CN114466356A - Task unloading edge server selection method based on digital twin - Google Patents

Task unloading edge server selection method based on digital twin Download PDF

Info

Publication number
CN114466356A
CN114466356A CN202210111471.4A CN202210111471A CN114466356A CN 114466356 A CN114466356 A CN 114466356A CN 202210111471 A CN202210111471 A CN 202210111471A CN 114466356 A CN114466356 A CN 114466356A
Authority
CN
China
Prior art keywords
server
user
task
unloading
information
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.)
Granted
Application number
CN202210111471.4A
Other languages
Chinese (zh)
Other versions
CN114466356B (en
Inventor
曾捷
宋雨欣
武腾
陈昌川
牟郸霖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202210111471.4A priority Critical patent/CN114466356B/en
Publication of CN114466356A publication Critical patent/CN114466356A/en
Application granted granted Critical
Publication of CN114466356B publication Critical patent/CN114466356B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/06—Authentication
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/08—Testing, supervising or monitoring using real traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/09—Management thereof

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Security & Cryptography (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Computer And Data Communications (AREA)

Abstract

The invention belongs to the technical field of wireless communication, and particularly relates to a task unloading edge server selection method based on digital twinning, which comprises the following steps: constructing a digital twin safety authentication system and initializing the system; a user sends a task unloading request to a server; the digital twin layer screens the servers according to the user task unloading request information to obtain the servers meeting the unloading request; unloading the tasks of the users to the screened servers; the invention fuses the digital twin technology with the mobile edge network, and the digital twin fully utilizes a large amount of collected storage data by continuously monitoring the running state of the mobile communication network, thereby relieving the problems of limited storage space and calculation training capacity of the user terminal and the like and assisting the user terminal to better complete the selection of the mobile edge server.

