CN110502328B - Offshore edge computing trusted cooperative task migration method - Google Patents
Offshore edge computing trusted cooperative task migration method Download PDFInfo
- Publication number
- CN110502328B CN110502328B CN201910803509.2A CN201910803509A CN110502328B CN 110502328 B CN110502328 B CN 110502328B CN 201910803509 A CN201910803509 A CN 201910803509A CN 110502328 B CN110502328 B CN 110502328B
- Authority
- CN
- China
- Prior art keywords
- node
- edge computing
- service
- queue
- edge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
- G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
-
- 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
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a marine Edge Computing credible collaborative task migration Method (A Method of trusted Computing Services for Markime Edge Computing Network, TCSMEC), which is based on collaborative filtering thought, adopts Network heartbeat information piggybacking technology, fuses node similarity relation, network behavior, link characteristics and neighbor similarity nodes to construct a local Trust Network, carries out recommendation based on characteristics such as service capability, service quality, load and Trust relation among the nodes, adopts a quartile Method to filter service quality evaluation information to improve recommendation effectiveness, realizes self-adaptive Edge Computing service access and rapid resource fusion through queue hierarchical control, and simulation experiments show that TCSMEC has obvious advantages in service energy consumption, load efficiency, link transmission rate and the like compared with ADOV and RandomWalk methods, and TCSMEC can efficiently realize marine weather Edge Computing credible task migration service in complex marine weather environment.
Description
Technical Field
The invention relates to the field of offshore edge computing, in particular to a method for migrating a trusted cooperative task in offshore edge computing.
Background
With the rapid development of technologies such as cloud computing, big data and internet of things and the popularization and application of mobile intelligent equipment, the applications of large flow, large connection and delay sensitivity such as marine internet of things, multimedia and instant messaging become main services of marine communication network flow, and higher requirements are provided for the service capability of marine communication services. The maritime wireless communication link is wide in coverage area and variable in environment, and the maritime communication node faces a plurality of resource constraints such as energy, calculation, storage and communication, so that the quality of service is difficult to guarantee when maritime wireless communication service is implemented.
Mature network technologies such as cellular networks, wireless metropolitan area networks, local area networks and the like based on base stations on land and capable of solving the problems are applied to ports and offshore area communication networks based on shore-based base stations as centers, however, the coverage area is limited, and the requirements for building marine communication networks cannot be met. At present, the existing offshore wireless broadband Network architecture research is mainly based on wireless Mesh, ad Hoc, DTN (Delay to Network), TD-LTE (Time Division Long Term Evolution), and the like, and the Network service bandwidth is relatively limited, and the requirement of offshore Network application service cannot be met due to insufficient Delay, high reliability, and the like. Meanwhile, the development of marine communication is obviously lagged behind land communication due to the complex and changeable marine environment, difficult construction and the like, and the development of a new generation of marine communication technology and system becomes a focus of attention in academia and industry.
The offshore edge computing network realizes low energy consumption of the offshore mobile terminal, high reliability of the offshore communication network and high efficiency of the offshore application business service by transferring part of computing tasks of the unmanned ship and the shipborne communication equipment to the edge cloud network. Because the edge calculation is closer to one side of a data source, an open platform integrating core capabilities of network, calculation, storage, application and the like is adopted, the network service can be responded more quickly, the requirements of the offshore edge calculation on instantaneity, reliability and high efficiency are met, and the wide attention is paid.
The main problems encountered in the construction of the offshore edge computing network are that the marine environment is extremely complex and changeable, most of the marine environment is unmanned island reef and is extremely uneven in distribution, the resource is limited, the construction is difficult, and the networking of land with the base station as the core is difficult to realize. Meanwhile, resources such as computation, storage, energy, bandwidth and the like of the offshore node are relatively limited, and phenomena such as chain breakage, single-point failure, over-high load and the like may exist when edge computing service is implemented. In order to construct a marine edge computing network and self-adaptively bear marine large-scale flow service, the edge computing network is constructed by taking a shore-based base station, an island reef base station and an intelligent relay floating platform as interactive cores, and edge computing task migration and resource scheduling are implemented based on mutual cooperation of marine nodes and edge nodes, wherein a key problem is how to construct an effective task migration mechanism.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a Method (A Method of trusted collaborative Services for landmark Edge Computing Network, TCSMEC) for migrating a marine Edge Computing trusted collaborative task.
