CN114726729B - Admission control method and system of wireless access network facing network slicing - Google Patents
Admission control method and system of wireless access network facing network slicing Download PDFInfo
- Publication number
- CN114726729B CN114726729B CN202210356501.8A CN202210356501A CN114726729B CN 114726729 B CN114726729 B CN 114726729B CN 202210356501 A CN202210356501 A CN 202210356501A CN 114726729 B CN114726729 B CN 114726729B
- Authority
- CN
- China
- Prior art keywords
- slice
- time
- admission control
- base station
- service
- 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
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
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W48/00—Access restriction; Network selection; Access point selection
- H04W48/08—Access restriction or access information delivery, e.g. discovery data delivery
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W48/00—Access restriction; Network selection; Access point selection
- H04W48/16—Discovering, processing access restriction or access information
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The application discloses an admission control method and a system of a wireless access network facing a network slice, wherein the admission control method of the wireless access network facing the network slice specifically comprises the following steps: initializing the time; acquiring an optimal slice level admission control decision; acquiring an optimal base station level admission control decision; updating a service request queue; acquiring time average profit; judging whether the maximum moment is reached; and if the maximum time is reached, outputting a time average profit, an optimal slice level admission control decision and an optimal base station level admission control decision. The method of the application can make the NSACF controller of the RAN slice correctly control the admission quantity of the service requests of different slices of users in a dynamic network, and reasonably arrange different base stations in the system to admit the requests and provide services for the requests on the premise of ensuring the stability of the RAN slice admission control system and the SLA of each slice, and simultaneously make the RAN slice system obtain the best time average profit.
Description
Technical Field
The application relates to the field of mobile communication, in particular to an admission control method and system of a wireless access network facing network slicing.
Background
Admission control in the field of communications means that while guaranteeing the quality of service (QoS) of a user, an access service request is arranged according to the current resource situation of the communications network, i.e. a decision is made to admit or reject a new service request in the communications network. A 5G network slices deployed Radio Access Networks (RANs), abbreviated as RAN slices. In RAN slicing, it is necessary to effectively control the number of access users and limit different kinds of slicing service requests of users, and mainly use a network slice admission control function (nsacp) controller to effectively manage overload, so as to avoid the influence of the congested request on slicing performance.
However, in the current RAN slice admission control method, the admission control level concerned only focuses on the admission of the base station to different users, and does not consider the specific network performance requirements of different slice services of the users. In the 5G era, each user may have different kinds and different numbers of requests at the same time, and since each network slice has different Service Level Agreements (SLAs), and user QoS is an important component of the SLAs, the RAN slicing system needs to increase slice-level admission control, so that the system scientifically and reasonably controls the admission numbers of different slice requests of the users. In addition, since different network slices are different for the deployment of base stations, this results in a limitation of base stations that are accessible to users, and thus there is a new impact on admission control at the base station level.
Aiming at the current RAN slice admission control technology, the following main defects exist:
1. the related art and methods of slice level admission control are lacking. In network slicing applications, at the same time, each user may have different kinds and different numbers of requests, corresponding to services of different kinds of slices, so that the RAN side system needs to scientifically and reasonably distinguish different slice requests of the user and determine the number of admission and rejection of each slice service request. The prior art mainly focuses on a RAN slice service guarantee mechanism and resource management thereof, and aims to realize an optimal slice deployment mode, but the admission control problem of different slice type requests of users is rarely concerned.
Second, cellular network admission control is shifted to a base station-user-slice three-layer admission control mode. In the application scene without network slice deployment, the base station directly decides whether to accept the request of the user; in the current network slicing deployment application scenario, the base station needs to perform admission control on a specific slicing request according to a specific slicing service of a user and provide service ensured by SLA. Since the deployment of slices is not fully covered, i.e. some base stations do not deploy the slice services intended by the users, the existing base station level admission control techniques do not fully take into account the impact of this; meanwhile, if the base station cannot guarantee the SLA of the slice service of the user and cannot accommodate the service request, many existing admission control technologies often ignore this key point.
3. Heterogeneous network admission control faces challenges. The evolution of cellular network architecture brings new challenges, and most of today's 5G networks are heterogeneous cellular networks, which not only comprise traditional macro base stations, but also add new micro base stations, pico base stations, home base stations and the like. The decision of base station level admission control becomes challenging due to the significant differences in transmit power of macro and micro base stations, pico base stations, etc.
4. The admission control problem of dynamic networks remains a problem. The previous admission control problem is simpler for the mobility of the user because the user can be served anywhere in the scenario. However, in the 5G RAN slice architecture, the deployment of slices to base stations is incomplete, and the movement of users may affect base stations that can accommodate the relevant service requests.
Therefore, how to provide an admission control method for combining and optimizing a slice level and a base station, so as to maximize the profit of the RAN slice admission control system is a problem that needs to be solved by those skilled in the art.
Disclosure of Invention
Aiming at the defects existing in the existing RAN slice admission control technology, the application provides an admission control method for optimizing slice level and base station combination for a dynamic heterogeneous network scene, so that the profit of the RAN slice admission control system is maximized. For the slice level admission control method, the NSACF controller of the RAN slice admission control system can scientifically and reasonably make a correct decision on the admission quantity of different slice requests of users, so that the admission quantity of different users and different types of slice requests at each moment is maximized while meeting slice SLA indexes and wireless resource total quantity constraint, and the system obtains the maximum income. For the base station level admission control method, the purpose is that the NSACF controller can scientifically and reasonably select different base stations to provide services for different slicing requests of users, and the cost paid by a control system for the base stations is ensured while the service speed as fast as possible is ensured.