Description

Task unloading edge server selection method based on digital twin
Technical Field
The invention belongs to the technical field of wireless communication, and particularly relates to a task unloading edge server selection method based on digital twins.
Background
In order to meet the increasing mobile data traffic demand, Mobile Edge Computing (MEC) is used as a key technology and key driver of the internet of things, computing and storage resources are provided at the edge, so that the edge of a wireless access network near an end user has IT (information technology) service and cloud computing capabilities, and the aims of providing a low-delay and high-bandwidth environment and accessing a wireless network in real time are fulfilled. The MEC has the potential to enable a wide range of applications and services, particularly those that are time sensitive, due to its superior characteristics of proximity/location/mobility awareness, real-time insight into wireless network information, high bandwidth, low latency, etc. In addition, the MEC may also collect more important information about user interests, location, preferences and behavior, thereby improving quality of service.
In the 6G era, with the explosive growth of application data volume, a single Mobile Edge Server (MES) cannot process multiple tasks simultaneously, and edge collaboration can effectively make up for the shortages of communication, calculation and storage resources of the single MES. The edge cooperation processes tasks in a distributed cooperative processing mode, idle network resources can be effectively utilized, and system power consumption and time overhead are saved. At present, many researches are combined with Artificial Intelligence (AI) to solve the problems of resource allocation, task unloading, caching and the like in an MEC scene. With the increasing number and types of the cooperative nodes, the safety of the cooperative nodes is not negligible, meanwhile, a large amount of data is trained, and the system needs to ensure that AI can be fully utilized to realize edge cooperation under the limitation of terminal resources.
The digital twin technology realizes the interactive fusion of a physical world and an information world, and draws a real communication world and a virtual space by fully utilizing data of a physical entity, updating and transmission of a sensor, operation history and the like so as to reflect the operation condition of the physical entity. The digital twin model is dynamic and can be combed based on real-time uploaded sampling data to complete physical modeling, control and prediction. Therefore, the establishment of the digital twin edge network can train and predict by using a large amount of collected data in the digital twin, thereby improving the overall performance of the network, relieving the resource limitation problem of the user terminal, ensuring the safety and the communication quality of the cooperative node, and assisting the user terminal to complete task unloading.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a task unloading edge server selection method based on digital twinning, which comprises the following steps: constructing a system model and a digital twin safety authentication system, and initializing the digital twin safety authentication system; a user sends a task unloading request to a server; the digital twin layer screens the servers according to the user task unloading request information to obtain the servers meeting the unloading request; and unloading the tasks of the users to the screened servers.
Preferably, the process of initializing the digital twin security authentication system includes: setting system parameter groups
Figure BDA0003495181460000021
And key group
Figure BDA0003495181460000022
In the system parameter group
Figure BDA0003495181460000023
Generating parameter S at random, selecting positive integer
Figure BDA0003495181460000024
As a primary cryptographic parameter; calculating a system public key K according to the parameter S and the positive integer wpubwS; selecting two hash functions
Figure BDA0003495181460000025
And
Figure BDA0003495181460000026
the privacy public parameter of the system is
Figure BDA0003495181460000027
Wherein
Figure BDA0003495181460000028
A group of system parameters is represented by a group of system parameters,
Figure BDA0003495181460000029
representing a key group, S representing a randomly generated parameter, KpubRepresenting the system public key, H1Representing a first hash function, H2Representing a second hash function.
Preferably, the process of screening the authenticated server by the digital twin layer according to the user task offloading request information includes:
s1: calculating a first classification information gain and a second classification information gain;
s2: comparing the first classification information gain with the second classification information gain, and if the first classification information gain is larger than the second classification information gain, executing a server screening strategy for verifying the channel gain after the security of the server is verified; if the gain of the first classification information is less than or equal to the gain of the second classification information, executing a server screening strategy for verifying the security of the server after verifying the gain of the channel;
s3: collecting all servers meeting the unloading requirements, and acquiring the number of services meeting the requirements; and judging whether the number of the services meeting the requirement meets the subtask distribution number, if so, ending, and if not, returning to the step S1.
Further, the process of calculating the first classification information gain and the second classification information gain includes:
the first classification information gain calculation process is as follows:
step 1: obtaining probability p of server being selectedsel；
Step 2: according to pselCalculating the Gini value (F) of the server;
and step 3: calculating a Gini _ index (F, sg) of the server according to the Gini value;
and 4, step 4: calculating a first classification information gain G based on the Gini value Gini (F) and the Gini coefficient Gini _ index (F, sg)sg(F)；
The second classification information gain calculation process includes:
step 1: acquiring a data set of a user side; sorting the data in the data set in ascending order, and using the intermediate value of two adjacent values of the sample to make the intermediate valueIs a division point xii(1≤i＜NF-1)，NFIs the number of sample values;
step 2: dividing the samples in the data set into two groups according to the classification mode of being larger than the division point and being smaller than the division point, and calculating the two classification kini coefficients Gini _ index (F, xi) of each division pointi)；
And 3, step 3: screening out the division point with the best classification effect according to the two classification damping coefficients
Figure BDA0003495181460000031
And 4, step 4: calculating a second classification information gain G according to the segmentation point with the best classification effect and the classification damping coefficientgs(F)。