The technical scheme for realizing the purpose of the invention is as follows:
a method for migrating a trusted cooperative task in offshore edge computing is characterized in that a network heartbeat information piggybacking technology is adopted, a similar trust relationship among adjacent nodes is fused, and based on behavior and real-time state characteristics of an edge computing service node, the edge computing service node is recommended to implement task migration, and the method comprises the following steps:
1) Supposing that N nodes exist in the offshore edge computing network, the adjacency relation matrix is N ij ＝(a ij ) n×n ，a ij ∈{0,1}，a ij =1 indicates that the node i and the node j are neighbors, and the edge computing service relation matrix is M kl ＝(b kl ) n×n ，b kl ∈{0,1}，b ik =1 indicates that the node i may request the edge computing service from the node k, and every t periods, the node sends a heartbeat packet containing its own feature information to a peripheral node to maintain connection information, and when receiving heartbeat information from a neighbor node, the offshore node q will perform the following operations:
constructing or updating its neighbor relation vector N ij | i＝q ＝{a q1 ,a q2 ,…,a qn }，
Building or updating its edge computing service relation vector M ik | i＝q ＝{b q1 ,b q2 ,…,b qn }，
Building or updating neighbor node trust relation vector
Constructing or updating edge computing service node service capability evaluation vector G ij | i＝q ＝{g q1 ,g q2 ,…,g qm In which τ is qj Satisfies the following formula (1):
the above formula(1) In, t s Indicating that the node receives the heartbeat data packet s times, rtt (i, j) indicating the communication delay between the node i and the node j, jac (i, j) indicating the jaccard coefficient of the edge node i and the edge node j, at j Represents the average online time of node j, t succ Indicates the number of times of forming a valid recommendation, t all Representing total recommendation times, e is a natural base number, and evaluating node service capability g qk Satisfies formula (2):
w k 、h k 、Rm k respectively representing the bandwidth, the CPU frequency and the memory size of an edge computing service node k;
2) When a neighbor node finishes edge calculation service, the neighbor node makes service quality evaluation and adds the service quality evaluation into a heartbeat data packet to inform other nodes, and after receiving the service quality evaluation from the neighbor node, a node q constructs or updates a service quality evaluation matrix S jk ＝(s jk ) n×n ，s jk ∈[0,1]；
3) The node q constructs or updates a priority queue L according to the service capability of the edge computing service node, the service quality evaluation of the neighbor node and the trust relationship;
4) When the node q needs edge computing service, selecting a proper edge computing service node from the priority queue to initiate a task migration request, wherein in order to reduce the search cost, the priority queue L adopts random hash search or round search;
5) After receiving the task migration request response, the node q migrates the task to the edge computing service node, and the edge computing service node aggregates service resources according to the task requirement, executes the computing task and returns the result to the node q;
6) The node q evaluates the edge calculation service s qk According to s qk For all the neighbors of the recommended edge computing service node k, updating the neighbor trust relationship through a formula (1)
Meanwhile, the node q piggybacks evaluation information of the task migration service in a heartbeat data packet to inform a neighbor node, and the neighbor node updates a service quality evaluation matrix S after receiving the feedback of the node q jk 。
The step 3) specifically comprises the following steps:
3-1) the node q calculates the service quality score s of the service node to the edge of any neighbor node according to the trust relationship of the neighbor nodes jk ∈S jk Making a correction, the corrected score u q,jk As formula (3):
u q,jk ＝b qk τ qj s jk (3)
3-2) compute the service node k for each edge, b qk Revised quality of service score u not equal to 0 q,jk Filtering scores which are too high or too low by adopting a quartile method;
3-3) after the filtering is finished, the node q calculates the priority eta of the edge calculation service node k k As in equation (4):
in the above formula (4), Q (k) is a load evaluation function of the edge computing service node k, p k For memory occupancy, ρ k For CPU utilization, d k For the bandwidth occupancy rate, p, rho and d are respectively a memory load threshold, a calculation load threshold and a bandwidth load threshold, and a node q calculates the priority eta of a service node k according to the edge k Eta. Mixing k >The edge calculation service node of alpha is arranged in a first priority queue L 1 Will beta<η k <The edge calculation service node of alpha is arranged in a second priority queue L 2 Eta. Mixing k <The edge compute service node of beta is enqueued in the third queue L 3 Where α and β are resolution thresholds, p, ρ, d, α, β ∈ [0,1 ]]，β<α，L 1 、L 2 、L 3 All are circular queues, and before the edge computing service node is added into the queue, the edge computing service node is added into the queueWhether or not there is an offline node (b) qx = 0), if so, the node is replaced, otherwise the end of the queue is added.