In order to solve the above problems, the present application provides a method for controlling admission of a radio access network facing to network slicing, which specifically includes the following steps: s1: initializing the time; s2: responding to the initialization of the completion time to acquire an optimal slice level admission control decision; s3: in response to obtaining the optimal slice level admission control decision, obtaining an optimal base station level admission control decision; s4: updating a service request queue in response to obtaining an optimal base station level admission control decision; s5: in response to performing the update of the service request queue, obtaining a time average profit; s6: judging whether the maximum moment is reached or not in response to the time average profit acquisition; if the maximum time is not reached, adding 1 to the time, and re-executing the steps S2-S5; s7: and if the maximum time is reached, outputting a time average profit, an optimal slice level admission control decision and an optimal base station level admission control decision.
As above, wherein the optimal slice level admission control decision is obtained in response to the initialization of the completion time instant, comprising the sub-steps of: taking the initialized moment, the service request quantity and the service request queue information as input; and obtaining the optimal slice level admission control decision according to the input.
As above, wherein n is set prior to obtaining the optimal slice level admission control decision ik (t) expressed as the number of service requests of user k with respect to slice i received by the system at time t, assuming that all of the slice service requests received by the system have base station reception satisfying their transmission rate constraints, i.e., there must be a base station J e J capable of providing services guaranteeing slice service level agreements, so 0.ltoreq.n ik (t)≤N ik (t), wherein N ik (t) is expressed as the number of service requests by user k with respect to slice i at time t.
As above, in the process of obtaining the optimal slice level admission control decision, the Lyapunov function L (t) related to the service request queue is defined first, and then the Lyapunov drift delta (t) is further defined, the rootDetermining a drift minus profit expression for each time t and an upper limit value of the drift minus profit expression according to Lyapunov drift delta (t), and for the number n of service requests of user k about slice i at time t which are accepted by the system ik (t) solving to obtain an optimal slice level admission control decision
As above, in the process of obtaining the optimal slice level admission control decision, the upper limit f (t) of the drift minus profit expression at each time t is specifically expressed as:
wherein B is defined asFor upper bound of number of service requests, J represents the number of base stations, I represents the number of network slices, +.>Indicating that base station j can be K e K for each user ij Slice i of (t) requesting the offered service speed, < >>Representing the service speed +.>Upper limit of (2). Wherein n is ik (t) is expressed as the number of service requests received by the system by user k with respect to slice i at time t, E { } is expressed as the average of the bracketed portions at different times, Q ik (t) is the service request queue for user k with respect to slice i at time t, Q (t) represents the vector of all service request queues, V is a trade-off factor, gamma i Request data volume, d, for each service of slice i ij (t) the cost of providing slice i service at base station j at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station j admits user k at time t and serves it with slice i.
As aboveWherein, in the process of obtaining the optimal slice level admission control decision, E { Σis applied to a non-constant part i∈I ∑ k∈K n ik (t)(Q ik (t)-Vγ i ) -Q (t) }, converting the average value at all times of minimization thereof into a value at each time of minimization thereof, specifically expressed as
Thereby directly obtaining the slice level admission control optimization variable { n } ik (t), solving result of K E K, I E I isI.e. to obtain the optimal slice level admission control decision.
As above, in the process of obtaining the optimal base station level admission control decision, for the non-constant partConverting the average value of the minimized time to the value of the minimized time, which is expressed as:
where V is a trade-off factor, d ij (t) is the cost of providing slice i service at base station j at time t,indicating that base station j can be K e K for each user ij The service speed, Q, of the slice i request provisioning of (t) ik (t) user k is a service request queue for slice i at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station j admits and provides user k at time tSlice I serves, J denotes the set of base stations, and I denotes the set of network slices.
And then, carrying out iterative solution by using a solution method based on a greedy strategy, thereby obtaining an optimal base station level admission control decision.
As above, wherein the service request queue Q ik The updating of (t) is specifically expressed as:
Q ik (t+1)＝max[Q ik (t)-ε ik (t),0]+n ik (t)
wherein Q is it (t+1) is the service request queue after updating, n ik (t) is expressed as the number of service requests by user k with respect to slice i that are admitted by the system at time t ε ik And (t) represents the service speed provided by the system for slice i of user k at time t.
As above, wherein the time average profitThe concrete steps are as follows:
where P (t) represents the total profit of the system at time t, t represents the time, and λ represents each time.
The admission control system of the wireless access network facing network slicing specifically comprises an initialization unit, an optimal slice level admission control decision acquisition unit, an optimal base station level admission control decision acquisition unit, a service queue updating unit, a time average profit acquisition unit, a judging unit and an output unit; an initialization unit for initializing the time; an optimal slice level admission control decision acquisition unit configured to acquire an optimal slice level admission control decision; an optimal base station level admission control decision acquisition unit configured to acquire an optimal base station level admission control decision; a service queue updating unit for updating the service request queue; the time average profit obtaining unit is used for obtaining time average profit; the judging unit is used for judging whether the maximum moment is reached, if the maximum moment is not reached, adding 1 to the moment, re-acquiring the optimal slice level admission control decision, acquiring the optimal base station level admission control decision, updating the service request queue and acquiring the time average profit; and the output unit is used for outputting time average profit, optimal slice level admission control decision and optimal base station level admission control decision if the maximum moment is reached.