Further, the server screening policy for performing channel gain verification after the server security verification is performed includes:
step 1: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication succeeds, executing the step 2;
step 2: acquiring task information of a user side, and dividing a task into a local task and an unloading task according to the task information;
and step 3: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and 4, step 4: setting a channel gain threshold according to the unloading task information; comparing the channel gain of the user side and the channel gain of the service side with a set channel gain threshold, and if the channel gain of the user side and the channel gain of the service side are more than or equal to the set channel gain threshold, the server meets the unloading requirement; otherwise the server does not meet the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
Further, the process of authenticating the user side and the edge server by adopting the digital twin security authentication system comprises the following steps:
step 11: inputting the user terminal and the server into a digital twin safety authentication system; the digital twin safety authentication system respectively distributes an identity identification code and a random identity identification code for the user terminal and the server;
step 12: the user terminal informs the digital twin layer DT of the unloading task and requests the DT layer to select a proper server MES;
step 13: user side using unloading request information Mm＝(RIDuser||IDm||Ts) And private key wH1(IDm) Calculating the signature ζm＝H2(M)·wH1(IDm) (ii) a Where, | | represents information data connection, TsRepresenting a time stamp, RIDuserIndicating the identity, ID, of the user terminalmRandom ID code representing server, w represents random positive integer, H1Representing a first hash function, H2Representing a second hash function, M representing offload request information;
step 14: the user side will { Mm,ζmSending the key to a server mthMES, and calculating a symmetric key K of a user side and the serveruser-mth＝e(wH1(RIDuser),H1(IDm) ); wherein e is
Figure BDA0003495181460000041
And
Figure BDA0003495181460000042
linear mapping of (2);
step 15: server receives { Mm,ζmAfter that, the timestamp T of the uninstall request message is checkedsIf not, the authentication fails, if so, the symmetric key K of the server is calculated according to the public parameters of the systemmth-user＝e(H1(RIDuser),wH1(IDm))；
Step 16: matching the symmetric key of the server with the symmetric key of the user side, failing authentication if the symmetric keys are not matched, and generating a check code C by the server if the symmetric keys are matchedm＝H2(Kmth-user||RIDuser||IDm) And check the message { RIDuser,IDm,CmSending the data to the user side;
and step 17: after receiving the check message, the user end generates a verification code Cuser＝H2(Kuser-mth||RIDuser||IDm) Verification of CmAnd CuserIf the server passes the verification, the server is a safe connection server, otherwise the server is not connectable.
Further, the process of setting the channel gain threshold according to the offloading task information includes:
step 1: obtaining BHz bandwidth and T time delay occupied by user segment of system modelDAccording to BHz and time delay TDCalculating the number of channels available to the user, i.e. N ═ BTD；
Step 2: acquiring a channel state information matrix g of a user side and a servermAnd calculating post-processing signal-to-noise ratio gamma of the user side and the server according to the channel state information matrix; the post-processing signal-to-noise ratio is formulated as:
Figure BDA0003495181460000051
wherein p isuIndicating the user side transmission power, ENmIs the channel estimation error noise, ZNmIs additive white Gaussian noise, | | gm||2Is the channel gain;
and step 3: calculating shannon capacity C (gamma) of post-processing signal-to-noise ratio;
C(γ)＝log2(1+γ)
and 4, step 4: correcting the post-processing signal-to-noise ratio to obtain a corrected post-processing signal-to-noise ratio; the correction formula is as follows:
Figure BDA0003495181460000052
and 5: calculating the error probability of the server by adopting a Q function Q (w) according to the shannon capacity C (gamma) and the corrected post-processing signal-to-noise ratio V (gamma);
Figure BDA0003495181460000053
wherein n represents the number of available channels for pilot; d represents the size of the task data;
and 6: calculating the data transmission rate of the user side according to the error probability;
and 7: according to the requirements of the ultra-reliable low-delay communication scene on time delay and error probability, the bit number of short packet transmission and the total transmission bandwidth, the threshold value | g of channel gain is solvedth||2。
Further, the formula for calculating the data transmission rate of the user side is as follows:
Figure BDA0003495181460000061
where ε is the error probability; n is the number of available channels for pilot; gamma is the post-processing signal-to-noise ratio.
Further, the requirements of the ultra-reliable low-delay communication scene on the time delay and the error probability include that the time delay is less than 0.5 millisecond, and the error probability is less than 10-5。
Preferably, the process of executing the server screening policy for verifying the security of the server after verifying the channel gain includes:
step 1: acquiring task information of a user side, and dividing a task into a local task and an unloading task according to the task information;
step 2: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and step 3: calculating channel gain according to the unloading task information; setting a channel gain threshold; comparing the calculated channel gain with a set channel gain threshold, and if the calculated channel gain is greater than or equal to the set channel gain threshold, executing the step 4; otherwise the server does not meet the unloading requirement;
and 4, step 4: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication succeeds, the server meets the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
The invention has the beneficial effects that:
1) the digital twin technology is fused with the mobile edge network, the digital twin fully utilizes a large amount of collected storage data by continuously monitoring the running state of the mobile communication network, the problems of limited storage space and calculation training capacity of a user terminal and the like are solved, and the user terminal is assisted to better complete the selection of the mobile edge server;
2) in the traditional safety verification, a user terminal must inquire and acquire data of each MES, and the digital twin technology can reduce the information exchange data volume of the user terminal and complete the safety confirmation of the MES with less extra resource consumption;
3) the selection of the cooperative mobile edge server is completed through the real-time data collected by the DT, the data security is fully guaranteed, meanwhile, a high-quality communication link is provided to realize the selection of the mobile edge server with high information gain of the user terminal, and the overall performance of the task unloading process of the digital twin edge network is improved.