The step 3-2) specifically comprises the following steps:
3-2-1) calculating a revised score u for a serving node k based on each neighbor to edge q,jk Arranging from small to large, taking the total number of scores as O, and taking
Correction score u q,jk Is marked as a lower bound c 1 Get the first
A correction score u q,jk Is marked as an upper bound c 2 Wherein, in the step (A),
represents the largest integer no greater than x;
3-2-2) removal
Score term of (3), calculate u q,jk ∈[c 1 ,c 2 ]Average value of (2)
The step 4) specifically comprises the following steps:
4-1) if L 1 If the queue is not empty, selecting an edge computing service node to initiate a request by adopting a random hash function, if a response is received, indicating that the edge computing service node receives a task migration request, completing the request, if the response is overtime, marking the node, and sequentially selecting the L where the node is located 1 The predecessor node or successor node of the queue initiates a request, if the response is still overtime, the two nodes are marked, the hash address H is generated again, the steps are repeated, and if the hash address H is generated again>L 1 The length of the queue, the slave L is abandoned 1 Selecting a node in the queue to initiate a request;
4-2) ifL 2 If the queue is not empty, node q will L 2 The edge computing service nodes of the queue sequentially initiate task migration requests according to the sequence of priority from large to small; if the response is received, the edge computing service node is indicated to receive the task migration request, and the request is completed; if the response is overtime, continuing to search in turn; number of times of looking for>3, if no node responds to the request, abandoning in L 2 Selecting an edge computing service node in the queue to initiate a request;
4-3) node q is at L 3 Queue random hash selection edge computing service node initiates a request, and L 1 The searching process of the queue is similar if the hash address H is regenerated for times>L 3 And if the length of the queue is long, the task migration request is abandoned.
The invention has the beneficial effects that: the invention provides a method for migrating a trusted cooperative task for offshore edge computing, which is based on a cooperative filtering thought and adopts a network heartbeat information piggybacking technology to fuse a node similarity relation, a network behavior, a link characteristic and a neighbor similarity node to construct a local trust network, and recommends based on characteristics such as service capability, service quality, load and trust relation among the nodes of an edge computing service node, adopts a quartile method to filter service quality evaluation information to improve recommendation effectiveness, and realizes self-adaptive edge computing service access and rapid resource fusion through queue hierarchical control.
Drawings
FIG. 1 is an illustration of a scenario of a marine edge computing collaboration service system;
FIG. 2 is a diagram of a topology of a marine edge computing collaboration service system;
FIG. 3 is a comparison graph of marine node connectivity in different marine meteorological environments;
FIG. 4 is a graph comparing the loads of different oceanographic environment base stations;
FIG. 5 is a graph comparing average loads of marine nodes in different marine meteorological environments;
FIG. 6 is a comparison graph of average transmission delays of links in different marine meteorological environments;
FIG. 7 is a comparison graph of success rate of delivery of migration tasks in different oceanographic environments;
FIG. 8 is a graph of mean residual energy comparison for a cooperative node;
FIG. 9 is a graph comparing average computation loads of cooperative nodes;
FIG. 10 is a graph comparing average link transmission rates of cooperative nodes;
FIG. 11 is a comparison graph of average residual energy of cooperative nodes in different oceanographic environments;
FIG. 12 is a graph comparing link transmission rates for different metocean environments.
Detailed Description
The invention is further illustrated but not limited by the following figures and examples.
Example (b):
an offshore edge computing cooperative service system is constructed by taking an offshore port as an example, and is shown in figure 1. The offshore edge computing network consists of a network Center Cloud Server (CDCS), a shore-based Center Base Station (CBS), an island/intelligent floating platform Relay Base Station (RBS) and a common Node (Node, N); the general nodes refer to autonomous communication equipment on unmanned ships and ships, and can form an ad hoc offshore communication network, and each node is accessed to at least one base station to use the offshore mobile communication network service. In order to maintain low management overhead, it is assumed that a heartbeat information period t =5s, when each offshore node sends heartbeat information, the heartbeat information is only shared by the offshore nodes located in the same subnet group, the number n of node neighbors is less than or equal to 60, and meanwhile, if the node does not feed back the heartbeat information in 3 periods, the node is considered to be offline.
The test time is 6 hours, the scale N of network nodes is less than or equal to 1500,1 network center, 1 shore-based center base station, 3 island relay base stations, 3 intelligent floating platforms and a plurality of common nodes. The access number of each base station is less than 240, the network topology is a distributed random topology structure, and nodes can dynamically and unrestrained join and leave the network. All nodes can randomly initiate task migration requests, the migration frequency f is less than 1/5 s, network information statistics analysis sampling is carried out every 5 seconds, experimental tests mainly explore the influence on the offshore edge computing cooperative service system under different oceanographic environments, and the test contrast setting is as follows:
1. a maritime edge computing cooperative service system under a windless environment on a sunny day;
2. a marine edge computing cooperative service system in a cloud environment;
3. a marine edge computing cooperative service system in a snowing environment;
4. the offshore edge computing cooperative service system under the rainfall environment.
