sequence.

The oscillating signals SN2 and SL4 on N2 and L4 are ordered in time to obtain:

S_ALL＝{(1s,DN21),(2s,DL41),(3s,DN22),(4s,DL42),...,(8s,DL44),(9s,DN25)}。

and sequentially outputting the fault signals in the S _ ALL to simulate the fault signals generated by random faults on N2 and L4 in a real network environment.

After the fault signal is simulated, the topology at which no fault has occurred is recorded as a baseline topology, denoted G0, according to step 102. Each fault signal in the S _ ALL is acquired in turn, and the corresponding network element and link are changed on the baseline topology, that is, a changed topology map after each fault occurs can be acquired.

According to step 301, a baseline topology G0 is obtained that records network element and link status at the initial state. Corresponding to the fault type, when no fault occurs, all network elements and links are in normal working state and are recorded as UP.

The baseline topology G0 { (N, UP), (L, UP) }.

Wherein (N, UP) { (N1, UP), (N2, UP), (N3, UP), (N4, UP) } indicates that the initial state of each network element is UP;

(L, UP) { (L1, UP), (L2, UP), (L3, UP), (L4, UP), (L5, UP) }, which indicates that the initial state of each link is UP;

the complete representation of the baseline topology is: g0 { (N, UP), (L, UP) } { (N1, UP), (N2, UP), (N3, UP), (N4, UP) }, { (L1, UP), (L2, UP), (L3, UP), (L4, UP), (L5, UP) }.

According to step 302, the fault signal is converted into a change of network element and link. The fault code in the fault data Dij corresponding to the fault signal Tij at each moment in S _ ALL { (T, D) } may indicate that a fault occurs DOWN or that a fault recovers UP, and after receiving the fault signal, the states of the network element and the link are modified to the states corresponding to the fault code. When a fault occurs, setting the states of the network element and the link to be DOWN; and when the failure is recovered, setting the states of the network element and the link to be UP.

If Dij indicates that the network element is in fault, the network element state at the moment Tij is (Nij, DOWN);

if Dij indicates that the network element fault is recovered, the network element state at the moment Tij is (Nij, UP);

if Dij indicates a link failure, the link state at time Tij is (Lij, DOWN);

if Dij indicates that the link failure is recovered, the link state at the time Tij is (Lij, UP);

finally, the fault signal is converted into a sequence of changes of network elements and links, denoted as G _ STATUS, which is in the form of G _ STATUS { (T11, N11, UP | DOWN),. -, (T12, L12, UP | DOWN),. -, (Tij, Nij, UP | DOWN),. -, (Tst, Lst, UP | DOWN), -. -, (Tkm, Nkm, UP | DOWN) }, wherein G _ STATUS represents a set of states of each network element and link in the current network topology, UP | DOWN represents that both may take one value, and the order of the network elements (N) and links (L) is arranged in the order in which the fault signal occurs.

In the present embodiment, the conversion is performed according to the fault signal sequence sets S _ ALL on N2 and L4 to obtain: g _ STATUS { (1s, N2, DOWN), (2s, L4, DOWN), (3s, N2, UP), (4s, L4, UP),. ·, (8s, L4, UP), (9s, N2, DOWN) }.

According to step 303, starting from the baseline topology G0, the topology change G _ STATUS at the time Tij is sequentially taken out from G _ STATUS, and is accumulated on the topology change map at the time of the previous fault occurrence, so as to calculate the network change topology map corresponding to each time Tij. In the present embodiment, for convenience of explanation, the network topology change map is represented by a topology snapshot Gij.

If G _ STATTUSij is (Nij, DOWN), then the state of the network element Nij in G0 is changed to DOWN;

if G _ STATTUSij is (Nij, UP), the state of the network element Nij in G0 is changed to UP;

if G _ STATTUSij is (Lij, DOWN), then the link Lij status in G0 is changed to DOWN;

if G _ STATTUSij is (Lij, UP), the link Lij status in G0 is changed to UP.

After ALL the fault signals in S _ ALL are acquired, the obtained topology snapshot Gij is as follows.