Drawings
FIG. 1 is a schematic diagram of a network architecture of the method of the present invention;
FIG. 2 is a schematic relational flow diagram of the method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A method for task offload edge server selection based on digital twinning, as shown in fig. 1, the method comprising: constructing a digital twin safety authentication system and initializing the system; a user provides a task uninstalling request to a server; the digital twin layer screens the servers according to the user task unloading request information to obtain the servers meeting the unloading request; and unloading the tasks of the users to the screened servers.
The constructed system model comprises: in the process of the task unloading edge server selection method based on the digital twin, the number of the mobile edge servers in the network is set as MtotalThe number of edge servers selected as cooperating mobile edge servers is Msel. The User sends a short data packet with the size of eta bits through N available channel numbers, and occupies BHz bandwidth and T time delayDThe relation between the total available channel number and the bandwidth and the time delay is BTD. The number of antennas of each MBS is L, and the signals received by the server mthMES are:
Figure BDA0003495181460000071
wherein
Figure BDA0003495181460000072
And with
Figure BDA0003495181460000073
Respectively a pilot signal and a data signal received by the mthMES,
Figure BDA0003495181460000074
is the average transmit power, g, of the UsermIs the channel state information matrix of User and mthMES,
Figure BDA0003495181460000075
and
Figure BDA0003495181460000076
is a transmission signal of the terminal device,
Figure BDA0003495181460000077
representing noise and other interference.
The task to be processed by the User can be expressed as
Figure BDA0003495181460000078
D is the size of the task data, lambda is the period of the CPU processing task, TTHIs the maximum acceptable delay. To reduce latency and improve energy efficiency, the User leaves the task portion for local processing and offloads the remaining tasks to the MESs for processing. The local processing task is expressed as
Figure BDA0003495181460000081
Wherein Dlocalα D (0 ≦ α ≦ 1) is the local processing task data size, λlocalIs the period of CPU processing of the local task, TlocalThe maximum acceptable delay for the local task.
Assume that the CPU cycle frequency estimate for a User processing a local task in a Digital Twin (DT) is
Figure BDA0003495181460000082
The error value between the CPU and the DT of the CPU cycle frequency is Δ flocalThen the estimated time required for the User to process the local task in DT is
Figure BDA0003495181460000083
It is assumed that the deviation Δ f between User and his DT can be obtained beforehandlocalThe actual calculated delay can be expressed as
Figure BDA0003495181460000084
Wherein the calculated time delay between the estimated and actual DT values
Figure BDA0003495181460000085
User assigns tasks on MESs, which will be divided into MselSubtasks, each of which can be represented as
Figure BDA0003495181460000086
Wherein
Figure BDA0003495181460000087
Processing the size of task data for the mthMBS; beta is a betamRatio of allocated offload tasks to overall offload tasks for mthMBS, λmIs the period of the mthMBS processing task CPU processing, TmMaximum acceptable time delay for mthMBS processing task
The time delay overhead of task offloading is mainly task uploading, task downloading and MES task processing. In reality, the size of the task download is far smaller than that of the task upload, so the time of the task download can be ignored. The time required for the task information to be transmitted from the User to the mthMES is
Figure BDA0003495181460000088
Suppose that mthmES in DT is used to process task WmHas a CPU cycle frequency of
Figure BDA0003495181460000089
In reality, the CPU cycle frequency error value of the mthMES in the mthmES and the DT is delta fmAssuming this value is a known quantity in advance, mthmES processes task W in DTmAn estimate of the time required is
Figure BDA00034951814600000810
The actual calculated delay can be expressed as
Figure BDA00034951814600000811
Wherein the calculated time delay between the estimated and actual values of DT
Figure BDA00034951814600000812
The total time for the mthMES to process the offload tasks is
Figure BDA00034951814600000813
The total time for all selected MESs to process the offload tasks is
Figure BDA0003495181460000091
Total time t for system to process task Wtotal＝max{tlocal,toff}。
Dynamic power of the dominant component in CPU power, since in CMOS circuits, the power is the same each time an operation is performed
Figure BDA0003495181460000092
Proportional, clock frequency f in low voltage operating statecAnd a supply voltage VcirApproximately linearly proportional, so the local task computation power consumption of User can be expressed as
Figure BDA0003495181460000093
The computational power consumption of the mthMES for processing the offload task is expressed as
Figure BDA0003495181460000094
κlocalAnd kappamRespectively, the energy coefficients of User and mthMES, which are affected by the chip structure. Total power consumption of User is Puser＝Mselpu+PlocalThe total power consumption of the selected MESs is
Figure BDA0003495181460000095
Total power consumption P of systemtotal＝Puser+Pmes。
In the digital twin safety authentication system, firstly, the system is initialized and a system parameter group is set
Figure BDA0003495181460000096
And key group
Figure BDA0003495181460000097
Make S random by
Figure BDA0003495181460000098
Generating and authenticating positive integer selected by system
Figure BDA0003495181460000099
Computing the public key K as the main cryptographic parameterpubTwo hash functions are selected, set
Figure BDA00034951814600000910
Setting up a mapping disclosure common parameter set
Figure BDA00034951814600000911
Identification code ID of mthMES generated by authentication systemmThe public key is H1(IDm) Corresponding private key is wH1(IDm) And key information { ID }m,H1(IDm),wH1(IDm) It is sent to mthmES. When User accesses the authentication system, it will be assigned a random ID RIDuserThe public key is H1(RIDuser) Corresponding private key is wH1(RIDuser) The system will integrate the key information
{RIDuser,H1(RIDuser),wH1(RIDuser) And sending the data to a corresponding User, wherein the User can change the random identity so as to ensure the identity and position information security in the task unloading process. The ID codes and random ID codes distributed to the User and the MES by the authentication system have valid periods and can be updated in real time.
As shown in fig. 2, the process of the digital twin layer screening the authenticated server according to the user task offload request information includes:
s1: calculating a first classification information gain and a second classification information gain;