The topology structure of the initial state offshore edge computing network system is shown in fig. 2, and based on the topology structure, if the connection attenuation strength RSSI between nodes meets RSSI > -100dB, the nodes are regarded as neighbor nodes capable of communicating and cooperating. Through experimental tests under weather environments such as clear weather, cloud and fog, snowfall, rainfall and the like on the simulated sea, the number of node neighbors is counted, as shown in the following figure 3, the horizontal axis is the number of the node neighbors, the vertical axis is the number of the nodes, the connection strength of the nodes under different weather environments generally obeys normal distribution, and the average values of the network node stable connection degrees are respectively 30, 20, 18 and 8. Therefore, the network node connectivity in different oceanic meteorological environments is in a descending trend, the sparsity of the number of neighbor nodes meets the requirements of a clear-sky environment, a cloud-fog environment, a snowfall environment and a rainfall environment, and the method for migrating the marine edge computing credible cooperative task specifically comprises the following steps:
1) Supposing that N nodes exist in the offshore edge computing network, the adjacency relation matrix is N ij ＝(a ij ) n×n ，a ij ∈{0,1}，a ij =1 indicates that the node i and the node j are neighbors, and the edge computing service relation matrix is M kl ＝(b kl ) n×n ，b kl ∈{0,1}，b ik =1 indicates that the node i can request the edge computing service to the node k, every t time interval, the node sends a heartbeat data packet containing its own characteristic information to the peripheral node to maintain the connection information, and after receiving the heartbeat information from the neighbor node, the offshore node q will execute the connection informationThe following operations are carried out:
constructing or updating its adjacency vector N ij | i＝q ＝{a q1 ,a q2 ,…,a qn }，
Building or updating its edge computing service relation vector M ik | i＝q ＝{b q1 ,b q2 ,…,b qn }，
Building or updating neighbor node trust relation vector
Constructing or updating edge computing service node service capability evaluation vector G ij | i＝q ＝{g q1 ,g q2 ,…,g qm }，
Wherein, tau qj Satisfies formula (1):
wherein, t s Representing that the node receives the heartbeat data packet s time, rtt (i, j) representing the communication delay between the node i and the node j, jac (i, j) representing the jaccard coefficient of the edge node i and the edge node j, at j Represents the average online duration, t, of node j succ Indicates the number of times of forming a valid recommendation, t all Representing the total number of recommendations, e is a natural base number,
in the formula for calculating (a) of (b),
service similarity and interactive behavior characteristics among nodes are fused,
the coulomb force between nodes is expressed, and the coulomb force is introduced to relieve low recommendation efficiency caused by sparsity of mutual information,
node service capability evaluation g qk Satisfies formula (2):
w k 、h k 、Rm k respectively representing the bandwidth, the CPU frequency and the memory size of the edge computing service node k.
2) When a neighbor node finishes edge calculation service, the neighbor node makes service quality evaluation and adds the service quality evaluation into a heartbeat data packet to inform other nodes, and after receiving the service quality evaluation from the neighbor node, a node q constructs or updates a service quality evaluation matrix S jk ＝(s jk ) n×n ，s jk ∈[0,1]。
3) The node q constructs or updates a priority queue L according to the service capability of the edge computing service node, the service quality evaluation of the neighbor node and the trust relationship, and the method comprises the following specific steps:
3-1) the node q calculates the service quality score s of the service node to the edge of any neighbor node according to the trust relationship of the neighbor nodes jk ∈S jk Making a correction, the corrected score u q,jk As in equation (3):
u q,jk ＝b qk τ qj s jk (3)
3-2) compute service node k for each edge, b qk Not equal to 0 revised quality of service score u q,jk Filtering scores which are too high or too low by adopting a quartile method, and specifically comprising the following steps of:
3-2-1) calculating a revised score u for the serving node k based on each neighbor to edge q,jk Arranging from small to large, taking the total number of scores as O, and taking
Correction score u q,jk Is marked as a lower bound c 1 Get it first
A correction score u q,jk Is marked asUpper bound of c 2 Wherein, in the step (A),
represents the largest integer no greater than x;
3-2-2) removal of
Score term of (3), calculate u q,jk ∈[c 1 ,c 2 ]Average value of (2)
3-3) after the filtering is finished, the node q calculates the priority eta of the edge computing service node k k As in equation (4):
wherein Q (k) is a load evaluation function of the edge computing service node k, p k For memory occupancy, ρ k For CPU utilization, d k For bandwidth occupancy, p, rho and d are respectively a memory load threshold, a calculation load threshold and a bandwidth load threshold, and a node q calculates the priority eta of a service node k according to the edge k Eta is to k >The edge calculation service node of alpha is listed in a first priority queue L 1 Will beta<η k <The edge calculation service node of alpha is arranged in a second priority queue L 2 Eta is to k <The edge compute service node of beta is enqueued in the third queue L 3 Where α and β are resolution thresholds, p, ρ, d, α, β ∈ [0,1 ]]，β<α，L 1 、L 2 、L 3 All are circular queues, and before the edge computing service node is added into the queue, whether an off-line node (b) exists in the queue is searched qx = 0), if so, replacing the node, otherwise, joining the tail of the queue.