G0＝{{(N1,UP),(N2,UP),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}

G1s＝{{(N1,UP),(N2,DOWN),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}

G2s＝{{(N1,UP),(N2,DOWN),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,DOWN),(L5,UP)}}

G3s＝{{(N1,UP),(N2,UP),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,DOWN),(L5,UP)}}

G4s＝{{(N1,UP),(N2,UP),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}

G5s＝{{(N1,UP),(N2,DOWN),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}

G6s＝{{(N1,UP),(N2,DOWN),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,DOWN),(L5,UP)}}

G7s＝{{(N1,UP),(N2,UP),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,DOWN),(L5,UP)}}

G8s＝{{(N1,UP),(N2,UP),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}

G9s＝{{(N1,UP),(N2,DOWN),...,(N4,UP)},{(L1,UP),(L2,UP),...,(L4,UP),(L5,UP)}}。

After the network change topology at the current time is obtained, according to step 103, the service affected by the fault signal at each time Tij is calculated through G _ status, and the fault service is recorded as Bij.

According to step 401, first, a network element or a link corresponding to a fault signal in each G _ status is obtained, where (Nij, UP | DOWN) corresponds to a network element Nij, and (Lij, UP | DOWN) corresponds to a link Lij.

Then according to step 402, all services passing through the failed network element or link are searched.

If the number is (Nij, UP | DOWN), all services passing through the network element Nij are inquired;

if it is (Lij, UP | DOWN), all traffic of the tea goes through the link Lij.

In the baseline topology state, the working route of each service has no fault, so there is no fault service, and in the state G1s after the first fault occurs, only the network element N2 has a fault, and other network elements or links have no fault. According to the working path of the service, the working path of the service B1 passes through N2, so B1 is a failure service, and the unaffected B2 is not a failure service at this time.

According to step 104, according to the topology snapshot Gij after each fault signal is sent, the route on the network topology change diagram at the corresponding moment of each service Bij is recalculated, and the route of the service is switched through the simulation transaction, and the simulation transaction is suspended and is not submitted. If the working route of the Bij appears a network element or a link DOWN for the first time, a standby route is searched; if the DOWN of a certain network element or a certain link appears in the backup route of the Bij for the first time, a new backup route is searched; if all network elements and links UP appear in the working route of the Bij for the first time and a standby route exists, switching the working route from the standby route and deleting the standby route; if the working or backup route of Bij is otherwise the case, then the route is not recomputed.

When the network topology state is G1s, the service affected by the DOWN of N2 is B1, the service is not affected by L4, the working route is switched to the standby route in B1, and if the recalculated standby route is (N1, L3, N4, L4, N3), the route of the service at this time changes as follows:

B1＝{W:(N1,L1,N2,L2,N3),P:(N1,L3,N4,L4,N3)}

B2＝{W:(N1,L3,N4,L4,N3),P:()}。

when the network topology state is G2s, a service which is not affected by N2, a DOWN of L4 affects services B1 and B2, a backup route recalculation occurs in B1, a working route is switched to a backup route in B2, and it is assumed that the recalculated backup routes are all (N1, L3, N4, L5, and N3), and at this time, the routes of the service change as follows:

B1＝{W:(N1,L1,N2,L2,N3),P:(N1,L3,N4,L5,N3)}

B2＝{W:(N1,L3,N4,L4,N3),P:(N1,L3,N4,L5,N3)}。

when the network topology state is G3s, the UP of N2 affects the service of B1, L4 does not affect the service, all network elements and links of the working route of B1 are restored to the UP state, and the standby route is switched to the working route, where the route change of the service is as follows:

B1＝{W:(N1,L1,N2,L2,N3),P:()}

B2＝{W:(N1,L3,N4,L4,N3),P:(N1,L3,N4,L5,N3)}。

when the network topology state is G4s, the service that is not affected by N2, the UP of L4 affects service B2, all network elements and links of the working route of B2 are restored to the UP state, and the standby route is switched to the working route, where the route change of the service is as follows:

B1＝{W:(N1,L1,N2,L2,N3),P:()}

B2＝{W:(N1,L3,N4,L4,N3),P:()}。

other time point changes and so on.

In another fault simulation scenario of the embodiment of the present invention, if a DOWN of N2 or L4 still exists between G2s and G3s, the route of the service does not respond to the topology change, because there is an existing backup route and the working route has been switched to the backup route, the backup route can be directly used without recalculating the backup route.

According to step 105, the traffic distribution Wij at each topology snapshot Gij is recorded until all topology states in G _ STATUS are traversed. And finishing traversing, and outputting a sequence W { (Tij, Wij) } of the traffic distribution along with time.