s2: comparing the first classification information gain with the second classification information gain, and if the first classification information gain is larger than the second classification information gain, executing a server screening strategy for verifying the channel gain after the security of the server is verified; if the gain of the first classification information is less than or equal to the gain of the second classification information, executing a server screening strategy for verifying the security of the server after verifying the gain of the channel;
s3: collecting all servers meeting the unloading requirements, and acquiring the number of services meeting the requirements; and judging whether the number of services meeting the requirement meets the subtask distribution number, if so, ending, and if not, returning to the step S1.
The server screening strategy for performing channel gain verification after server security verification comprises:
step 1: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication succeeds, executing the step 2;
step 2: acquiring task information of a user side, and dividing a task into a local task and an unloading task according to the task information;
and step 3: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and 4, step 4: setting a channel gain threshold according to the unloading task information; comparing the channel gain of the user side and the channel gain of the service side with a set channel gain threshold, and if the channel gain of the user side and the channel gain of the service side are more than or equal to the set channel gain threshold, the server meets the unloading requirement; otherwise the server does not meet the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
The process of authenticating the user terminal and the edge server by adopting the digital twin safety authentication system comprises the following steps:
step 11: inputting the user terminal and the server into a digital twin safety authentication system; the digital twin security authentication system respectively distributes an identity identification code and a random identity identification code for the user terminal and the server;
step 12: the user terminal informs the digital twin layer DT of the unloading task and requests the DT layer to select a proper server MES;
step 13: user side using unloading request information Mm＝(RIDuser||IDm||Ts) And private key wH1(IDm) Calculating the signature ζm＝H2(M)·wH1(IDm) (ii) a Wherein, | | representsInformation data connection, TsRepresenting a time stamp, RIDuserIndicating the identity, ID, of the user terminalmRandom ID code representing server, w represents random positive integer, H1Representing a first hash function, H2Representing a second hash function, M representing offload request information;
step 14: the user side will { Mm,ζmSending the key to a server mthMES, and calculating a symmetric key K of a user side and the serveruser-mth＝e(wH1(RIDuser),H1(IDm) ); wherein e is
Figure BDA0003495181460000111
And
Figure BDA0003495181460000112
linear mapping of (2);
step 15: server receives { Mm,ζmAfter that, the timestamp T of the uninstall request message is checkedsIf not, the authentication fails, if so, the symmetric key K of the server is calculated according to the public parameters of the systemmth-user＝e(H1(RIDuser),wH1(IDm))；
Step 16: matching the symmetric key of the server with the symmetric key of the user side, failing authentication if the symmetric keys are not matched, and generating a check code C by the server if the symmetric keys are matchedm＝H2(Kmth-user||RIDuser||IDm) And check the message { RIDuser,IDm,CmSending the data to the user side;
and step 17: after receiving the check message, the user end generates a verification code Cuser＝H2(Kuser-mth||RIDuser||IDm) Verification of CmAnd CuserIf the server passes the verification, the server is a safe connection server, otherwise the server is not connectable.
The process of setting the channel gain threshold according to the offloading task information includes:
step 1: obtaining BHz bandwidth and T time delay occupied by user segment of system modelDAccording to BHz and time delay TDCalculating the number of channels available to the user, i.e. N ═ BTD；
Step 2: acquiring a channel state information matrix g of a user terminal and a servermAnd calculating post-processing signal-to-noise ratio gamma of the user side and the server according to the channel state information matrix; the post-processing signal-to-noise ratio is formulated as:
Figure BDA0003495181460000113
wherein p isuIndicating the user side transmission power, ENmIs the channel estimation error noise, ZNmIs additive white Gaussian noise, | | gm||2Is the channel gain;
and step 3: calculating the shannon capacity C (gamma) of the signal-to-noise ratio;
C(γ)＝log2(1+γ)
and 4, step 4: correcting the post-processing signal-to-noise ratio to obtain a corrected post-processing signal-to-noise ratio; the correction formula is as follows:
Figure BDA0003495181460000121
and 5: calculating the error probability of the server by adopting a Q function Q (w) according to the shannon capacity C (gamma) and the corrected post-processing signal-to-noise ratio V (gamma);
the expression of the Q function is:
Figure BDA0003495181460000122
the formula for calculating the error probability of the server is:
Figure BDA0003495181460000123
wherein n represents the number of available channels for pilot; d represents the size of the task data;
step 6: calculating the data transmission rate of the user side according to the error probability; the transmission rate is calculated by the formula:
Figure BDA0003495181460000124
and 7: according to the requirements of the ultra-reliable low-delay communication scene on time delay and error probability, the bit number of short packet transmission and the total transmission bandwidth, the threshold value | g of channel gain is solvedth||2。
Further, the requirements of the ultra-reliable low-delay communication scene on the time delay and the error probability comprise that the time delay is less than 0.5 millisecond, and the error probability is less than 10-5。
The process of executing the server screening policy for verifying the security of the server after verifying the channel gain comprises:
step 1: acquiring task information of a user side, and dividing a task into a local task and an unloading task according to the task information;
step 2: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and step 3: calculating channel gain according to the unloading task information; setting a channel gain threshold; comparing the calculated channel gain with a set channel gain threshold, and if the calculated channel gain is greater than or equal to the set channel gain threshold, executing the step 4; otherwise the server does not meet the unloading requirement;
and 4, step 4: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication succeeds, the server meets the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
The set channel gain threshold is: the channel parameter between User and mthMES is gmThen, then believeThe channel gain is gm||2Setting the channel gain threshold to | | >h||2. When the distribution number of the subtasks, the maximum value of the time delay and the minimum value of the error probability are known, the signal-to-noise ratio of the system after processing can be calculated, and then the channel gain threshold value | | | g is determinedth||2。
