CN110086677B - Method for configuring time-varying intention in intention network - Google Patents
Method for configuring time-varying intention in intention network Download PDFInfo
- Publication number
- CN110086677B CN110086677B CN201910390467.4A CN201910390467A CN110086677B CN 110086677 B CN110086677 B CN 110086677B CN 201910390467 A CN201910390467 A CN 201910390467A CN 110086677 B CN110086677 B CN 110086677B
- Authority
- CN
- China
- Prior art keywords
- time
- intention
- time slice
- configuration
- intent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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
- H04L41/08—Configuration management of networks or network elements
- H04L41/0893—Assignment of logical groups to network elements
-
- 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
- H04L41/0896—Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities
-
- 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
- H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
- H04L41/5041—Network service management, e.g. ensuring proper service fulfilment according to agreements characterised by the time relationship between creation and deployment of a service
- H04L41/5051—Service on demand, e.g. definition and deployment of services in real time
-
- 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
- H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
- H04L41/5041—Network service management, e.g. ensuring proper service fulfilment according to agreements characterised by the time relationship between creation and deployment of a service
- H04L41/5054—Automatic deployment of services triggered by the service manager, e.g. service implementation by automatic configuration of network components
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a time-varying intention configuration method in an intention network, and provides a heuristic algorithm, wherein the algorithm is a greedy algorithm (MSP) based on maximized single time slice configuration and used for solving a configuration problem supporting multi-strategy time-varying intention, and the MSP algorithm mainly comprises a time slice selection algorithm and a single time slice configuration algorithm. And determining the sequence of the configuration of each time slice by adopting a time slice selection algorithm according to the total weight of the intention in the unconfigured time slices, and then configuring the intention in each time slice by using a single time slice configuration algorithm while considering the resource usage of the whole effective time. The method can be suitable for configuring the time-varying intents with different priorities, and has the characteristics of short time consumption for solving the configuration scheme and high intention configuration success rate.
Description
Technical Field
The invention belongs to the technical field of intention networks, and particularly relates to a time-varying intention configuration method in an intention network.
Background
With the gradual and deep research on Software Defined Networking (SDN) and Network Function Virtualization (NFV), the demand of network users for simplifying network usage and management is increasingly urgent, and the new network structure, such as the intended network (IBN), which allows users to only inform the network of the required network targets, and the automatic implementation and maintenance of the network becomes a hotspot of the research in the field of software defined networking.
The network structure can provide an intention interface for network users, and the network users express the network requirements to be realized through the intention interface. The user informs the network of its own needs, i.e., intentions, which the network automatically fulfills and maintains. The user simply tells the network what is needed and does not worry about how to do so. Because the intention network provides a uniform network interface at the user side, the network is further abstracted, so that a network user can more conveniently formulate network strategies by using an intention language, and the network strategies are automatically realized after the constraints of the network side of the intention network on various resources and other conditions are taken into consideration.
Challenges remain in the design of the intended network: how to reasonably design an intention configuration scheme in the case of limited physical resources at the bottom layer-more intention configurations are configured. Because the configuration of the intent is not simply a translation of the intent expressed by the network user into the corresponding underlying network language, the injection of flow tables in the SDN environment is complete. In practical environments, the underlying physical resources are usually limited, and it is very important to realize the intentions of network users as much as possible and reasonably use the network resources under the condition of limited resources. More, many of the current intents are based on group granularity, i.e., one intention involves more than one network node (EP), but rather a network node group (EPG) consisting of a plurality of network nodes, such as subscriber hosts. Especially for time-varying intentions, i.e. the operation of the EPG in the intention may differ at different times. For such intended configurations, existing configuration schemes are not perfect, and there is room for further improvement in the use of underlying physical resources.
One of the conventional methods for configuring dynamic intents is to configure dynamic intents, i.e., intents including variable intents, using a finite state machine. When configuring time-varying intents using finite state machines, the finite state machine holding the state change for each time-varying intention, reconfigures the intention according to the state of the network when the intention passes from one state to the other state in its finite state machine. The configuration intention can be selected in real time according to the change of the network state when the intention state changes by using a finite state machine; when the method is used, the times of recalculation of the intention configuration are increased due to the increase of the times of switching the intention state, and the calculation resources are consumed; the number of intents increases, i.e., the number of finite state machines that need to be maintained is reduced by employing intent coalescing, but the number of finite state machines that need to be maintained may also increase.
In another method, in some researches, time-varying intentions are configured in a figure form in a time sequence, and all intentions are shown together in a drawing mode in the figure form, so that the relationship among the intentions is clearly understood. All time-varying intents are first time-spent, and since the time-dimension transform nodes of each intention are known, the time slices are divided in such a way that the intersections are taken at the start and end times of each intention to divide the time slices. And then, the divided intentions in each time slice are shown in each time slice graph. Then, starting from the first divided time slice, the required resources are reserved for the intents configured before the current time slice at the future time. The general process is to start with the first time slice of the partition, maximize the configuration of the current time slice, and configure the intent as valid in the previous time slice as possible in the future until the last time slice is configured. In the method, fixing starts from the first time slice, so that the intention configured at the previous time slice is configured at the current time slice and the intention configured at the current time slice is not configured with more weight, and the condition of influencing the intention configuration number on the whole time slice occurs; and the resource use condition of the time-varying intention in the whole time domain is not considered, so that the situation that the intention resource occupies great differences in different time slices to influence the intention configuration number is presented.