The application has the following beneficial effects:
the method of the application can make the NSACF controller of the RAN slice correctly control the admission quantity of the service requests of different slices of users in a dynamic network, and reasonably arrange different base stations in the system to admit the requests and provide services for the requests on the premise of ensuring the stability of the RAN slice admission control system and the SLA of each slice, and simultaneously make the RAN slice system obtain better time average profit.
In addition, the method can effectively adapt to the application scene of the differential deployment of the network slice, and can adapt to the situation that the base station covered by the slice is limited compared with the prior admission control scheme. In addition, the method is suitable for the dynamic heterogeneous network, has better expandability, and can be more and more obvious along with the increase of the scale of the RAN slice compared with other methods.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings may be obtained according to these drawings for a person having ordinary skill in the art.
Fig. 1 is a schematic diagram of a service provisioning scenario of a radio access network for network slice deployment according to an embodiment of the present application;
fig. 2 is an internal structural diagram of an admission control system of a radio access network for network slicing according to an embodiment of the present application;
fig. 3 is a flowchart of a method for controlling admission of a radio access network for network slicing according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The application mainly solves the problem of admission control of the RAN slice, realizes the joint design of slice level admission control and base station level admission control in the system, and ensures that the RAN slice admission control system in the prior art obtains the highest time average profit.
Scene assumption: considering the design of a service guarantee scenario of a radio access network in which network slices are deployed, as shown in fig. 1, one slice may be deployed on a plurality of base stations, and one base station may also deploy a plurality of network slices. Definitions i, j, k represent network slices, base stations and subscriber units, respectively. Let i= {1,2,..i., i.. I is the set of network slices, j= {1,2,..j } is a set of base stations, k= {1,2,..k, K } is a set of users, and the total number of three sets is I, J, K, respectively. For deployment relationship of slice and base station, define I j For the set of slices deployed at base station j, its number is I j . Similarly, define J i A set of base stations deployed for slice i, the number of which is J i . The user location, the type and number of slicing service requests of the user may all change dynamically over time. The time-varying interval is denoted as T, the time of change is denoted as 0,1,2, once again, T. At time t, corresponding slicing services are provided by admission control of different slicing requests of each user. Each user will move randomly according to a certain rule, being in a different position at each instant t. During the connection, each time instant is equally divided into a plurality of time slots, each time slot being denoted by τ.
Example 1
As shown in fig. 2, the admission control system of a radio access network facing network slicing provided by the present application can make the RAN slicing system in the prior art obtain the best admission control decision and the highest time average profit.
The system of the application specifically comprises: an initialization unit 210, an optimal slice level admission control decision acquisition unit 220, an optimal base station level admission control decision acquisition unit 230, a service queue update unit 240, a time average profit acquisition unit 250, a judgment unit 260, and an output unit 270.
The initializing unit 210 is configured to perform time initialization.
An optimal slice level admission control decision acquisition unit 220 is connected to the initialization 210 for acquiring an optimal slice level admission control decision.
The best base station level admission control decision obtaining unit 230 is connected to the best slice level admission control decision obtaining unit 220, for obtaining the best base station level admission control decision.
The service queue updating unit 240 is connected to the optimal base station level admission control decision obtaining unit 230, and is configured to update the service request queue.
The time average profit obtaining unit 250 is connected to the service queue updating unit 240 for obtaining time average profit.
The judging unit 260 is connected to the time average profit obtaining unit 250 and the optimal slice level admission control decision obtaining unit 220, respectively, for judging whether the maximum time is reached. And if the maximum moment is not reached, adding 1 to the moment, and re-acquiring the optimal slice level admission control decision.
The output unit 270 is connected to the judging unit 260, and is configured to output the time average profit, the best slice level admission control decision, and the best base station level admission control decision if the maximum time is reached.
Example two
As shown in fig. 3, the method for controlling admission of a radio access network for network slicing provided by the present application maximizes the profit of the existing RAN slice admission control system. The objective of profit maximization of the RAN slice admission control system is to carry out slice-level admission control and base station-level admission control at each moment under the constraint of maximum value and stability of user request, so as to maximize the time average profit of the system.
Based on the above thought, the application provides a radio access network admission control method for network slicing, which specifically comprises the following steps:
step S310: the time is initialized.
Specifically, where the time of initialization, let t=0.
Step S320: in response to the initialization of the completion time, an optimal slice level admission control decision is obtained.
Slice level admission control decisions relating to the variable n ik A problem of the number of admission requested by slice i of user k of (t). The present embodiment defines a drift-minus-profit expression based on lyapunov optimization theory, and minimizes the upper limit of the drift-minus-profit expression at each time t, thereby obtaining an optimal slice level admission control decision.
Wherein step S320 specifically includes the following sub-steps:
step S3201: the time of initialization, the number of service requests, and service request queue information are input.
Specifically, the initialization time t, the number of service requests { N }, will be ik (t), K e K, I e I, and service request queue Q (t) information as input.
N ik (t) is expressed as the number of service requests of user k with respect to slice i at time t, N ik (t) is dynamic and varies with time. Suppose that user k's slice i service request number N ik (t) is independent and equidistributed, since the number of service requests of users with respect to slices is often limited, N ik The distribution of (t) is bounded.
Q ik (t) is the service request queue for user k with respect to slice i at time t, Q (t) represents the vector of all service request queues.