4) When the node q needs edge computing service, a suitable edge computing service node is selected from the priority queue to initiate a task migration request, and L is used for reducing search overhead 1 、L 3 Using random hash search, L 2 The method adopts the round searching and comprises the following specific steps:
4-1) if L 1 If the queue is not empty, selecting an edge computing service node to initiate a request by adopting a random hash function, if a response is received, indicating that the edge computing service node receives a task migration request, completing the request, if the response is overtime, marking the node, and sequentially selecting the L where the node is located 1 The predecessor node or successor node of the queue initiates a request, if the response is still overtime, the two nodes are marked, the hash address H is generated again, the steps are repeated, and if the hash address H is generated again>L 1 The length of the queue is then abandoned from L 1 And selecting a node in the queue to initiate the request.
4-2) if L 2 If the queue is not empty, node q will L 2 The edge computing service nodes of the queue initiate task migration requests in sequence from large to small according to the priority, if responses are received, the edge computing service nodes are indicated to receive the task migration requests, the requests are completed, if the responses are overtime, round searching is continued, and if the round searching times are counted>3, if no node responds to the request, abandoning in L 2 And selecting an edge computing service node in the queue to initiate a request.
4-3) node q is at L 3 Queue random hash selection edge computing service node initiates a request, and L 1 The searching process of the queue is similar if the times of regenerating the Hash address H>L 3 And if the queue is long, the task migration request is abandoned.
5) After receiving the task migration request response, the node q migrates the task to the edge computing service node, the edge computing service node aggregates service resources according to the task requirement, executes the computing task, and returns the result to the node q; when the edge computing service node executes the migration task, the migration request source, the transmission link information and the like are stored, and when data loss and incomplete information occur, emergency scheduling is directly performed from the priority link.
6) Node q makes evaluation s for the edge calculation service qk According to s qk For all the neighbors of the recommended edge computing service node kUpdating neighbor trust relationships via equation (1)
Meanwhile, the node q piggybacks evaluation information of the task migration service in a heartbeat data packet to notify a neighbor node, and the neighbor node updates a service quality evaluation matrix S after receiving feedback of the node q jk 。
7) As shown in fig. 4, with the addition of nodes in the network, the initial base station node has good network environment and service capability, and the network nodes all adopt a casual vehicle taking strategy of intelligent pig game to attract more node visits, so that the visit volume of the base station node is increased with the increase of the visit nodes, when the base station node load reaches a threshold value, the base station unloads part of tasks of the base station node to a cloud or migrates the tasks to other base stations according to priority queues to complete data distribution, the base station load is accelerated and decelerated, and finally the base station node load gradually tends to be stable, as can be seen from fig. 4, under the marine meteorological environments such as sunny days, cloud fog, snow fall, rainfall and the like, the base station bandwidth load rates are respectively stabilized at 0.83, 0.90, 0.91 and 0.97; in the marine meteorological environments such as sunny days, cloud and fog, snowfall, rainfall and the like, the result of the bandwidth load test experiment of the general nodes of the edge computing collaborative service network at sea is shown in fig. 5, the maximum bandwidth load rates are respectively 0.071, 0.10, 0.095 and 0.109, and the bandwidth load rates at steady time are respectively 0.062, 0.063, 0.094 and 0.095;
6-8 show that the task delivery success rate reaches above 0.9 under the influence of different oceanographic environments; the maximum delay of the general nodes of the offshore edge computing cooperative service network is respectively 34ms, 33ms, 32ms and 31ms, and the average delay is respectively 30ms, 32ms, 33ms and 33ms.
8) And (5) analyzing the energy consumption of the algorithm.
8-1) simulating a marine meteorological environment by taking sunny days as constraint conditions, and carrying out comparative analysis on the TCSMEC and AODV and RandomWalk methods. As shown in fig. 8-10, the average remaining energy of the cooperative node is 34%, 8% and 21%, respectively; the link average transmission rates are 0.25MB/s, 0.23MB/s and 0.24MB/s; the average load of the cooperative nodes is respectively 6%, 21% and 15%;
8-2) simulating a marine meteorological environment by using cloud and mist, snowfall, rainfall and the like as constraint conditions, and comparing and analyzing the marine edge calculation credible cooperative task migration method. As shown in fig. 11-12, the average remaining energy of the cooperative nodes is 22%, 18% and 3%, respectively; the average load of the cooperative nodes is respectively 6%, 9% and 10%; the link average transmission rates are 0.25MB/s, 0.25MB/s and 0.25MB/s.