Disclosure of Invention
In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for configuring a time-varying intent in an intent network, which solves the above-mentioned problems in the background art.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that: a method of configuring a time-varying intent in an intent network, comprising the steps of:
s1, reading and dividing the time slice objects Tn in the conflict-free time-varying intention set S1;
wherein, the set of time-varying intentions S1 ═ int1, int2, int3,. cndot, intn }, int ∈ S1, I is intent, I ═ 1,2,3,. cndot, I;
a slice object Tn ∈ time, and time { T1, T2, T3,. eta., Tn }, N is a subscript of the slice object, N ═ 1,2,3,. eta, N, time is a set of slice objects;
each time slice object Tn stores an intention set acaIntentlist to be configured, an intention set aceIntentlist to be configured successfully and an intention set notAcIntentlist to be not configured;
s2, initializing a subintention object subint intended to be in each time slice object;
wherein, each subavenient object subint stores an EP pair object stpair and an intention id of the intention inti;
s3, initializing the object stpair for each EP, selecting a corresponding path for the object stpair for each EP through a path selection algorithm, and storing the path in the corresponding object stpair for each EP;
s4, setting the chop object Time currently selected and the Time slice object queue to be configured; initializing a set Aintents and a set Nintents;
wherein the elements in the set ainters are all intent ids successfully configured in the processed slice object; the elements in the set Nintents are all intent ids that have not been successfully configured in the processed slice object;
s5, judging whether all the time slice objects in the time are processed;
if yes, go to step S13;
if not, go to step S6;
s6, judging whether the choose Time of the currently selected Time slice object is empty and whether the queue of the Time slice object to be configured is empty or not;
if yes, go to step S7;
if not, go to step S8;
s7, selecting a Time slice object Tn in the timer through a Time slice selection algorithm according to the path stored in the stpair, storing the Time slice object Tn in the timeout Time, and entering the step S9;
s8, taking the Time slice Ti out of the queue, storing the Time slice Ti in the cache Time, and entering the step S9;
s9, judging whether the chorose Time is configured;
if yes, the choose Time is set to be null, and the step S5 is returned to;
if not, go to step S10;
s10, configuring the intention in the chorose Time through a single Time slice configuration algorithm according to the path saved in the stpair, updating Aintents and Nintents and the acedIntentlist and notAcIntentlist in the Aintents and Nintents, marking the chorose Time as configured, and entering the step S11;
s11, judging whether the intention of the chorose Time and the intention of the adjacent Time slice object are not configured, and whether the intersection of the acedIntentlist and the notAcIntentlist in the chorose Time and the aceIntentlist in the adjacent Time slice object are simultaneously established;
if yes, go to step S12;
if not, the choose Time is set to be null, and the step S5 is returned;
s12, storing the Time slice objects adjacent to the choose Time into the queue, emptying the choose Time and returning to the step S5;
s13, traversing all processed time slice objects in the time set, obtaining an intention set successfully configured according to the accepted Intentlist and the notalcedIntentlist of each time slice object S2, and completing the configuration of the time-varying intention in the intention network.
Further, the method for selecting a corresponding path for each EP pair object stpair through the path selection algorithm in step S3 specifically includes:
a1, according to the maximum required bandwidth Z of the intention in the time slice object, removing the links of which the link bandwidth is less than the required bandwidth Z in the network topology to obtain a network topology structure which is composed of the topology which is greater than the required bandwidth Z and is a path;
a2, determining the diameter d in the current network topology;
a3, determining a path which meets the bandwidth requirement and has the length not exceeding d for each intention in the time slice object;
a4, processing the path number according to the determined path number, and storing the processed path number in the corresponding EP pair object stpair.
Further, the step a4 is specifically:
when the determined path number is smaller than a threshold k and larger than 0, directly storing the determined path in a corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm;
when the determined number of paths is 0, setting the sub-intention no-path mark for the object stabir at the corresponding EP;
and when the determined number of paths is greater than a threshold k, generating random 01 masks for all the determined paths, taking out the paths corresponding to the masks of 1, storing the paths in the corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm.
Further, the method for selecting the time slice object Tn in the timer through the time slice selection algorithm in the step S7 specifically includes:
b1, selecting all the time slice objects which are not configured yet in the time set;
b2, judging whether the number of the selected unconfigured time slice objects is larger than a threshold value j;
if yes, go to step B3;
if not, go to step B4;
b3, sorting the taken time slice objects from big to small according to the total weight of intentions in the time slice objects, taking the first j time slice objects out, and entering the step B4;
b4, determining the number of the configuration intents in each time slice object by using the time slice object selection solving model, selecting the time slice object Tn with the largest number of the configuration intents, and finishing the selection of the time slice object Tn in the timeset.
Further, the time slice object selection solution model in step B4 is:
max∑i∈ints(t)Wi×Ii
wherein, Ii∈{0,1}，
Ki,j∈{0,1}，
j＝1,2,3,...,m；
pi,j,k∈{0,1}，
In the formula, WiIs the weight of intent i;
Iiif the variable is a variable 0-1 for representing the intention i, if the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
ints (t) is all intents valid for a time period t;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying EPj in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
Further, the method for configuring the choose Time through the single Time slice configuration algorithm in step S10 specifically includes:
c1, judging whether the intention id corresponding to the intention in the accountlist of the chorose Time is the same as the intention id corresponding to the intention in the notacocIntentlist of the adjacent processed Time slice object;
if yes, go to step C2;
if not, go to step C3;
c2, removing the intention corresponding to the same intention id in notac Intentry of the adjacent Time slice object in acaose Time acointlist, and entering step C3;
c3, judging whether the intention id corresponding to the intention in the accontlist of the chorose Time is the same as the intention id corresponding to the intention in the accontlist of the adjacent processed Time slice object;
if yes, go to step C4;
if not, go to step C5;
c4, marking an intention corresponding to the same intention id in the acoIntentlist of the adjacent processed Time slice object in the acoIntentlist of the chorose Time, and entering the step C5;
and C5, configuring unconfigured intents in the current acoIntlist of the choose Time through an intention configuration solving model, and completing intention configuration of the choose Time.