Step S3202: and obtaining the optimal slice level admission control decision according to the input.
Setting n before obtaining the optimal slice level admission control decision ik (t) is expressed as user k about the slice at time ti number of service requests admitted by the system. Here, it is assumed that all slice service requests admitted by the system have base station admission that can meet the transmission rate QoS constraint thereof, i.e. the base station J e J must exist to provide the service of guaranteeing slice SLA. Obviously, 0.ltoreq.n ik (t)≤N ik (t). Thus, the vector { n } ik (t), K epsilon K, I epsilon I } is a slice level admission control optimization variable of a radio access network service guarantee scene of network slice deployment. The best slice level admission control decision is actually for n ik (t) solving.
Based on Lyapunov optimization theory, the method firstly defines a Lyapunov function L (t) related to a service request queue as follows:
wherein Q is ik And (t) is a service request queue for user k with respect to slice i at time t.
Further, a one-step condition lyapunov shift Δ (t) is defined as:
wherein Q (t) represents the vector of all service request queues, L (t) represents the lyapunov function of the service request queue, L (t+1) represents the lyapunov function of the service request queue corresponding to time t+1, and E { } represents the average value of the bracketed part at different times in the solution.
Still further, the drift-minus-profile expression at each time t is:
where delta (t) represents lyapunov drift, V is a trade-off factor, a non-negative constant, used to trade-off between service request queue and profit. Q (t) represents the vector of all service request queues, P (t) represents the total profit of the system at time t, and E { } represents the average of the bracketed parts at different times.
Specifically, the total profit P (t) of the RAN system at time t is the total slicing service data amount U (t) of all users minus the total cost D (t) of the system at time t to satisfy the user slicing service request. I.e. the total profit P (t) of the system at time t is specifically expressed as:
P(t)＝U(t)-D(t)
specifically, the total income of the system at the time t is all slice service data volume of all users, specifically expressed as:
wherein, gamma i For each service request data volume of slice i, n ik (t) is expressed as the number of service requests by the system for slice i for user k at time t.
The cost function D (t) of the system at the time t is the total cost for satisfying the user slice service request, and is specifically expressed as:
wherein d is ij And (t) is the cost of providing slice i service at base station j at time t.
d ij The size of (t) is divided into three cases: if slice i is not deployed at base station j, i.eThen at any time there is zero cost, i.e. d ij (t) =0; if slice i is deployed on base station J, i.e., J e J i But at time t base station j does not admit any user's slice i request, i.e. +.>Then a fixed cost d is charged ij (t)＝d 0 The method comprises the steps of carrying out a first treatment on the surface of the If slice i is deployed at base station j, i.ej∈J i And at time t base station j admits the user set of slice i not empty, i.e. +.>The system is +.>And a certain cost is charged. Thus d ij (t) is expressed as
Wherein beta is i Is a constant representing the cost of the system providing a unit slice i service, a ijk (t) ∈ {0,1} is a representation that base station j admits user k at time t and serves it with slice i,indicating that base station j can be K e K for each user ij Slice i of (t) requests the speed of service provided.
Further, the upper limit of the draft-minus-profile expression at each time t is specifically expressed as:
wherein B is defined asFor upper bound of number of service requests, J represents the number of base stations, I represents the number of network slices, +.>Indicating that base station j can be K e K for each user ij Slice i of (t) requesting the offered service speed, < >>Representing the sameService speed->Upper limit of (2). Wherein n is ik (t) is expressed as the number of service requests received by the system by user k with respect to slice i at time t, E { } is expressed as the average of the bracketed portions at different times, Q ik (t) is the service request queue for user k with respect to slice i at time t, Q (t) represents the vector of all service request queues, V is a trade-off factor, gamma i Request data volume, d, for each service of slice i ij (t) the cost of providing slice i service at base station j at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station j admits user k at time t and serves it with slice i.
At each time t, the value of the drift-minus-profile expression is no greater than f (t) under any feasible slice-level admission control method.
The main solution idea of slice-level admission control decisions is therefore to minimize f (t) in each instant t. The NSACF controller is enabled to observe related system information, namely an initial time t, service queue information Q (t) and service request quantity { N }, wherein the initial time t is a time period ik (t), k.epsilon.K, i.epsilon.I } is known, and assume { a } ijk (t) } has a fixed value, so that f (t) has a partial formula that is constant.
For the non-constant part E { Σof the above formula i∈I ∑ k∈K n ik (t)(Q ik (t)-Vγ i ) Q (t) } can be converted to a value that minimizes the average of all its moments, by:
based on this, the nsaacf controller can directly obtain the slice level admission control optimization variable { n } ik (t), solving result of K E K, I E I is
Step S3203: outputting the obtained optimal slice level admission control decision.
Wherein the result obtained is solvedI.e. the optimal slice level admission control decision.
Step S330: in response to obtaining the optimal slice level admission control decision, obtaining an optimal base station level admission control decision.
Base station level admission control decisions relating to variable a ijk The base station j of (t) has a choice question for the admission of and servicing the request of slice i of user k. The main idea is still to minimize the upper limit value f (t) of the drift-minus-profile expression.