To summarize:
the maritime wireless communication link has wide coverage area and variable environment, and the maritime communication node faces a plurality of resource constraints such as energy, calculation, storage, communication and the like, so that the quality of service is difficult to guarantee when the maritime wireless communication service is implemented. The invention discloses a method for migrating a trusted cooperative task in marine edge computing. The method is based on a collaborative filtering thought, a network heartbeat information piggybacking technology is adopted, a local trust network is constructed by fusing node similarity relation, network behavior, link characteristics and neighbor similarity nodes, recommendation is carried out based on characteristics such as service capability, service quality, load and trust relation between nodes of edge computing service nodes, service quality evaluation information is filtered by a quartile method to improve recommendation effectiveness, adaptive edge computing service access and resource rapid fusion are achieved through queue hierarchical control, simulation experiments show that compared with ADOV and RandomWalk methods, service energy consumption, load efficiency, link transmission rate and the like of TCSMEC are obviously superior, and the TCSMEC can efficiently achieve offshore edge computing credible collaborative task migration service in a complex marine meteorological environment.
Claims (4)
1. A maritime edge computing credible cooperative task migration method is characterized in that a network heartbeat information piggybacking technology is adopted, similar trust relationships among adjacent nodes are fused, and based on behavior and real-time state characteristics of edge computing service nodes, the edge computing service nodes are recommended to implement task migration, and the method comprises the following steps:
1) Supposing that N nodes exist in the offshore edge computing network, the adjacency relation matrix is N ij ＝(a ij ) n×n ，a ij ∈{0,1}，a ij =1 indicates that the node i and the node j are neighbors, and the edge computing service relation matrix is M kl ＝(b kl ) n×n ，b kl ∈{0,1}，b ik =1 indicates that the node i may request the edge computing service from the node k, and every t time interval, the node sends a heartbeat data packet containing its own characteristic information to a peripheral node to maintain connection information, and when receiving heartbeat information from a neighboring node, the marine node q will perform the following operations:
constructing or updating its adjacency vector N ij | i＝q ＝{a q1 ,a q2 ,…,a qn }，
Building or updating its edge computing service relation vector M ik | i＝q ＝{b q1 ,b q2 ,…,b qn }，
Building or updating neighbor node trust relation vector
Constructing or updating edge computing service node service capability evaluation vector G ij | i＝q ＝{g q1 ,g q2 ,…,g qm }，
Wherein, tau qj Satisfies the following formula (1):
in the above formula (1), t s Indicating that the node receives the heartbeat data packet s times, rtt (i, j) indicating the communication delay between the node i and the node j, jac (i, j) indicating the jaccard coefficient of the edge node i and the edge node j, at j Represents the average online duration, t, of node j succ Indicates the number of effective recommendations formed, t all Representing total recommended times, e is a natural base number, and evaluating the service capacity of the node g qk Satisfies formula (2):
w k 、h k 、Rm k respectively representing the bandwidth, the CPU frequency and the memory size of an edge computing service node k;
2) When a neighbor node finishes edge computing service, the neighbor node makes service quality evaluation and adds the service quality evaluation to a heartbeat data packet to inform other nodes, and after receiving the service quality evaluation from the neighbor node, a node q constructs or updates a service quality evaluation matrix S jk ＝(s jk ) n×n ，s jk ∈[0,1]；
3) The node q constructs or updates a priority queue L according to the service capability of the edge computing service node, the service quality evaluation of the neighbor node and the trust relationship;
4) When the node q needs edge computing service, selecting a proper edge computing service node from the priority queue to initiate a task migration request, wherein in order to reduce the search cost, the priority queue L adopts random hash search or round search;
5) After receiving the task migration request response, the node q migrates the task to the edge computing service node, the edge computing service node aggregates service resources according to the task requirement, executes the computing task, and returns the result to the node q;
6) Node q makes evaluation s for the edge calculation service qk According to s qk Updating the neighbor trust relationship of all the neighbors of the recommended edge computing service node k through the formula (1)
Meanwhile, the node q piggybacks evaluation information of the task migration service in a heartbeat data packet to notify a neighbor node, and the neighbor node updates a service quality evaluation matrix S after receiving feedback of the node q jk 。
2. The offshore edge computing trusted cooperative task migration method according to claim 1, wherein the step 3) specifically comprises the following steps:
3-1) the node q calculates the service quality score s of the service node to the edge of any neighbor node according to the trust relationship of the neighbor nodes jk ∈S jk Making a correction, the corrected score u q,jk As in equation (3):
u q,jk ＝b qk τ qj s jk (3)
3-2) compute the service node k for each edge, b qk Revised quality of service score u not equal to 0 q,jk Filtering scores which are too high or too low by adopting a quartile method;
3-3) after the filtering is finished, the node q calculates the priority eta of the edge computing service node k k As in equation (4):
in the above formula (4), Q (k) is the load evaluation function of the edge computing service node k, p k For memory occupancy, ρ k For CPU utilization, d k For bandwidth occupancy, p, rho and d are respectively a memory load threshold, a calculation load threshold and a bandwidth load threshold, and a node q calculates the priority eta of a service node k according to the edge k Eta is to k >The edge calculation service node of alpha is arranged in a first priority queue L 1 Will beta<η k <The edge calculation service node of alpha is listed in a second priority queue L 2 Eta is to k <The edge compute service node of beta is enqueued in the third queue L 3 Where α and β are resolution thresholds, p, ρ, d, α, β ∈ [0,1 ]]，β<α，L 1 、L 2 、L 3 All are circular queues, and before the edge computing service node is added into the queue, whether an offline node (b) exists in the queue is searched qx = 0), if so, the node is replaced, otherwise the end of the queue is added.