Whether User asserts MES as a securely connectable server and the channel gain threshold are two criteria for selecting the cooperative MESs. The cooperative MESs select event data set is marked as F, a decision tree algorithm is adopted, two criteria are verified through different sequences, the result classification information gains G (F) are different, and the cooperative MESs select event data set can be classified into
Figure BDA0003495181460000131
The value of the damping is expressed as
Figure BDA0003495181460000132
pselIs the probability that the MES was selected.
For the method of verifying MES security before channel gain, the Keyny coefficient can be expressed as:
Figure BDA0003495181460000133
with an information gain of Gsg(F)＝Gini(F)-Gini_index(F,sg)，FsgData set of a method for verifying the gain of a channel after verifying the security of a MES, FgsA data set for a method of verifying MES security after verifying channel gain.
For the method for verifying MES safety after verifying channel gain, because the channel gain is a numerical attribute, data needs to be sorted in ascending order at first, and then the middle value of two adjacent values of a sample is used as a segmentation point xi from small to large in sequencei(1≤i＜NF-1)，NFFor the number of sample values, the samples in the data set are divided into two groups according to the classification mode of being larger than the division point and being smaller than the division point, and each group is calculatedDichotomous damping coefficient of cut point
Figure BDA0003495181460000134
The division point with highest purity and best classification effect is
Figure BDA0003495181460000135
The optimal information gain of the method is as follows:
Ggs(F)＝Gini(F)-Gini_index(F,ξoptimum)
in order to make the final selection more effective, a division method which can provide the maximum information gain to the system should be selected when the DT is classified. Comparative Gsg(F) And Ggs(F) If, if
Figure BDA0003495181460000141
Selecting a method for verifying MES safety and then verifying channel gain; if G issg(F)＜Ggs(F) Selecting a method for verifying MES safety after verifying channel gain; if G issg(F)＝Ggs(F) Optionally, one of the two methods may be used as the collaboration server selection method, in which case the present invention sets the method to verify the MES security after verifying the channel gain.
The DT selects MES cooperation users which simultaneously meet the security verification and the channel gain verification to carry out task unloading according to the attributes (such as data quantity and delay threshold value) of real-time tasks transmitted by the User and the sequence of server judgment criteria, and the User unloads the tasks to each MES cooperation process according to the suggestion of the DT.
The above-mentioned embodiments, which further illustrate the objects, technical solutions and advantages of the present invention, should be understood that the above-mentioned embodiments are only preferred embodiments of the present invention, and should not be construed as limiting the present invention, and any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A task offload edge server selection method based on digital twinning is characterized by comprising the following steps: constructing a system model and a digital twin safety authentication system, and initializing the digital twin safety authentication system; a user sends a task unloading request to a server; the digital twin layer screens the servers according to the user task unloading request information to obtain the servers meeting the unloading request; and unloading the tasks of the users to the screened servers.
2. The method for task offload edge server selection based on digital twinning as claimed in claim 1, wherein the process of initialization of the digital twinning security authentication system comprises: setting system parameter groups
Figure FDA0003495181450000011
And key group
Figure FDA0003495181450000012
In the system parameter group
Figure FDA0003495181450000013
Generating parameter S at random, selecting positive integer
Figure FDA0003495181450000014
As a primary cryptographic parameter; calculating a system public key K according to the parameter S and the positive integer wpubwS; selecting two hash functions H1:
Figure FDA0003495181450000015
And H2:
Figure FDA0003495181450000016
The privacy public parameter of the system is
Figure FDA0003495181450000017
Wherein
Figure FDA0003495181450000018
A group of system parameters is represented, and,
Figure FDA0003495181450000019
representing a key group, S representing a randomly generated parameter, KpubRepresenting the system public key, H1Representing a first hash function, H2Representing a second hash function.
3. The method as claimed in claim 1, wherein the step of screening the authenticated server by the digital twin layer according to the user task offload request information comprises:
s1: calculating a first classification information gain and a second classification information gain;
s2: comparing the first classification information gain with the second classification information gain, and if the first classification information gain is larger than the second classification information gain, executing a server screening strategy for verifying the channel gain after the security of the server is verified; if the gain of the first classification information is less than or equal to the gain of the second classification information, executing a server screening strategy for verifying the security of the server after verifying the gain of the channel;
s3: collecting all servers meeting the unloading requirements, and acquiring the number of services meeting the requirements; and judging whether the number of the services meeting the requirement meets the subtask distribution number, if so, ending, and if not, returning to the step S1.
4. The method of claim 3, wherein the step of calculating the first classification information gain and the second classification information gain comprises:
the first classification information gain calculation process is as follows:
step 1: obtaining probability p of server being selectedsel；
Step 2: according to pselCalculating the Gini value (F) of the server;
and step 3: calculating a Gini _ index (F, sg) of the server according to the Gini value;
and 4, step 4: calculating a first classification information gain G based on the Gini value Gini (F) and the Gini coefficient Gini _ index (F, sg)sg(F)；
The second classification information gain calculation process includes:
step 1: acquiring a data set of a user side; and sorting the data in the data set in ascending order, and using the intermediate value of two adjacent values of the sample as a division point xi in sequence from small to largei(1≤i＜NF-1)，NFIs the number of sample values;
step 2: dividing the samples in the data set into two groups according to the classification mode of being larger than the division point and being smaller than the division point, and calculating the two classification kini coefficients Gini _ index (F, xi) of each division pointi)；
And step 3: screening out the division point with the best classification effect according to the two classification damping coefficients