Further, the intention configuration solution model in the step C5 is:
wherein the content of the first and second substances,
Ii∈{0,1}，
Ki,j∈{0,1}，
j＝1,2,3,...,m；
pi,j,k∈{0,1}，
i∈{0,1}，
in the formula, WiIs the weight of intent i;
Wjis the weight of intent j;
α is the resource usage factor;
Ribandwidth resources used for intent i;
Rqbandwidth resources used for intent q;
Iiif the variable is a variable 0-1 for representing the intention i, if the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
iin order to represent the 0-1 variable of the added intention priority configuration variable, the configuration time 1 of the intention i does not need to be added, and the configuration time 0 of the intention i needs to be added;
ints (t) is all intents valid for a time period t;
conf (t) is a sub-intent of the partially configured intent in the time period t;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kto characterize the variable 0-1 that satisfies the kth path of EPj in intent i, a1 selects that path and a 0 deselects itSelecting the path;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
Further, the step S13 is specifically:
traversing all processed time slice objects in the timeset, taking the union of the acedIntentlist in all the time slice objects to form a set C1, taking the union of the notacocdIntentlist in all the time slice objects to form a set C2, taking the intersection of the set C1 and the set C2 as C3, and forming an intention set S2 for successful configuration by subtracting the set C3 from the set C1.
The invention has the beneficial effects that:
(1) suitable for configuring time-varying intents of different priorities. The MSP algorithm provided by the invention can maximize the sum of successfully configured intention weights, and the intention priority corresponds to the intention weight, so that the method can be suitable for the intention configurations with different priority weights, and the intention with high priority is configured as much as possible.
(2) The time for solving the configuration scheme is short. In the method, the filtering operation is carried out before the intention configuration of each time slice, so that the intention participating in the configuration in each time slice is reduced, and the time is reduced compared with the existing configuration method.
(3) The intention configuration success rate is high. The algorithm provided by the invention considers the weight of the time-varying intention and the resource consumption of the time-varying intention in the effective time, so that the configuration method can be reasonably obtained under the condition of meeting the constraint of physical resources, and the configuration success rate is improved compared with the conventional configuration method under the same condition.
Drawings
FIG. 1 is a flow chart of a time-varying intention configuration method in an intention network according to the present invention.
Fig. 2 is a flow chart of a path selection algorithm in the present invention.
FIG. 3 is a flow chart of a time slice selection algorithm in the present invention.
FIG. 4 is a flow chart of a single time slice configuration algorithm in the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
The technical problem to be solved by the invention is as follows: on the premise that the expression intention is a time-varying intention related to connection, service function and QoS through the network, the resource of an underlying network is limited, and the intention granularity is known to be group granularity, the time-varying intention is configured, and the configuration number of the time-varying intention is increased as much as possible. For the purpose of expressing connectivity and types of function chains through network services, the configuration problem of the time-varying intent is essentially the problem of whether a path can be found within a limited period of time of the physical resources of the respective network, which will be intended to be realized. The method comprehensively considers the resource occupation situation of each intention in the whole effective time, and configures the intentions by taking the maximum intention configuration quantity as a target in the whole effective time, and configures the configuration quantity of the time-varying intentions as much as possible.
In order to maximize the number of intended configurations in the entire time domain, the following 0-1 integer law optimization model is therefore set:
max∑n∈TP∑i∈ints(n)Wi×Ii,n(1)
wherein the content of the first and second substances,
wherein n is any time slice in TP time period, and when i ∈ nts (n), corresponding ∑ Wi×IiIs the product of the weight of the intent i in the slice object n and the 0-1 variable of the intent i;
Wiis the weight of intent i;
Ii,na variable 0-1 for representing whether the intention i is configured in the effective part of the time slice object n is configured to be 1, otherwise, the variable is 0;
ints (n) is all intents valid in time slice n;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying EPj in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,na bandwidth requirement configured for intention i within a time period t;
CAPl,nbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
In the constraint condition (2), in order to increase the association of time-varying intentions to the time periods, the influence of the last time period on the time period, namely the intention of successful configuration in the last time slice, needs to be increased, and the intention still existing in the current time slice also needs to guarantee successful configuration; the valid portion of the intent that was not successfully configured in the previous time slice does not need to be configured again when it is also valid in the current time slice. Equation (3) also for each time period, successful configuration requires that each EP pair in the intent can be successfully configured. The root of the energy allocation for each EP is of course whether a path can be found that meets the requirements. Thus the constraints for each EP pair are as (6). In the case that the underlying physical resources are limited, the constraint of link bandwidth needs to be considered, that is, for each link, the usage amount of bandwidth may not exceed the bandwidth capacity of each link, as shown in formula (8), the sum of all path bandwidths using link l in the time period is less than the bandwidth capacity CAP of link llIt is noted that, since there are SFC-related intents, the EP may differ in bandwidth across different service functions, so that the bandwidth requirement for each intention is a bandwidth vector, p in the equationi,j,kThe corresponding bandwidth is the bandwidth corresponding to the intention i on the link i.