3. The offshore edge computing trusted cooperative task migration method according to claim 2, wherein the step 3-2) specifically comprises the following steps:
3-2-1) calculating a revised score u for a serving node k based on each neighbor to edge q,jk Arranging from small to large, taking the total number of scores as O, and taking
Correction score u q,jk Is marked as the lower bound c 1 Get it first
Correction score u q,jk Is marked as an upper bound c 2 Wherein, in the step (A),
represents the largest integer no greater than x;
3-2-2) removal of
The score of (1) is calculated q,jk ∈[c 1 ,c 2 ]Average value of (2)
4. The offshore edge computing trusted cooperative task migration method according to claim 1, wherein the step 4) specifically comprises the following steps:
4-1) if L 1 If the queue is not empty, selecting an edge computing service node to initiate a request by adopting a random hash function, if a response is received, indicating that the edge computing service node receives a task migration request, completing the request, if the response is overtime, marking the node, and sequentially selecting the L where the node is located 1 The predecessor node or successor node of the queue initiates a request, if the response is still overtime, the two nodes are marked, the hash address H is regenerated, the steps are repeated, and if the hash address H is regenerated for times>L 1 The length of the queue is then abandoned from L 1 Selecting a node in the queue to initiate a request;
4-2) if L 2 If the queue is not empty, node q will L 2 The edge computing service nodes of the queue sequentially initiate task migration requests according to the sequence of priority from large to small; if the response is received, the edge computing service node is indicated to receive the task migration request, and the request is completed; if the response is overtime, continuing to search in turn; number of times of searching for wheel>3, if no node responds to the request, abandoning in L 2 Selecting an edge computing service node in the queue to initiate a request;
4-3) node q is at L 3 Queue random hash selection edge computing service node initiates a request, and L 1 The searching process of the queue is similar if the times of regenerating the Hash address H>L 3 And if the queue is long, the task migration request is abandoned.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910803509.2A CN110502328B (en) | 2019-08-28 | 2019-08-28 | Offshore edge computing trusted cooperative task migration method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910803509.2A CN110502328B (en) | 2019-08-28 | 2019-08-28 | Offshore edge computing trusted cooperative task migration method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110502328A CN110502328A (en) | 2019-11-26 |
| CN110502328B true CN110502328B (en) | 2022-11-15 |
Family
ID=68589800
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910803509.2A Active CN110502328B (en) | 2019-08-28 | 2019-08-28 | Offshore edge computing trusted cooperative task migration method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110502328B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111092755B (en) * | 2019-12-05 | 2022-05-03 | 桂林电子科技大学 | Edge service migration simulation method based on resource occupation |
| CN111556457A (en) * | 2020-04-27 | 2020-08-18 | 山东省科学院海洋仪器仪表研究所 | Task migration method and system of underwater self-organizing network based on edge gateway |
| CN111741448B (en) * | 2020-06-21 | 2022-04-29 | 天津理工大学 | Clustering AODV (Ad hoc on-demand distance vector) routing method based on edge computing strategy |
| CN112153605B (en) * | 2020-08-25 | 2023-04-18 | 中国电子科技集团公司电子科学研究院 | Time-frequency resource scheduling method and device for offshore communication system |
| CN112235385B (en) * | 2020-10-09 | 2022-08-30 | 嘉兴学院 | Offshore edge computing low-overhead cloud edge intelligent cooperative ally member discovery method |
| CN113115323B (en) * | 2021-04-15 | 2023-04-07 | 西北工业大学深圳研究院 | Distributed multi-hop ad hoc network secure communication method and device based on unmanned aerial vehicle transmission |
| CN113422803B (en) * | 2021-05-19 | 2022-11-25 | 武汉工程大学 | Seamless migration method for intelligent unmanned aerial vehicle inspection task based on end edge cloud cooperation |
| CN113852931B (en) * | 2021-08-26 | 2024-04-02 | 北京邮电大学 | Offshore wireless communication method and device, storage medium and electronic equipment |