The specific calculation of the time Tij traffic distribution Wij is as follows: if Bij recalculates the standby route, increasing the flow of Bij on the new standby route and reducing the flow of Bij on the original working route or standby route; if the backup route of Bij is switched back to the working route, increasing the flow of Bij on the working route and reducing the flow of Bij on the backup route; if the working route or the standby route of the Bij is in other conditions, the flow is not recalculated, and the flow change caused by the topology change is reduced.

G1s, if the service affected by the DOWN of N2 is B1, and the service is not affected by L4, the working route is switched to the backup route in B1, and the recalculated backup route is (N1, L3, N4, L4, N3), the service routing state at this time is:

B1＝{W:(N1,L1,N2,L2,N3),P:(N1,L3,N4,L4,N3)}

B2＝{W:(N1,L3,N4,L4,N3),P:()}。

comparing W and P at B1 before and after G1s, the flow rate of B1 should be decreased at L1, N2, and L2, and the flow rate of B1 should be increased at L3, N4, and L4.

G2s, the service which is not affected by N2, the DOWN of L4 affects services B1 and B2, the backup route recalculation occurs in B1, the working route is switched to the backup route in B2, it is assumed that the recalculated backup routes are all (N1, L3, N4, L5, N3), and the service routing state at this time is:

B1＝{W:(N1,L1,N2,L2,N3),P:(N1,L3,N4,L5,N3)}

B2＝{W:(N1,L3,N4,L4,N3),P:(N1,L3,N4,L5,N3)}。

comparing W and P at B2 before and after G2s, the flow rate at B2 should be decreased at L4 and the flow rate at B2 should be increased at L5.

G3s, the UP of N2 affects the service B1, L4 does not affect the service, all network elements and links of the working route of B1 are restored to the UP state, the standby route is switched to the working route, and the service route state at this time is:

B1＝{W:(N1,L1,N2,L2,N3),P:()}

B2＝{W:(N1,L3,N4,L4,N3),P:(N1,L3,N4,L5,N3)}。

comparing W of B1 with P before and after G3s, L1, N2, L2 should increase the flow rate of B1, and L3, N4, L4 should decrease the flow rate of B1.

G4s, the service that is not affected by N2, the UP of L4 affects service B2, all network elements and links of the working route of B2 recover to the UP state, the standby route is switched to the working route, and the service route state at this time is:

B1＝{W:(N1,L1,N2,L2,N3),P:()}

B2＝{W:(N1,L3,N4,L4,N3),P:()}。

comparing W and P at B2 before and after G4s, the flow rate at B2 should be increased at L4 and the flow rate at B2 should be decreased at L5.

Other time point changes and so on.

In another fault simulation scenario of the embodiment of the present invention, if a DOWN of N2 or L4 still exists between G2s and G3s, the route of the service does not respond to the topology change, and since there is an existing backup route and the working route is switched to the backup route, the traffic distribution does not change.

According to the step 106, after the flow simulation and prediction data of all fault states are acquired, all the suspended simulation transactions in the step 105 are rolled back, the network change topology Gij is restored to the state of G0, the current simulation or prediction is completed, or the next simulation or prediction is continued by returning to the step 101.

By using the method for simulating and predicting the network traffic provided in embodiment 1, the network traffic simulation and prediction under the condition of a fault is performed on the network topology model provided in this embodiment, and the service traffic change condition of the service under the condition that the network is dynamically changed due to the fault is obtained without affecting the hardware and service running state in the network.

Example 3:

on the basis of the method for simulating and predicting the network traffic provided in the foregoing embodiments 1 to 2, the present invention further provides a device for simulating and predicting the network traffic, which can be used for implementing the method described above, as shown in fig. 6, the device is a schematic structural diagram of the device in the embodiments of the present invention. The apparatus for network traffic simulation and prediction of the present embodiment includes one or more processors 21 and memory 22. In fig. 6, one processor 21 is taken as an example.

The processor 21 and the memory 22 may be connected by a bus or other means, such as the bus connection in fig. 6.

The memory 22, which is a non-volatile computer-readable storage medium for a network traffic simulation and prediction method, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the network traffic simulation and prediction methods in embodiments 1-2. The processor 21 executes various functional applications and data processing of the device for network traffic simulation and prediction, that is, implements the methods for network traffic simulation and prediction of embodiments 1 to 2, by running nonvolatile software programs, instructions, and modules stored in the memory 22.

The memory 22 may include high speed random access memory and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some embodiments, the memory 22 may optionally include memory located remotely from the processor 21, and these remote memories may be connected to the processor 21 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

Program instructions/modules are stored in the memory 22 and, when executed by the one or more processors 21, perform the methods of network traffic simulation and prediction in embodiments 1-2 described above, e.g., perform the various steps shown in fig. 1-4 described above.