The nsacp is enabled to observe relevant system information, namely an initial time t, service queue information Q (t), and positions of all users. And according to { n } which has been determined in step S320 ik (t) } some of the expressions in f (t) in step S3202 are constants, and the part to be solved isConverting the average value of the minimized all the moments into the value of the minimized each moment, wherein the specific expression is as follows:
where V is a trade-off factor, d ij (t) is the cost of providing slice i service at base station j at time t,indicating that base station j can be K e K for each user ij The service speed, Q, of the slice i request provisioning of (t) ik (t) user k is a service request queue for slice i at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station J admits user k at time t and serves it with slice I, J indicating the set of base stations and I indicating the set of network slices.
And then carrying out iterative solution by utilizing a solution method based on a greedy strategy.
Wherein step S330 specifically includes the following substeps:
step S3301: the initialization time, the user position and the service request queue information are used as input.
Specifically, the time t, the user position, and the service request queue information Q (t) are input.
Step S3302: an unacceptable combination is determined from the input.
Specifically, determining the unacceptable combination includes determining whether base station j can accommodate the slice i request of user k. It can be determined in detail from the transmission rate whether base station j can accommodate the slice i request of user k.
And before determining the transmission rate, determining the downlink rate obtained by the slice i of the user k requesting access to the base station j.
Wherein each user moves randomly according to a certain rule and is at a different position at each time t. During the connection, each time instant is equally divided into a plurality of time slots, each time slot being denoted by τ. Definition c jk (t, τ) the unit bandwidth achievable rate of user k connection with base station j at the long time scale of time t. The slice i of user k requests access to the downlink rate r obtained by base station j ijk (t, τ) is specifically expressed as:
r ijk (t,τ)＝w i c jk (t,τ)
wherein w is i It is the wireless bandwidth allocated by the system in the base station to provide the user with the slice i service, the value of which is constant throughout the time, the size of which is determined by the traffic properties of the slice itself.
Further, defining a minimum transmission rate for providing a single slice i service asThe transmission rate QoS constraint is as follows: on a long time scale, the transmission rate satisfies only: />Base station j may accommodate user k's slice i request。
It is worth noting that the sliced SLA contains multiple performance indicators, and the present application only considers the transmission speed QoS constraints.
Variable a in response to determining whether base station j can accommodate the request for slice i for user k ijk (t) ∈ {0,1} is to indicate that base station j admits user k at time t and serves it with slice i. Obviously, only when slice i is deployed on base station J, i.e., J e J i When a is ijk (t) may be equal to 1. Then the set of users for which base station j admits slice i can be denoted as K ij (t)＝{k∈K|a ijk (t)＝1}，j∈J i I is E I, the number is K ij (t)。
Still, if slice i is not deployed on base station j, base station j may not accommodate all user k's requests for slice i, i.e., a ijk (t)＝0,Then i, j, k are described as unacceptable combinations.
Based on the above, an unacceptable slice i and a request for slice i are determined.
Wherein the combination is not acceptable, i.e. a ijk (t)＝0,
Step S3303: based on the determined non-admissible combination, an optimal base station level admission control decision is obtained.
Specifically, iterative solution is performed based on a greedy strategy, each iteration being performed to minimizeFor the purpose, the base station j is optimized one by one to accommodate the variable a of the user k at time t and to provide it with the services of the slice i ijk And (t) until h (t) converges, thereby obtaining an optimal base station level admission control decision.
Further, for { a }, each iteration cycle ijk (t) } analyzing one by one and judging in two cases:
(1) if a in the last iteration period ijk (t) =0, and i, j, k are not an unacceptable combination, and after admission, the transmission rate QoS condition can be met and h (t) is reduced, then the iteration period a ijk (t) =1. Otherwise, maintain a ijk (t) =0.
(2) If a in the last iteration period ijk (t) =1, h (t) can be reduced after rejection, the iteration period a ijk (t) =0. Otherwise, maintain a ijk (t) =1.
After all { a } ijk After (t) }, it is determined whether h (t) of the two iteration cycles converges. If not, continuing iteration; if the convergence is carried out, stopping iteration, and obtaining a base station level admission control optimization variable { a }, by the NSACF controller ijk (t), K epsilon K, I epsilon I, J epsilon J) the best solution result, which is the best base station level admission control decision.
Step S3304: outputting the optimal base station level admission control decision.
Step S340: in response to obtaining the best base station level admission control decision, the service request queue is updated.
Service request queue { Q ik The stability of (t), k.epsilon.K, i.epsilon.I, directly affects the stability of the system.
Specifically, the nsaf controller centrally manages different slice requests of each user, each slice request of each user is independently used as a queue, and the processing of each queue does not affect other queues. Q (Q) ik The dynamic update of (t) is specifically expressed as:
Q ik (t+1)＝max[Q ik (t)-ε ik (t),0]+n ik (t)
wherein Q is ik (t) is a service request queue for user k with respect to slice i at time t, n ik (t) is expressed as the number of service requests by the system for slice i for user k at time t,it is the service speed, a, provided by the system for slice i of user k at time t ijk (t) is a signal indicating that the base station j is connected at time tReceive user k and provide it with slice i service,/-, for>Indicating that base station j can be K e K for each user ij Slice i of (t) requests the speed of service provided. Wherein { Q ik (0) K e K, I e I } = 0. Obviously, the total number of queues is I K. Each queue should have some stability and should not be an infinitely growing queue. When->Q ik (t) is stable.
Step S350: in response to making an update to the service request queue, a time-averaged profit is obtained.
Specifically, time average profitThe concrete steps are as follows:
where P (t) represents the total profit of the system at time t, t represents the time, and λ represents each time.
Step S360: in response to obtaining the time average profit, it is determined whether the maximum time is reached.