The problem of the intended configuration, by nature, is that of the resource allocation class, the problem of most complex resource allocation classes being np-hard. Furthermore, as can be seen from the objective optimization model (1), when the effective intentions of the same intention in different time periods are different, the constraint of bandwidth resources exists at the same time, the problem that the maximum value of the number of time-varying intention configurations in the whole continuous time needs to be solved also belongs to the np-hard problem, and the optimal solution is difficult to be obtained in polynomial time. Meanwhile, the time-varying intention configuration algorithm mainly aims to maximize the intention number in each time slice of the intention number configuration algorithm, but the number of the divided time slices has great influence on the solving time of the multi-time slice joint optimization problem. Therefore, the invention provides a heuristic algorithm, which is a greedy algorithm (MSP) based on maximizing single time period configuration and is used for solving the configuration problem supporting multiple strategy intents.
The MSP Algorithm mainly consists of a time slice selection Algorithm (TPS) and a single time slice configuration Algorithm (CST). And determining the sequence of the configuration of each time slice by adopting a time slice selection algorithm according to the total weight of the intention in the unconfigured time slices, and then configuring the intention in each time slice by using a single time slice configuration algorithm while considering the resource usage of the whole effective time. In order to reflect the relation of valid parts of the same intention in different time slices in a single time slice configuration algorithm, the invention introduces an intention priority coefficient for controlling the influence of the intention valid parts which are configured in other time slices on the intention valid parts in the current time slice; the introduction of the resource usage coefficient reflects the impact of the resource usage of the intent in its effective time on the intent configuration.
In order to implement the optimization model, the invention provides a method for configuring time-varying intents in an intention network as shown in fig. 1, which comprises the following steps:
s1, reading and dividing the time slice objects Tn in the conflict-free time-varying intention set S1;
wherein, the set of time-varying intentions S1 ═ int1, int2, int3,. cndot, intn }, int ∈ S1, I is intent, I ═ 1,2,3,. cndot, I;
a slice object Tn ∈ time, and time { T1, T2, T3,. eta., Tn }, N is a subscript of the slice object, N ═ 1,2,3,. eta, N, time is a set of slice objects;
the time slice object holds the valid portion of the intent at that time, i.e., the child intent. Such as: intention a, intention b, intention c in the intention set; intent a is valid in time slices 1,2, intent b is valid in time slices 2,3, and intent c is valid in time slice 3, then time slice object 1 holds the valid portion of intent a in time slice 1 (a sub-intent of intent a), time slice object 2 holds intent b, intent a holds the valid portion of intent a in time slice 2 (a sub-intent of intent b, a sub-intent of intent a), and time slice object 3 holds the valid portion of intent c (a sub-intent of intent c).
Each time slice object Tn stores an intention set acaIntentlist to be configured, an intention set aceIntentlist to be configured successfully and an intention set notAcIntentlist to be not configured;
the duration of the time period is saved in each time slice object;
s2, initializing a subintention object subint intended to be in each time slice object;
each subacute object subunt stores an EP object stpair of the intention inti, an intention id, a bandwidth requirement list bdlist corresponding to the subacute and a specific behavior list actionlist;
s3, initializing the object stpair for each EP, selecting a corresponding path for the object stpair for each EP through a path selection algorithm, and storing the path in the corresponding object stpair for each EP;
s4, setting the chop object Time currently selected and the Time slice object queue to be configured; initializing a set Aintents and a set Nintents;
wherein the elements in the set ainters are all intent ids successfully configured in the processed slice object; the elements in the set Nintents are all intent ids that have not been successfully configured in the processed slice object;
s5, judging whether all the time slice objects in the time are processed;
if yes, go to step S13;
if not, go to step S6;
s6, judging whether the choose Time of the currently selected Time slice object is empty and whether the queue of the Time slice object to be configured is empty or not;
if yes, go to step S7;
if not, go to step S8;
s7, selecting a Time slice object Tn in the timer through a Time slice selection algorithm according to the path stored in the stpair, storing the Time slice object Tn in the timeout Time, and entering the step S9;
s8, taking the Time slice Ti out of the queue, storing the Time slice Ti in the cache Time, and entering the step S9;
s9, judging whether the chorose Time is configured;
if yes, the choose Time is set to be null, and the step S5 is returned to;
if not, go to step S10;
s10, configuring the intention in the chorose Time through a single Time slice configuration algorithm according to the path saved in the stpair, updating Aintents and Nintents and the acedIntentlist and notAcIntentlist in the Aintents and Nintents, marking the chorose Time as configured, and entering the step S11;
s11, judging whether the intention of the chorose Time and the adjacent Time slice object is not configured, and whether intersection of an acedIntentlist and a notAcIntentlist in the chorose Time and the aceIntentlist in the adjacent Time slice object are simultaneously established or not;
if yes, go to step S12;
if not, the choose Time is set to be null, and the step S5 is returned;
s12, storing the Time slice objects adjacent to the choose Time into the queue, emptying the choose Time and returning to the step S5;
s13, traversing all processed time slice objects in the time set, obtaining an intention set successfully configured according to the accepted Intentlist and the notalcedIntentlist of each time slice object S2, and completing the configuration of the time-varying intention in the intention network.
In step S3, since the path of each EP pair is used as a variable in the model, when the topology size and the connectivity of the topology nodes are high or the number of EP pairs increases, the path variable increases sharply, which seriously affects the time for model solution. Therefore, the invention adopts a Path selection algorithm (RPS) to properly reduce the number of paths of EP pairs participating in algorithm calculation.
Therefore, as shown in fig. 2, the method for selecting a corresponding path for each EP pair object stpair through the path selection algorithm specifically includes:
a1, according to the maximum required bandwidth Z of the intention in the time slice object, removing the links of which the link bandwidth is less than the required bandwidth Z in the network topology to obtain a network topology structure which is composed of the topology which is greater than the required bandwidth Z and is a path;
a2, determining the diameter d in the current network topology;
a3, determining a path with a corresponding bandwidth requirement and a length not exceeding d for each intention in the time slice object;
in addition to meeting the above bandwidth requirements, specific service functions need to be met;
a4, processing the path number according to the determined path number, and storing the processed path number in the corresponding EP pair object stpair.