Those of ordinary skill in the art will appreciate that all or part of the steps of the various methods of the embodiments may be implemented by associated hardware as instructed by a program, which may be stored on a computer-readable storage medium, which may include: a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and the like.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (11)

1. A method for network traffic simulation and prediction is characterized in that:

generating a set of fault signal sequences for at least one network element or link;

acquiring fault signals in fault signal sequences of all network elements and links according to a time sequence, accumulating each acquired fault signal on network topology, and generating a change topological graph after each fault occurs;

acquiring fault service corresponding to each fault signal according to the time sequence of the occurrence of the faults;

calculating the route change of the fault service after each fault occurs according to the change topological graph after each fault occurs, generating a simulation transaction, and suspending the simulation transaction without submitting the simulation transaction;

recording the flow of each fault service after each simulation transaction is executed, acquiring the change of the service flow, and predicting the change of flow distribution in the network;

and after the execution of the simulation transaction is finished, rolling back the suspended simulation transaction.

2. The method for network traffic simulation and prediction according to claim 1, wherein the generating a set of fault signal sequences for at least one network element or link specifically comprises:

generating a set of random fault time sequences for at least one network element or link using a time sequence generating function;

generating at least one piece of fault data for each faulty network element or faulty link;

corresponding each fault data to a time point in the fault time sequence of the network element or the link to generate a fault signal sequence of the network element or the link;

and serially arranging the fault data in the fault signal sequences of all the network elements and the links according to the time sequence.

3. The method for network traffic simulation and prediction according to claim 2, wherein generating a set of random time series of failures for at least one network element or link further comprises:

and adjusting the distribution density of the fault time in the fault time sequence of the network element or the link according to the fault probability curve of each network element or link, so that the fault occurrence probability of the network element or the link in a unit time period is in accordance with the fault probability curve of the network element or the link.

4. The method of claim 1, wherein said associating each failure data with a time point in the time series of failures for the network element or the link further comprises:

the types of the fault signals comprise fault occurrence and fault recovery, and the two types of the fault occurrence and the fault recovery alternately appear according to a time sequence.

5. The method for simulating and predicting network traffic according to claim 1, wherein accumulating each acquired fault on the network topology specifically includes:

acquiring a topology when no fault occurs as a baseline topology;

sequentially acquiring each fault signal, and converting the fault signals into state changes of network elements or links;

starting from the baseline topology, according to the state change of the network element or link corresponding to each fault signal, corresponding state change is carried out on the basis of the previous change topological graph, and the changed change topological graph corresponding to each fault signal is obtained.

6. The method for network traffic simulation and prediction according to claim 1, wherein the acquiring the fault service corresponding to each fault signal specifically includes:

acquiring a network element or a link corresponding to each fault signal;

and searching all services passing through the network element or the link corresponding to the fault signal as fault services.

7. The method for network traffic simulation and prediction according to claim 1, wherein the calculating the route change of the failed traffic after each failure occurs specifically comprises:

if the working route of the fault service has a network element or link fault for the first time, generating a first standby route for the fault service;

if the first standby route of the fault service has a network element or link fault for the first time, generating a second standby route for the fault service;

if the working route of the fault service has the condition that all network elements and links have no fault for the first time, the standby route of the fault service is switched back to the working route.

8. The method for network traffic simulation and prediction according to claim 1, wherein the obtaining of the change of the service traffic specifically includes:

and according to the route change of the fault service, reducing the traffic of the service on the route of the fault service, and increasing the traffic of the service on the new route to which the service is transferred.

9. The method for network traffic simulation and prediction according to claim 1, wherein predicting the change in the traffic distribution in the network specifically comprises:

recalculating the bandwidth occupancy rate of the service according to the change of the service flow corresponding to the fault information;

and recalculating the traffic distribution of the service in the changed network topology according to the network topology change corresponding to the fault information.

10. The method of network traffic simulation and prediction of claim 9, characterized by:

the traffic distribution includes bottleneck nodes of traffic flow and idle points in the network topology.

11. An apparatus for network traffic simulation and prediction, comprising:

comprising at least one processor and a memory, said at least one processor and memory being connected by a data bus, said memory storing instructions executable by said at least one processor, said instructions, after execution by said processor, being adapted to perform the method of network traffic simulation and prediction according to any of claims 1-10.

Images