Specifically, it is determined whether the time t set in step S310 reaches a preset maximum time, and if the maximum time is not reached, the steps S320-350 are re-executed at time t+1 until the maximum time is reached. If the set time t reaches the preset maximum time, step S370 is executed.
Step S370: outputting time average profit, optimal slice level admission control decision and optimal base station level admission control decision.
The application has the following beneficial effects:
(1) The method of the application can make the NSACF controller of the RAN slice correctly control the admission quantity of the service requests of different slices of users in a dynamic network, and reasonably arrange different base stations in the system to admit the requests and provide services for the requests on the premise of ensuring the stability of the RAN slice admission control system and the SLA of each slice, and simultaneously make the RAN slice system obtain better time average profit.
(2) The method can effectively adapt to the application scene of the differential deployment of the network slice, and can adapt to the situation that the base station covered by the slice is limited compared with the prior admission control scheme. In addition, the method is suitable for the dynamic heterogeneous network, has better expandability, and can be more and more obvious along with the increase of the scale of the RAN slice compared with other methods.
Although the examples referred to in the present application are described for illustrative purposes only and not to be limiting of the application, modifications, additions and/or deletions to the embodiments may be made without departing from the scope of the application.
The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (4)
1. The admission control method of the wireless access network facing the network slice is characterized by comprising the following steps:
s1: initializing the time;
s2: responding to the initialization of the completion time to acquire an optimal slice level admission control decision;
s3: in response to obtaining the optimal slice level admission control decision, obtaining an optimal base station level admission control decision;
s4: updating a service request queue in response to obtaining an optimal base station level admission control decision;
s5: in response to performing the update of the service request queue, obtaining a time average profit;
s6: judging whether the maximum moment is reached or not in response to the time average profit acquisition;
if the maximum time is not reached, adding 1 to the time, and re-executing the steps S2-S5;
s7: outputting a time average profit, an optimal slice level admission control decision and an optimal base station level admission control decision if the maximum moment is reached;
wherein in response to the initialization of the completion time instant, an optimal slice level admission control decision is obtained, comprising the sub-steps of:
taking the initialized moment, the service request quantity and the service request queue information as input;
acquiring an optimal slice level admission control decision according to the input;
wherein n is set before the optimal slice level admission control decision is obtained ik (t) expressed as the number of service requests of user k for slice i that are admitted by the system at time t, assuming that all of the slice service requests admitted by the system have base station admission, i.e., base station exists, that satisfies its transmission rate constraintCan provide services guaranteeing slice service level agreements, so that n is more than or equal to 0 ik (t)≤N ik (t), wherein N ik (t) is expressed as the number of service requests of user k with respect to slice i at time t;
in the process of obtaining the optimal slice level admission control decision, firstly, a Lyapunov function L (t) related to a service request queue is defined, then, a Lyapunov drift delta (t) is further defined, a drift minus profit expression at each t moment is determined according to the Lyapunov drift delta (t), the upper limit value of the drift minus profit expression is determined, and the number n of service requests of a user k related to a slice i at the t moment in the upper limit value are admitted by a system ik (t) solving to obtain an optimal slice level admission control decision;
in the process of obtaining the optimal slice level admission control decision, the upper limit f (t) of the drift minus profit expression at each time t is specifically expressed as:
wherein B is defined as For a collection of users,for user k, upper bound on the number of service requests of slice I, J denotes the number of base stations, I denotes the number of network slices,/->Indicating that base station j can be +/for each user>Is requesting the speed of service provided by slice i, +.>Representing the service speed +.>Upper limit of (2), where n ik (t) is expressed as the number of service requests received by the system by user k with respect to slice i at time t, E { } is expressed as the average of the bracketed portions at different times, Q ik (t) is the service request queue for user k with respect to slice i at time t, Q (t) represents the vector of all service request queues, V is a trade-off factor, gamma i Request data volume, d, for each service of slice i ij (t) the cost of providing slice i service at base station j at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station j admits user k at time t and provides slice i service for it;
d ij the size of (t) is divided into three cases: if slice i is not deployed at base station j, i.eThen at any time there is zero cost, i.e. d ij (t) =0; if slice i is deployed on base station j, i.e. +.>But at time t base station j does not admit any user's slice i request, i.e. +.>Then a fixed cost d is charged ij (t)＝d 0 The method comprises the steps of carrying out a first treatment on the surface of the If slice i is deployed on base station j, i.e. +.>And at time t base station j admits the user set of slice i not empty, i.e +.>The system is +.>Collecting cost; thus d ij (t) is expressed as
Wherein beta is i Is a constant representing the cost of the system providing a unit slice i service, a ijk (t) ∈ {0,1} is a representation that base station j admits user k at time t and serves it with slice i,indicating that base station j can be +/for each user>The service speed requested to be provided by the slice i;
in the process of obtaining the optimal slice level admission control decision, for non-constant partsConverting the average value of the minimized time to the value of the minimized time, which is expressed as:
thereby directly obtaining the slice level admission control optimization variableThe solution result of (2) is
In the process of obtaining the optimal base station level admission control decision, for non-constant partConverting the average value of the minimized time to the value of the minimized time, which is expressed as:
where V is a trade-off factor, d ij (t) is the cost of providing slice i service at base station j at time t,indicating that base station j can be +/for each user>The service speed, Q, provided by the slice i request ik (t) user k is a service request queue for slice i at time t, variable a ijk (t)E {0,1} is e { indicates that base station j admits user k at time t and provides it with slice i service, }, a }>Representing the set of base stations, +.>Representing a set of network slices;
and then, carrying out iterative solution by using a solution method based on a greedy strategy, thereby obtaining an optimal base station level admission control decision.