Figure FDA0003495181450000021
And 4, step 4: calculating a second classification information gain G according to the division point with the best classification effect and the classification damping coefficientgs(F)。
5. The method of claim 3, wherein performing a server screening strategy that verifies server security prior to channel gain comprises:
step 1: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication is successful, executing the step 2;
step 2: acquiring task information of a user side, and dividing tasks into local tasks and unloading tasks according to the task information;
and step 3: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and 4, step 4: setting a channel gain threshold according to the unloading task information; comparing the channel gains of the user side and the service side with a set channel gain threshold, and if the channel gains of the user side and the service side are more than or equal to the set channel gain threshold, the server meets the unloading requirement; otherwise the server does not meet the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
6. The method as claimed in claim 5, wherein the process of authenticating the user side and the edge server by using the digital twin security authentication system comprises:
step 11: inputting the user terminal and the server into a digital twin safety authentication system; the digital twin safety authentication system respectively distributes an identity identification code and a random identity identification code for the user terminal and the server;
step 12: the user terminal informs the digital twin layer DT of the unloading task and requests the DT layer to select a proper server MES;
step 13: user side using unloading request information Mm＝(RIDuser||IDm||Ts) And private key wH1(IDm) Calculating the signature ζm＝H2(M)·wH1(IDm) (ii) a Where, | | represents information data connection, TsRepresenting a time stamp, RIDuserIdentity, ID, indicating the user's endmRandom ID code representing server, w represents random positive integer, H1Representing a first hash function, H2Representing a second hash function, M representing offload request information;
step 14: the user side will { Mm,ζmSending the key to a server mthMES, and calculating a symmetric key K of a user side and the serveruser-mth＝e(wH1(RIDuser),H1(IDm) ); wherein e is
Figure FDA0003495181450000031
And
Figure FDA0003495181450000032
linear mapping of (2);
step 15: server receives { Mm,ζmAfter that, the timestamp T of the uninstall request message is checkedsIf not, the authentication fails, if so, the symmetric key K of the server is calculated according to the public parameters of the systemmth-user＝e(H1(RIDuser),wH1(IDm))；
Step 16: matching the symmetric key of the server with the symmetric key of the user side, failing authentication if the symmetric keys are not matched, and generating a check code C by the server if the symmetric keys are matchedm＝H2(Kmth-user||RIDuser||IDm) And will check the message { RIDuser,IDm,CmSending the data to the user side;
and step 17: after receiving the check message, the user end generates a verification code Cuser＝H2(Kuser-mth||RIDuser||IDm) Verification CmAnd CuserIf the server passes the verification, the server is a safe connection server, otherwise the server is not connectable.
7. The method for selecting a digital twin-based task offload edge server according to claim 5, wherein the process of setting the channel gain threshold according to the offload task information comprises:
step 1: obtaining BHz bandwidth and T time delay occupied by user segment of system modelDAccording to BHz and time delay TDCalculating the number of channels available to the user, i.e. N ═ BTD；
Step 2: acquiring a channel state information matrix g of a user terminal and a servermAnd calculating post-processing signal-to-noise ratio gamma of the user side and the server according to the channel state information matrix; the post-processing signal-to-noise ratio is formulated as:
Figure FDA0003495181450000041
wherein p isuIndicating the user side transmission power, ENmIs the channel estimation error noise, ZNmIs additive white Gaussian noise, | | gm||2Is the channel gain;
and step 3: calculating shannon capacity C (gamma) of post-processing signal-to-noise ratio;
C(γ)＝log2(1+γ)
and 4, step 4: correcting the post-processing signal-to-noise ratio to obtain a corrected post-processing signal-to-noise ratio; the correction formula is as follows:
Figure FDA0003495181450000042
and 5: calculating the error probability of the server by adopting a Q function Q (w) according to the shannon capacity C (gamma) and the corrected post-processing signal-to-noise ratio V (gamma);
Figure FDA0003495181450000043
wherein n represents the number of available channels for pilot; d represents the size of the task data;
step 6: calculating the data transmission rate of the user side according to the error probability;
and 7: according to the requirements of the ultra-reliable low-delay communication scene on time delay and error probability, the bit number of short packet transmission and the total transmission bandwidth, the threshold value | g of channel gain is solvedth||2。
8. The method as claimed in claim 7, wherein the formula for calculating the data transmission rate at the user end is:
Figure FDA0003495181450000051
where ε is the error probability; n is the number of available channels for pilot; gamma is the post-processing signal-to-noise ratio.
9. The method as claimed in claim 7, wherein the requirements of the ultra-reliable low-latency communication scenario for latency and error probability include latency below 0.5 ms and error probability below 10 ms-5。
10. The method as claimed in claim 3, wherein the step of executing the server screening strategy for verifying the security of the server after verifying the channel gain comprises:
step 1: acquiring task information of a user side, and dividing tasks into local tasks and unloading tasks according to the task information;
step 2: acquiring information of unloading tasks, wherein the information comprises the task distribution number, the maximum tolerable time delay and the minimum error probability;
and step 3: calculating channel gain according to the unloading task information; setting a channel gain threshold; comparing the calculated channel gain with a set channel gain threshold, and if the calculated channel gain is greater than or equal to the set channel gain threshold, executing the step 4; otherwise the server does not meet the unloading requirement;
and 4, step 4: authenticating the user side and the edge server by adopting a digital twin safety authentication system, wherein if the authentication of the user side and the edge server fails, the edge server does not meet the task unloading requirement; if the authentication is successful, the server meets the unloading requirement;
and 5: and performing collective output on all servers meeting the requirements.
CN202210111471.4A 2022-01-29 2022-01-29 Task unloading edge server selection method based on digital twin Active CN114466356B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210111471.4A CN114466356B (en) 2022-01-29 2022-01-29 Task unloading edge server selection method based on digital twin