| CN114518948A (en) * | 2022-02-21 | 2022-05-20 | 南京航空航天大学 | Large-scale microservice application-oriented dynamic perception rescheduling method and application |
| CN117150438B (en) * | 2023-10-31 | 2024-02-06 | 成都汉度科技有限公司 | Communication data fusion method and system based on edge calculation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104298557A (en) * | 2014-06-05 | 2015-01-21 | 中国人民解放军信息工程大学 | SOA dynamic load transferring method and system |
| WO2016188116A1 (en) * | 2015-05-25 | 2016-12-01 | 华南理工大学 | Credibility detection-based security routing protocol in vehicular ad hoc network |
| WO2017219476A1 (en) * | 2016-06-22 | 2017-12-28 | 邦彦技术股份有限公司 | Data transmission method and apparatus |
| CN109379739A (en) * | 2018-09-28 | 2019-02-22 | 嘉兴学院 | A kind of credible cooperating service method of sea wireless Mesh netword |
-
2019
- 2019-08-28 CN CN201910803509.2A patent/CN110502328B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104298557A (en) * | 2014-06-05 | 2015-01-21 | 中国人民解放军信息工程大学 | SOA dynamic load transferring method and system |
| WO2016188116A1 (en) * | 2015-05-25 | 2016-12-01 | 华南理工大学 | Credibility detection-based security routing protocol in vehicular ad hoc network |
| WO2017219476A1 (en) * | 2016-06-22 | 2017-12-28 | 邦彦技术股份有限公司 | Data transmission method and apparatus |
| CN109379739A (en) * | 2018-09-28 | 2019-02-22 | 嘉兴学院 | A kind of credible cooperating service method of sea wireless Mesh netword |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110502328A (en) | 2019-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110502328B (en) | Offshore edge computing trusted cooperative task migration method | |
| CN110851429B (en) | Edge computing credible cooperative service method based on influence self-adaptive aggregation | |
| Chen et al. | Task replication for vehicular cloud: Contextual combinatorial bandit with delayed feedback | |
| CN111131421B (en) | Method for interconnection and intercommunication of industrial internet field big data and cloud information | |
| CN110493360A (en) | The mobile edge calculations discharging method of system energy consumption is reduced under multiserver | |
| CN113794494B (en) | Edge computing system and computing unloading optimization method for low-orbit satellite network | |
| CN110753319B (en) | Heterogeneous service-oriented distributed resource allocation method and system in heterogeneous Internet of vehicles | |
| CN109614215A (en) | Stream scheduling method, device, equipment and medium based on deeply study | |
| Munir et al. | Intelligent service fulfillment for software defined networks in smart city | |
| CN112235385B (en) | Offshore edge computing low-overhead cloud edge intelligent cooperative ally member discovery method | |
| Liu | Intelligent routing based on deep reinforcement learning in software-defined data-center networks | |
| CN113391647B (en) | Multi-unmanned aerial vehicle edge computing service deployment and scheduling method and system | |
| CN113422803B (en) | Seamless migration method for intelligent unmanned aerial vehicle inspection task based on end edge cloud cooperation | |
| CN107454009B (en) | Data center-oriented offline scene low-bandwidth overhead traffic scheduling scheme | |
| Xiao et al. | A novel task allocation for maximizing reliability considering fault-tolerant in VANET real time systems | |
| Zhou et al. | Adaptive Routing Strategy Based on Improved Double Q‐Learning for Satellite Internet of Things | |
| CN108833294B (en) | Low-bandwidth-overhead flow scheduling method for data center wide area network | |
| CN114268575A (en) | Self-adaptive three-dimensional transmission method and system in heaven-earth integrated information network | |
| Huang et al. | A hierarchical pseudonyms management approach for software-defined vehicular networks | |
| CN111885113A (en) | Self-adaptive selection and resource allocation method for anchor nodes in social network | |
| Gu et al. | AI-Enhanced Cloud-Edge-Terminal Collaborative Network: Survey, Applications, and Future Directions | |
| CN116133082A (en) | Multi-hop clustering method for improving topology duration of aviation ad hoc network | |
| Yang et al. | Collaborative optimization scheduling of maritime communication for mobile edge architecture | |
| CN110636137B (en) | Cluster-based NFV service chain optimization method in ocean view monitoring network | |
| CN114710200A (en) | Satellite network resource arrangement method and system based on reinforcement 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 |