2. The admission control method for a network slice oriented radio access network of claim 1 wherein the service request queue Q ik The updating of (t) is specifically expressed as:
Q ik (t+1)＝max[Q ik (t)-ε ik (t),0]+n ik (t)
wherein Q is it (t+1) is an updated service request queue, n ik (t) is expressed as the number of service requests by user k with respect to slice i that are admitted by the system at time t ε ik And (t) represents the service speed provided by the system for slice i of user k at time t.
3. The admission control method for a network slice oriented radio access network of claim 1 wherein time-averaged profitThe concrete steps are as follows:
where P (t) represents the total profit of the system at time t, t represents the time, and λ represents each time.
4. The admission control system of the wireless access network facing the network slice is characterized by comprising an initialization unit, an optimal slice level admission control decision acquisition unit, an optimal base station level admission control decision acquisition unit, a service queue updating unit, a time average profit acquisition unit, a judgment unit and an output unit;
an initialization unit for initializing the time;
an optimal slice level admission control decision acquisition unit configured to acquire an optimal slice level admission control decision;
an optimal base station level admission control decision acquisition unit configured to acquire an optimal base station level admission control decision;
a service queue updating unit for updating the service request queue;
the time average profit obtaining unit is used for obtaining time average profit;
the judging unit is used for judging whether the maximum moment is reached, if the maximum moment is not reached, adding 1 to the moment, re-acquiring the optimal slice level admission control decision, acquiring the optimal base station level admission control decision, updating the service request queue and acquiring the time average profit;
the output unit is used for outputting time average profit, optimal slice level admission control decision and optimal base station level admission control decision if the maximum moment is reached;
in an optimal slice level admission control decision acquisition unit, an optimal slice level admission control decision is acquired, comprising the sub-steps of:
taking the initialized moment, the service request quantity and the service request queue information as input;
acquiring an optimal slice level admission control decision according to the input;
wherein n is set before the optimal slice level admission control decision is obtained ik (t) expressed as the number of service requests of user k for slice i that are admitted by the system at time t, assuming that all of the slice service requests admitted by the system have base station admission, i.e., base station exists, that satisfies its transmission rate constraintCan provide guaranteed slice service level coordinationThe service of the proposal, therefore 0.ltoreq.n ik (t)≤N ik (t), wherein N ik (t) is expressed as the number of service requests of user k with respect to slice i at time t;
in the process of obtaining the optimal slice level admission control decision, firstly, a Lyapunov function L (t) related to a service request queue is defined, then, a Lyapunov drift delta (t) is further defined, a drift minus profit expression at each t moment is determined according to the Lyapunov drift delta (t), the upper limit value of the drift minus profit expression is determined, and the number n of service requests of a user k related to a slice i at the t moment in the upper limit value are admitted by a system ik (t) solving to obtain an optimal slice level admission control decision;
in the process of obtaining the optimal slice level admission control decision, the upper limit f (t) of the drift minus profit expression at each time t is specifically expressed as:
wherein B is defined as For a collection of users,for user k, upper bound on the number of service requests of slice I, J denotes the number of base stations, I denotes the number of network slices,/->Indicating that base station j can be +/for each user>Is requesting the speed of service provided by slice i, +.>Representing the service speed +.>Upper limit of (2), where n ik (t) is expressed as the number of service requests received by the system by user k with respect to slice i at time t, E { } is expressed as the average of the bracketed portions at different times, Q ik (t) is the service request queue for user k with respect to slice i at time t, Q (t) represents the vector of all service request queues, V is a trade-off factor, gamma i Request data volume, d, for each service of slice i ij (t) the cost of providing slice i service at base station j at time t, variable a ijk (t) ∈ {0,1} is to indicate that base station j admits user k at time t and provides slice i service for it;
d ij the size of (t) is divided into three cases: if slice i is not deployed at base station j, i.eThen at any time there is zero cost, i.e. d ij (t) =0; if slice i is deployed on base station j, i.e. +.>But at time t base station j does not admit any user's slice i request, i.e. +.>Then a fixed cost d is charged ij (t)＝d 0 The method comprises the steps of carrying out a first treatment on the surface of the If slice i is deployed on base station j, i.e. +.>And at time t base station j admits the user set of slice i not empty, i.e +.>The system is +.>Collecting cost; thus d ij (t) is expressed as
Wherein beta is i Is a constant representing the cost of the system providing a unit slice i service, a ijk (t) ∈ {0,1} is a representation that base station j admits user k at time t and serves it with slice i,indicating that base station j can be +/for each user>The service speed requested to be provided by the slice i;
in the process of obtaining the optimal slice level admission control decision, for non-constant partsConverting the average value of the minimized time to the value of the minimized time, which is expressed as:
thereby directly obtaining the slice level admission control optimization variableThe solution result of (2) isObtaining an optimal slice level admission control decision;