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210111471.4A CN114466356B (en) 2022-01-29 2022-01-29 Task unloading edge server selection method based on digital twin

Publications (2)

Publication Number Publication Date
CN114466356A true CN114466356A (en) 2022-05-10
CN114466356B CN114466356B (en) 2022-10-14

Family

ID=81410836

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210111471.4A Active CN114466356B (en) 2022-01-29 2022-01-29 Task unloading edge server selection method based on digital twin

Country Status (1)

Country Link
CN (1) CN114466356B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110933118A (en) * 2020-02-20 2020-03-27 深圳市城市交通规划设计研究中心股份有限公司 Edge computing gateway secure communication method, system, terminal equipment and server
CN111447619A (en) * 2020-03-12 2020-07-24 重庆邮电大学 Joint task unloading and resource allocation method in mobile edge computing network
CN112492626A (en) * 2020-12-07 2021-03-12 南京邮电大学 Method for unloading computing task of mobile user
US20210314417A1 (en) * 2020-04-03 2021-10-07 Toyota Motor Engineering & Manufacturing North America, Inc. Digital twin-based edge server switching decision
CN113572804A (en) * 2021-04-29 2021-10-29 重庆工程职业技术学院 Task unloading system, method and device based on edge cooperation
CN113590232A (en) * 2021-08-23 2021-11-02 南京信息工程大学 Relay edge network task unloading method based on digital twinning

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110933118A (en) * 2020-02-20 2020-03-27 深圳市城市交通规划设计研究中心股份有限公司 Edge computing gateway secure communication method, system, terminal equipment and server
CN111447619A (en) * 2020-03-12 2020-07-24 重庆邮电大学 Joint task unloading and resource allocation method in mobile edge computing network
US20210314417A1 (en) * 2020-04-03 2021-10-07 Toyota Motor Engineering & Manufacturing North America, Inc. Digital twin-based edge server switching decision
CN112492626A (en) * 2020-12-07 2021-03-12 南京邮电大学 Method for unloading computing task of mobile user
CN113572804A (en) * 2021-04-29 2021-10-29 重庆工程职业技术学院 Task unloading system, method and device based on edge cooperation
CN113590232A (en) * 2021-08-23 2021-11-02 南京信息工程大学 Relay edge network task unloading method based on digital twinning

Also Published As

Publication number Publication date
CN114466356B (en) 2022-10-14

Similar Documents

Publication Publication Date Title
Lu et al. Low-latency federated learning and blockchain for edge association in digital twin empowered 6G networks
US11017322B1 (en) Method and system for federated learning
Yang et al. Edge coordinated query configuration for low-latency and accurate video analytics
CN112492626B (en) Method for unloading computing task of mobile user
CN110941667A (en) Method and system for calculating and unloading in mobile edge calculation network
CN111132175A (en) Cooperative computing unloading and resource allocation method and application
CN111629052B (en) Content caching method, node, equipment and storage medium based on MEC
CN113315978B (en) Collaborative online video edge caching method based on federal learning
CN114968404A (en) Distributed unloading method for computing task with position privacy protection
CN114301935A (en) Reputation-based method for selecting edge cloud collaborative federated learning nodes of Internet of things
Krouka et al. Communication-efficient split learning based on analog communication and over the air aggregation
CN116669111A (en) Mobile edge computing task unloading method based on blockchain
Merluzzi et al. Energy-efficient classification at the wireless edge with reliability guarantees
Li et al. Anycostfl: Efficient on-demand federated learning over heterogeneous edge devices
CN111866181B (en) Block chain-based task unloading optimization method in fog network
CN114466356B (en) Task unloading edge server selection method based on digital twin
CN117252253A (en) Client selection and personalized privacy protection method in asynchronous federal edge learning
Zhang et al. A deep reinforcement learning approach to multiple streams’ joint bitrate allocation
Ji et al. Efficiency-Boosting Federated Learning in Wireless Networks: A Long-Term Perspective
Shi et al. CoLEAP: Cooperative learning-based edge scheme with caching and prefetching for DASH video delivery
CN115378788A (en) Block chain performance self-adaptive optimization method based on hierarchical consensus and reinforcement learning
CN115022684A (en) Video stream self-adaptive transmission method based on deep reinforcement learning under QUIC protocol
CN116489683B (en) Method and device for unloading computing tasks in space-sky network and electronic equipment
CN117221122B (en) Asynchronous layered joint learning training method based on bandwidth pre-allocation
CN116911382A (en) Asynchronous aggregation and privacy protection method in resource-limited federal edge learning

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