The step a4 specifically includes:
when the determined path number is smaller than a threshold k and larger than 0, directly storing the determined path in a corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm;
when the determined number of paths is 0, setting the sub-intention no-path mark for the object stabir at the corresponding EP;
and when the determined number of paths is greater than a threshold k, generating random 01 masks for all the determined paths, taking out the paths corresponding to the masks of 1, storing the paths in the corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm.
The path of each EP pair in the path selection algorithm RPS adopts the shortest path algorithm with the length of k, and the length ensures that the length of the path does not exceed d, so that the condition that the found path is too long and a loop exists can be avoided. After the paths are selected, a random 01 mask is generated for each EP pair, where the number of 1's in the mask is k, the value of k being the selected path for the next algorithmic computation. The random selection mode can better ensure that paths of a plurality of EP pairs do not select the same link, and a single link becomes a bottleneck link, thereby influencing the successful configuration number of intents in the time slice object.
The time slice selection algorithm in step S7 described above is an embodiment of a greedy strategy for the time-varying intent configuration algorithm. It should be noted that at the beginning of the time-varying intent configuration algorithm, the intent valid part to be configured in each time slice has been preprocessed, that is, the intents in each time slice, the intents with the number of 0 obtained from the path selection algorithm, are included in the non-configurable set in advance. When the unconfigured time periods are sorted according to the intention weight, no intention with an actual path number of 0 participates. That is, the algorithm will not have the j intents with the most selected intention weights, and the number of intents which can be successfully configured is 0, so that the algorithm can be ensured to be correctly carried out
Therefore, as shown in fig. 3, the method for selecting the time slice object Tn in the timeset by the time slice selection algorithm specifically includes:
b1, selecting all the time slice objects which are not configured yet in the time set;
b2, judging whether the number of the selected unconfigured time slice objects is larger than a threshold value j;
if yes, go to step B3;
if not, go to step B4;
b3, sorting the taken time slice objects from big to small according to the total weight of intentions in the time slice objects, taking the first j time slice objects out, and entering the step B4;
b4, determining the number of the configuration intents in each time slice object by using the time slice object selection solving model, selecting the time slice object Tn with the largest number of the configuration intents, and finishing the selection of the time slice object Tn in the timeset.
The intention configuration in the TPS algorithm is consistent with the configuration idea of the time-varying intention configuration main flow, that is, the maximized intention configuration is realized, so the time slice object selection solution model in step B4 is:
max∑i∈ints(t)Wi×Ii(9)
wherein the content of the first and second substances,
in the formula, WiIs the weight of intent i;
Iiif the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
ints (t) is all intents valid for a time period t;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying EPj in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
The solver model can be classified into a 0-1 integer programming problem, and at present, a plurality of mature solving methods exist for the 0-1 integer programming problem, such as a branch-and-bound method, a cut plane method, a 0-1 programming implicit number method and the like. Therefore, the present invention adopts the solver of the modern optimization problem, and can solve the problem quickly.
The single slice configuration algorithm CST in step S10 is to include the intended resource usage in a slice to influence the intended configuration. When configuring the intended active portion in a single time slice, there can be problems with: firstly, after selecting the starting point of the configuration time slice, how to ensure that the configured intentions are still configured in other time slices which are effective; the second is that because of the time-varying nature of the time-varying intent, the bandwidth demand in time slice 1 is easily met but the bandwidth demand in time slice 2 cannot, how to configure an intent within a single time slice takes into account the resource demand over the entire time slice that the intent is intended to be valid to determine which intents should be configured.
Aiming at the two problems, the invention provides an intended configuration algorithm in a single time slice, and the configuration problem of the single time slice is re-modeled in the algorithm. The main purpose of the single intra-time-slice intention configuration algorithm is to increase the priority of the configuration of the intention of the active part configured in other time slices to the intention of the active part in the time slices to be processed, and simultaneously, when configuring the intention in each time slice, the resource usage situation of the intention in all the active time is taken into consideration. The following is a flow chart of the CST algorithm, wherein, the last configured time slice refers to the time slice for putting the current time slice into the queue, and when the current time slice is put into the queue, the corresponding sub-intention set which has been successfully configured in the time slice and the sub-intention set which cannot be configured are also saved, so that the intention of the prior configuration can be marked and the intention which cannot be configured can be directly removed when the next time slice is configured.
Therefore, as shown in fig. 4, the method for configuring the choose Time by using a single Time slice configuration algorithm specifically includes:
c1, judging whether the intention id corresponding to the intention in the accountlist of the chorose Time is the same as the intention id corresponding to the intention in the notacocIntentlist of the adjacent processed Time slice object;
if yes, go to step C2;
if not, go to step C3;
c2, removing the intention corresponding to the same intention id in notac Intentry of the adjacent Time slice object in acaose Time acointlist, and entering step C3;
c3, judging whether the intention id corresponding to the intention in the accontlist of the chorose Time is the same as the intention id corresponding to the intention in the accontlist of the adjacent processed Time slice object;
if yes, go to step C4;
if not, go to step C5;
c4, marking an intention corresponding to the same intention id in the acoIntentlist of the adjacent processed Time slice object in the acoIntentlist of the chorose Time, and entering the step C5;
and C5, configuring unconfigured intents in the current acoIntlist of the choose Time through an intention configuration solving model, and completing intention configuration of the choose Time.