in the process of obtaining the optimal base station level admission control decision by the optimal base station level admission control decision obtaining unit, for the non-constant partConverting the average value of the minimized time to the value of the minimized time, which is expressed as:
where V is a trade-off factor, d ij (t) is the cost of providing slice i service at base station j at time t,indicating that base station j can be +/for each user>The service speed, Q, provided by the slice i request ik (t) user k is a service request queue for slice i at time t, variable a ijk (t) ∈ {0,1} is a representation that base station j admits user k at time t and provides it with slice i service, +.>Representing the set of base stations, +.>Representing a set of network slices;
and then, carrying out iterative solution by using a solution method based on a greedy strategy, thereby obtaining an optimal base station level admission control decision.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210356501.8A CN114726729B (en) | 2022-04-06 | 2022-04-06 | Admission control method and system of wireless access network facing network slicing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210356501.8A CN114726729B (en) | 2022-04-06 | 2022-04-06 | Admission control method and system of wireless access network facing network slicing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114726729A CN114726729A (en) | 2022-07-08 |
| CN114726729B true CN114726729B (en) | 2023-08-29 |
Family
ID=82242272
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210356501.8A Active CN114726729B (en) | 2022-04-06 | 2022-04-06 | Admission control method and system of wireless access network facing network slicing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114726729B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018131453A1 (en) * | 2017-01-10 | 2018-07-19 | 株式会社Ｎｔｔドコモ | Ran connection control method and base station |
| CN109842910A (en) * | 2017-11-28 | 2019-06-04 | 中国移动通信有限公司研究院 | A kind of processing method and access network elements of network slice |
| EP3806402A1 (en) * | 2019-10-10 | 2021-04-14 | Commissariat à l'énergie atomique et aux énergies alternatives | Admission control process for slices in a virtualised telecommunications network and the congestion liable to be generated between the services deployed on said slices |
| CN113411862A (en) * | 2021-06-18 | 2021-09-17 | 北京邮电大学 | Cache placement and user access method and device in dynamic cellular network |
| CN114070750A (en) * | 2021-11-03 | 2022-02-18 | 南京理工大学 | 5G network slice resource allocation method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7305431B2 (en) * | 2002-09-30 | 2007-12-04 | International Business Machines Corporation | Automatic enforcement of service-level agreements for providing services over a network |
| US20220022044A1 (en) * | 2020-07-15 | 2022-01-20 | Northeastern University | Methods for Multi-Access Edge Computing Network Slicing in 5G Networks |
-
2022
- 2022-04-06 CN CN202210356501.8A patent/CN114726729B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018131453A1 (en) * | 2017-01-10 | 2018-07-19 | 株式会社Ｎｔｔドコモ | Ran connection control method and base station |
| CN109842910A (en) * | 2017-11-28 | 2019-06-04 | 中国移动通信有限公司研究院 | A kind of processing method and access network elements of network slice |
| EP3806402A1 (en) * | 2019-10-10 | 2021-04-14 | Commissariat à l'énergie atomique et aux énergies alternatives | Admission control process for slices in a virtualised telecommunications network and the congestion liable to be generated between the services deployed on said slices |
| CN113411862A (en) * | 2021-06-18 | 2021-09-17 | 北京邮电大学 | Cache placement and user access method and device in dynamic cellular network |
| CN114070750A (en) * | 2021-11-03 | 2022-02-18 | 南京理工大学 | 5G network slice resource allocation method |
Non-Patent Citations (1)
| Title |
|---|
| 移动网络中基于人工智能的端到端网络切片技术研究;李静等;《信息通信技术与政策》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114726729A (en) | 2022-07-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11184289B2 (en) | Systems and methods for managing network traffic with a network operator | |
| EP2304899B1 (en) | Priority-based admission control in a network with variable channel data rates | |
| WO2014101243A1 (en) | Load balancing method and network control node | |
| CN111314997B (en) | Service type-based access network slice selection method and system | |
| TW200818945A (en) | Load balancing apparatus and method in wireless network hotspots | |
| CN106792835B (en) | Control method and system for wireless equipment sharing bandwidth | |
| EP3725133B1 (en) | Method and apparatus for front-haul network dependent wireless network control | |
| CN110267343B (en) | Paging management method, network data analysis function and network side equipment | |
| CN111556508B (en) | Stackelberg game multi-operator dynamic spectrum sharing method facing large-scale IoT access | |
| WO2020212640A1 (en) | Resource allocation in a network slice | |
| WO2011066807A1 (en) | Method, device and system for resource allocation | |
| Banchs et al. | A novel radio multiservice adaptive network architecture for 5G networks | |
| Khorov et al. | SEBRA: SAND-enabled bitrate and resource allocation algorithm for network-assisted video streaming | |
| CN102111890A (en) | Priority regulation method and relative equipment | |
| KR20180008701A (en) | Resource allocation method, apparatus, system, and base station | |
| CN103686826A (en) | Admission control method of carrier aggregation system | |
| WO2015035573A1 (en) | Network transmission method and apparatus | |
| CN114726729B (en) | Admission control method and system of wireless access network facing network slicing | |
| WO2011087935A2 (en) | Method of controlling resource usage in communication systems | |
| Zhang et al. | A hard and soft hybrid slicing framework for service level agreement guarantee via deep reinforcement learning | |
| WO2020136512A1 (en) | Prioritizing users based on revenue during congestion | |
| CN107426809B (en) | WVN power and cache allocation method based on virtual user queue model | |
| Khatibi et al. | Elastic slice-aware radio resource management with AI-traffic prediction | |
| CN109803352B (en) | Resource allocation method and device for fog wireless access network | |
| CN102665217B (en) | Wireless communication channel resource optimization method based on wireless multimedia network |
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 |