The intention configuration solving model in the above step C5 is:
wherein the content of the first and second substances,
in the formula, WiIs the weight of intent i;
Wjis the weight of intent j;
α is the resource usage factor;
Ribandwidth resources used for intent i;
Rqbandwidth resources used for intent q;
Iiif the variable is a variable 0-1 for representing the intention i, if the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
iin order to represent the 0-1 variable of the added intention priority configuration variable, the configuration time 1 of the intention i does not need to be added, and the configuration time 0 of the intention i needs to be added;
ints (t) is all intents valid for a time period t;
conf (t) is a sub-intent of the partially configured intent in the time period t;
EPs (i) is all EP pairs in intention i;
Ki,j0-of j to characterize the EP that needs to be configured in intent i1 variable, 1 for successful configuration of the EP pair;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying the EP pair j in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
Target formula (25), RiIs the set of resource consumption of intention i in the time slice object where it is located, setting resource consumption in the present invention refers to the product of the time of the intention size in each time slice object and the occupied bandwidth, namely:
Riin order to aim at the sum of the products of the bandwidth demand and the time length of all the EP pairs in each time period, the product of the bandwidth and the time delay is taken as one of the indexes of the network performance in the communication field, and the capacity of a data link is expressed. The product of the length of occupied time and the size of the used bandwidth used in the intention can then be expressed as the resource usage during that time period. And adding the resource use conditions of the time periods to obtain the resource use amount of the intention in the whole time period. Therefore, the invention adopts the resource usage amount of the intention to represent the resource occupation of the intention to the whole time quantum, and avoids the situation that the bandwidth resources used by the intention in the effective time quantum have great difference. For example, intention A is that the bandwidth requirements for all EP pairs in slot 1 are 10Mbps but the bandwidth requirements in its active slots 3, 4, 5 are 100Mbps, so when the intention in slot 1 is configured, intention A is successfully configured but when the intention in slot 1 is configuredIt is difficult to complete the configuration within the other active time slices of intention a. In such a case, increasing the resource consumption of the intent may increase the effectiveness of the intent configuration.
In the target formula (16)iIs a newly added artificial variable, namely an intention priority configuration variable, which is used for distinguishing the intention. For sub-intents in a time slice that needs to be configured, there may be two sub-intents, one being the portion of the first configured intent that is valid in the current time slice, referred to as the first occurring intent; one is the part of the intention in the current time slice that has been configured in the other time slices, called the partially configured intention. Configuration is only intended to be completed if it is intended to be successfully configured within all of the active time slices in which it resides. Therefore, in order to make the partially configured intents as possible configurable within the current time slice, it is necessary to increase the "priority" of the partially configured intents.
In such a case, if a strong constraint exists that the partially configured intent variable I that is forced to be equal to 1, again the number of intents for configuration of the time period cannot be maximized. Based on this, the invention uses the soft constraint of formula (17) to make the configuration possible with the intention of partial configuration as much as possible: where the constraint is for the first-appearing intention, i.e. when the right side of the equation is 0, the intention cannot be met, the sum of Ii in the left equationiAre both 0. When the right side of the equation is 1, Ii andionly one variable value is 1, and the reasonable setting of the variable weight in the optimization target will
The value setting is much less than IiThe weight of (2) can ensure that the intention variable is 1 and the intention priority configuration variable is 0 on the premise of maximizing the intention configuration. For a partially configured intent, the constraint is still equation (18), i.e., for a partially configured intent if its variable IiIs 1 at this timeiAnd IiThere are no constraints between variables, the larger the value in the optimization objective, the more necessaryiAnd IiAll are 1. in this case, the intention weights are the sameIn the case where only one of the partially configured intention and the first-appearing intention is configured, the partially configured intention is preferentially selected, that is, a "priority" that the partially configured intention is different from the first-appearing intention is achieved.
In the optimization objective formula (16), in order to maximize the intended configuration as the optimization objective of primary importance, the intended weight is a main factor, and it is necessary to set a coefficient of the resource usage to be much smaller than that of the intended weight. Because the intention weight and the resource usage are different in magnitude, normalization processing needs to be performed on the weight and the resource usage in the optimization target, the proportion of the current intention weight to the total intention weight of the time slice is used as the intention weight, and the proportion of the current intention resource usage to the intention usage weight in the current time slice is used as the intention resource usage.
The other constraint conditions are consistent with those of a solver in the time slice selection algorithm.
In step S13, specifically, the method includes:
traversing all processed time slice objects in the timeset, taking the union of the acedIntentlist in all the time slice objects to form a set C1, taking the union of the notacocdIntentlist in all the time slice objects to form a set C2, taking the intersection of the set C1 and the set C2 as C3, and forming an intention set S2 for successful configuration by subtracting the set C3 from the set C1.
The invention has the beneficial effects that:
(1) suitable for configuring time-varying intents of different priorities. The MSP algorithm provided by the invention can maximize the sum of successfully configured intention weights, and the intention priority corresponds to the intention weight, so that the method can be suitable for the intention configurations with different priority weights, and the intention with high priority is configured as much as possible.
(2) The time for solving the configuration scheme is short. In the method, the filtering operation is carried out before the intention configuration of each time slice, so that the intention participating in the configuration in each time slice is reduced, and the time is reduced compared with the existing configuration method.
(3) The intention configuration success rate is high. The algorithm provided by the invention considers the weight of the time-varying intention and the resource consumption of the time-varying intention in the effective time, so that the configuration method can be reasonably obtained under the condition of meeting the constraint of physical resources, and the configuration success rate is improved compared with the conventional configuration method under the same condition.
Claims (5)
1. A method for configuring a time-varying intent in an intent network, comprising the steps of:
s1, reading and dividing the time slice objects Tn in the conflict-free time-varying intention set S1;
wherein, the set of time-varying intentions S1 ═ int1, int2, int 3.., int }, int ∈ S1, I is intent, I ═ 1,2, 3.., I;
a slice object Tn ∈ time, and time { T1, T2, T3,. eta., Tn }, N is a subscript of the slice object, N ═ 1,2,3,. eta, N, time is a set of slice objects;
each time slice object Tn stores an intention set acaIntentlist to be configured, an intention set aceIntentlist to be configured successfully and an intention set notAcIntentlist to be not configured;
s2, initializing a subintention object subint intended to be in each time slice object;
wherein, each subavenient object subint stores the EP pair object stpair and intent id of the intent i;
s3, initializing the object stpair for each EP, selecting a corresponding path for the object stpair for each EP through a path selection algorithm, and storing the path in the corresponding object stpair for each EP;
s4, setting the chop object Time currently selected and the Time slice object queue to be configured; initializing a set Aintents and a set Nintents;
wherein the elements in the set ainters are all intent ids successfully configured in the processed slice object; the elements in the set Nintents are all intent ids that have not been successfully configured in the processed slice object;
s5, judging whether all the time slice objects in the time are processed;
if yes, go to step S13;
if not, go to step S6;
s6, judging whether the choose Time of the currently selected Time slice object is empty and whether the queue of the Time slice object to be configured is empty or not;
if yes, go to step S7;
if not, go to step S8;
s7, selecting a Time slice object Tn in the timer through a Time slice selection algorithm according to the path stored in the stpair, storing the Time slice object Tn in the timeout Time, and entering the step S9;
s8, taking the Time slice Ti out of the queue, storing the Time slice Ti in the cache Time, and entering the step S9;
s9, judging whether the chorose Time is configured;
if yes, the choose Time is set to be null, and the step S5 is returned to;
if not, go to step S10;
s10, configuring the intention in the chorose Time through a single Time slice configuration algorithm according to the path saved in the stpair, updating Aintents and Nintents and the acedIntentlist and notAcIntentlist in the Aintents and Nintents, marking the chorose Time as configured, and entering the step S11;
s11, judging whether the intention in the Time slice object adjacent to the chope Time is not configured, and whether an intersection exists between the accuntentlist in the chope Time and the accuntentlist in the Time slice object adjacent to the accuntentlist in the chope Time, and whether an intersection exists between the notacentlist in the chope Time and the accontentlist in the Time slice object adjacent to the notacentlist in the chope Time is established at the same Time;
if yes, go to step S12;
if not, the choose Time is set to be null, and the step S5 is returned;
s12, storing the Time slice objects adjacent to the choose Time into the queue, emptying the choose Time and returning to the step S5;
s13, traversing all processed time slice objects in the time set, obtaining an intention set successfully configured according to the accepted Intentlist and the notalcedIntentlist of each time slice object S2, and completing the configuration of the time-varying intention in the intention network;
the step S13 specifically includes:
traversing all processed time slice objects in the timeset, taking a union of the acquired intantlist in all the time slice objects to form a set C1, taking a union of the notacocdintantlist in all the time slice objects to form a set C2, taking the intersection of the set C1 and the set C2 as C3, and forming a successfully configured intention set S2 by subtracting the set C3 from the set C1;
the method for configuring the choose Time through the single Time slice configuration algorithm in step S10 specifically includes:
c1, judging whether the intention id corresponding to the intention in the accountlist of the chorose Time is the same as the intention id corresponding to the intention in the notacocIntentlist of the adjacent processed Time slice object;
if yes, go to step C2;
if not, go to step C3;
c2, removing the intention corresponding to the same intention id in notac Intentry of the adjacent Time slice object in acaose Time acointlist, and entering step C3;
c3, judging whether the intention id corresponding to the intention in the accontlist of the chorose Time is the same as the intention id corresponding to the intention in the accontlist of the adjacent processed Time slice object;
if yes, go to step C4;
if not, go to step C5;
c4, marking an intention corresponding to the same intention id in the acoIntentlist of the adjacent processed Time slice object in the acoIntentlist of the chorose Time, and entering the step C5;
and C5, configuring unconfigured intents in the current acoIntlist of the choose Time through an intention configuration solving model, and completing intention configuration of the choose Time.
2. The method for configuring time-varying intents in an intention network according to claim 1, wherein the method for selecting a corresponding path for each EP pair object stpair in the step S3 through a path selection algorithm specifically comprises:
a1, according to the maximum required bandwidth Z of the intention in the time slice object, removing the links of which the link bandwidth is less than the required bandwidth Z in the network topology to obtain a network topology structure which is composed of the topology which is greater than the required bandwidth Z and is a path;
a2, determining the diameter d in the current network topology;
a3, determining a path which meets the bandwidth requirement and has the length not exceeding d for each intention in the time slice object;
a4, processing the paths according to the determined path number, and storing the processed paths in the corresponding EP pair object stpair;
the step a4 specifically includes:
when the determined path number is smaller than a threshold k and larger than 0, directly storing the determined path in a corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm;
when the determined number of paths is 0, setting the sub-intention no-path mark for the object stpair in the corresponding EP;
and when the determined number of paths is greater than a threshold k, generating random 01 masks for all the determined paths, taking out the paths corresponding to the masks of 1, storing the paths in the corresponding EP pair object stpair as input data in a time slice selection algorithm and a single time slice configuration algorithm.
3. The method for configuring time-varying intents in an intention network according to claim 1, wherein the method for selecting a time slice object Tn in a timer by a time slice selection algorithm in the step S7 is specifically as follows:
b1, selecting all the time slice objects which are not configured yet in the time set;
b2, judging whether the number of the selected unconfigured time slice objects is larger than a threshold value j;
if yes, go to step B3;
if not, go to step B4;
b3, sorting the taken time slice objects from big to small according to the total weight of intentions in the time slice objects, taking the first j time slice objects out, and entering the step B4;
b4, determining the number of the configuration intents in each time slice object by using the time slice object selection solving model, selecting the time slice object Tn with the largest number of the configuration intents, and finishing the selection of the time slice object Tn in the timeset.
4. The method for configuring time-varying intents in an intention network according to claim 3, wherein the time slice object selection solution model in the step B4 is:
max∑i∈ints(t)Wi×Ii
wherein, Ii∈{0,1}，
In the formula, WiIs the weight of intent i;
Iiif the variable is a variable 0-1 for representing the intention i, if the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
ints (t) is all intents valid for a time period t;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying the EP pair j in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
5. The method for configuring time-varying intents in an intention network according to claim 1, wherein the intention configuration solution model in the step C5 is:
wherein, Ii+i＝Ki,1∩Ki,2...Ki,m，
In the formula, WiIs the weight of intent i;
Wjis the weight of intent j;
α is the resource usage factor;
Ribandwidth resources used for intent i;
Rqbandwidth resources used for intent q;
Iiif the variable is a variable 0-1 for representing the intention i, if the variable is 1, the intention i is successfully configured, and if the variable is 0, the intention i is not configurable;
iin order to represent the 0-1 variable of the added intention priority configuration variable, the configuration time 1 of the intention i does not need to be added, and the configuration time 0 of the intention i needs to be added;
ints (t) is all intents valid for a time period t;
conf (t) is a sub-intent of the partially configured intent in the time period t;
EPs (i) is all EP pairs in intention i;
Ki,jto characterize the 0-1 variable of the EP pair j that needs to be configured in intent i, if 1, then the EP pair is successfully configured;
PATi,jall paths for EP pairs j that can be configured in intent i;
pi,j,kin order to represent a variable 0-1 of the kth path satisfying EPj in the intention i, if the variable is 1, the path is selected, and if the variable is 0, the path is not selected;
path (l) is all paths through link l;
BWi,ta bandwidth requirement configured for intention i within a time period t;
CAPl,tbandwidth capacity for link l in time period t;
links are all links in the current network;
TP is the set of all time periods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910390467.4A CN110086677B (en) | 2019-05-10 | 2019-05-10 | Method for configuring time-varying intention in intention network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910390467.4A CN110086677B (en) | 2019-05-10 | 2019-05-10 | Method for configuring time-varying intention in intention network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110086677A CN110086677A (en) | 2019-08-02 |
| CN110086677B true CN110086677B (en) | 2020-07-17 |
Family
ID=67419663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910390467.4A Active CN110086677B (en) | 2019-05-10 | 2019-05-10 | Method for configuring time-varying intention in intention network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110086677B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114095330B (en) | 2020-07-29 | 2023-05-02 | 华为技术有限公司 | Intention negotiation method and device |
| CN114125034B (en) * | 2020-08-31 | 2023-12-08 | 华为技术有限公司 | Network performance data subscription method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10530697B2 (en) * | 2015-02-17 | 2020-01-07 | Futurewei Technologies, Inc. | Intent based network configuration |
-
2019
- 2019-05-10 CN CN201910390467.4A patent/CN110086677B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110086677A (en) | 2019-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hong et al. | Achieving high utilization with software-driven WAN | |
| CN104396187B (en) | The method and device that Bandwidth guaranteed and work are kept | |
| US6795399B1 (en) | Link capacity computation methods and apparatus for designing IP networks with performance guarantees | |
| CN112822109B (en) | SDN core network QoS route optimization method based on reinforcement learning | |
| CN110086677B (en) | Method for configuring time-varying intention in intention network | |
| EP1005194A2 (en) | Router placement methods and apparatus for designing IP networks with performance guarantees | |
| EP1005195A2 (en) | Network topology optimization methods and apparatus for designing IP networks with performance guarantees | |
| CN107294852B (en) | Network routing method using topology dispersed short path set | |
| CN103746852B (en) | Service routing configuration method and network management equipment | |
| CN109614215A (en) | Stream scheduling method, device, equipment and medium based on deeply study | |
| JP2004336549A (en) | Method and device for band control | |
| CN107454628A (en) | Compete the grouping scheduling method based on statistics load in multiple access | |
| Natalino et al. | Machine learning aided orchestration in multi-tenant networks | |
| CN114071582A (en) | Service chain deployment method and device for cloud-edge collaborative Internet of things | |
| CN109672621A (en) | A kind of method and apparatus selecting transmission path for vpn service | |
| Wu et al. | Network congestion avoidance through packet-chaining reservation | |
| US7417999B1 (en) | Priority propagation in a multi-level scheduling hierarchy | |
| Mitzel et al. | A study of reservation dynamics in integrated services packet networks | |
| CN107360483B (en) | Controller load balancing algorithm for software defined optical network | |
| CN110677352A (en) | Single-switch single-controller transmission control method in software defined network | |
| CN113965616B (en) | SFC mapping method based on VNF resource change matrix | |
| Chuprikov et al. | Priority queueing for packets with two characteristics | |
| Bensalem et al. | Towards optimal serverless function scaling in edge computing network | |
| US9185042B2 (en) | System and method for automated quality of service configuration through the access network | |
| Narman et al. | Multi class traffic analysis of single and multi-band queuing system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |