A collaborative method and system for data analysis tasks

Abstract

The present invention discloses a collaborative method and system for data analysis tasks. Production data collected by a data source is distributed to each client through a multicast distribution tree. The client initiates a data analysis request, which is transmitted in reverse along the multicast distribution tree. If each router responsible for upward transmission has data analysis capabilities, it embeds its own router information in the data analysis request in sequence. The DR router finally receives the data analysis request and then sends a data analysis task arrangement message. The router that receives the data analysis task arrangement message determines whether to execute the data analysis request based on the embedded router information. If so, it executes the data analysis request. The present invention only requires routers in some networks to have data analysis capabilities, realizes distributed execution of data analysis, reduces networking costs, and improves the efficiency of data analysis.

Description

Technical Field

[0001] The present application belongs to the field of data analysis technology, and in particular relates to a collaborative method and system for data analysis tasks. Background Art

[0002] In daily production scenarios, video equipment is increasingly used to collect video data, and production is carried out by analyzing the video data, which has created a large demand for artificial intelligence analysis of production data.

[0003] Current AI analysis for machine vision often relies on centralized processing in the cloud, or collaborative processing between the cloud and the edge. This approach is relatively centralized and insufficiently distributed, failing to meet the demands of large-scale AI analysis. The equipment responsible for AI analysis often has to handle a wide range of tasks, placing high demands on the performance and resulting in high costs. Furthermore, AI analysis tasks take a long time to process, resulting in poor real-time performance. Summary of the Invention

[0004] The purpose of this application is to provide a collaborative method and system for data analysis tasks, which allows some network devices in the network that have data analysis capabilities to collaborate on data analysis, overcome the performance bottleneck problem of centralized processing in the existing technology, and reduce the total network cost as much as possible.

[0005] In order to achieve the above objectives, the technical solutions of this application are as follows:

[0006] A collaborative method for data analysis tasks, applied to a production data multicast network, the production data multicast network including a data source, a router, and a client, the collaborative method for data analysis tasks comprising:

[0007] The production data collected by the data source is distributed to each client through the multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree.

[0008] The gateway router connected to the first client records the data analysis request and then passes it back along the multicast distribution tree. Each router responsible for passing the data back up, if it has data analysis capabilities, embeds its own router information in the data analysis request.

[0009] Before a data analysis request reaches a DR router, if the first router it passes through is currently executing the task corresponding to the data analysis request, the request is stopped and the first router executes the data analysis request. If no router is currently executing the task corresponding to the data analysis request, the DR router eventually receives the data analysis request and then sends a data analysis task scheduling message. The destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address.

[0010] The router that receives the data analysis task arrangement message determines whether to execute the data analysis request itself according to the embedded router information. If so, it executes the data analysis request; otherwise, it passes it downward.

[0011] Furthermore, the gateway router connected to the first client records the data analysis request, including:

[0012] The gateway router connected to the first client records the task corresponding to the data analysis request;

[0013] The collaborative method for data analysis tasks further includes:

[0014] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the transmission of the data analysis request of the second client is stopped, and the gateway router forwards the data analysis result to the second client.

[0015] Furthermore, the gateway router connected to the first client records the data analysis request, including:

[0016] The gateway router connected to the first client records the task corresponding to the data analysis request and the IP address of the first client;

[0017] The gateway router unicasts the data analysis result to the first client;

[0018] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the IP address of the second client is added to the record;

[0019] The gateway router unicasts the data analysis result to the first client and the second client.

[0020] Furthermore, if each router responsible for transmitting data upward has data analysis capabilities, it will embed its own router information in the data analysis request in sequence, including:

[0021] The router information embedded in the data analysis request is arranged in the embedding order to form a router list;

[0022] The determining, based on the embedded router information, whether to execute the data analysis request by itself includes:

[0023] If the router itself is a router with data analysis capability that is closest to the first client in the router list, the data analysis request is executed.

[0024] Furthermore, the collaborative method for data analysis tasks further includes:

[0025] When the second router that is executing the data analysis request needs to stop the data analysis request, it sends a handover query request upward and downward along the multicast distribution tree;

[0026] If the router that receives the handover query request has data analysis capabilities, it will embed its own router information in the handover query request and continue to pass it on;

[0027] The DR router and gateway router that finally receive the handover query request send the router information embedded in the handover query request to the second router. The second router selects a router with data analysis capability to execute the data analysis request based on the received router information.

[0028] This application also proposes a collaborative system for data analysis tasks, which is applied to a production data multicast network. The production data multicast network includes a data source, a router, and a client. The collaborative system for data analysis tasks includes:

[0029] The production data collected by the data source is distributed to each client through the multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree.

[0030] The gateway router connected to the first client records the data analysis request and then passes it back along the multicast distribution tree. Each router responsible for passing the data back up, if it has data analysis capabilities, embeds its own router information in the data analysis request.

[0031] Before a data analysis request reaches a DR router, if the first router it passes through is currently executing the task corresponding to the data analysis request, the request is stopped and the first router executes the data analysis request. If no router is currently executing the task corresponding to the data analysis request, the DR router eventually receives the data analysis request and then sends a data analysis task scheduling message. The destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address.

[0032] The router that receives the data analysis task arrangement message determines whether to execute the data analysis request itself according to the embedded router information. If so, it executes the data analysis request; otherwise, it passes it downward.

[0033] Furthermore, the gateway router connected to the first client records the data analysis request, including:

[0034] The gateway router connected to the first client records the task corresponding to the data analysis request;

[0035] The collaborative method for data analysis tasks further includes:

[0036] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the transmission of the data analysis request of the second client is stopped, and the gateway router forwards the data analysis result to the second client.

[0037] Furthermore, the gateway router connected to the first client records the data analysis request, including:

[0038] The gateway router connected to the first client records the task corresponding to the data analysis request and the IP address of the first client;

[0039] The gateway router unicasts the data analysis result to the first client;

[0040] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the IP address of the second client is added to the record;

[0041] The gateway router unicasts the data analysis result to the first client and the second client.

[0042] Furthermore, if each router responsible for transmitting data upward has data analysis capabilities, it will embed its own router information in the data analysis request in sequence, including:

[0043] The router information embedded in the data analysis request is arranged in the embedding order to form a router list;

[0044] The determining, based on the embedded router information, whether to execute the data analysis request by itself includes:

[0045] If the router itself is a router with data analysis capability that is closest to the first client in the router list, the data analysis request is executed.

[0046] Furthermore, the collaborative system for the data analysis task further includes:

[0047] When the second router that is executing the data analysis request needs to stop the data analysis request, it sends a handover query request upward and downward along the multicast distribution tree;

[0048] If the router that receives the handover query request has data analysis capabilities, it will embed its own router information in the handover query request and continue to pass it on;

[0049] The DR router and gateway router that finally receive the handover query request send the router information embedded in the handover query request to the second router. The second router selects a router with data analysis capability to execute the data analysis request based on the received router information.

[0050] This application proposes a collaborative method and system for data analysis tasks that only requires routers in some networks to have data analysis capabilities, enabling distributed execution of data analysis and reducing networking costs. By collecting and sorting a list of routers with data analysis capabilities, data analysis can be performed locally. Upon encountering a router that is currently executing a data analysis request, the request is stopped, avoiding repeated router selection and accelerating the data analysis process. This invention reduces networking costs, fully utilizes the data analysis capabilities of routers, and improves the efficiency of data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a network diagram for the application scenario of this application.

[0052] Figure 2 Flowchart of the collaborative approach for the data analysis task in this application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0054] In the description of this application, the terms "first", "second", etc. are only used to distinguish the description and do not indicate the order of arrangement, nor can they be understood as indicating or implying relative importance.

[0055] In the description of this application, the terms "include", "comprising", etc. indicate the existence of the described features, integers, steps, operations, elements and / or components, but do not exclude the existence or addition of one or more other features, steps, operations, elements, components and / or their collections.

[0056] Refer to the attached Figure 1 Multiple data sources are deployed at the production site to collect production data. Data sources can be cameras or other sensors. The production data collected by the data sources is sent to various clients via the network. The network consists of network connection devices, namely various routers and switches used for networking. This application uses routers as an example for illustration. Each router has a unique router ID for the entire network and supports multicast. Some routers have built-in data analysis modules that can be used to perform data analysis tasks, such as artificial intelligence analysis, recognition and classification of collected view images, etc. Among these routers are designated routers (DRs), gateway routers (GWs) that connect to clients or data sources, and other routers that connect to these DRs and GWs. All routers constitute the entire network. Production data sent by the data source is sent via multicast, using multicast entries such as (S, G). Clients receive the production data via multicast. The multicast network runs the common PIM SM multicast routing protocol. A multicast distribution tree has been established from the data source to the client. Data analysis requests are initiated by clients. All messages in the network contain router IDs, and each router ID is unique across the entire network and does not conflict.

[0057] In one embodiment, Figure 1 As shown, a collaborative method for data analysis tasks is proposed, which is applied to a production data multicast network. The production data multicast network includes a data source, a router, and a client. The collaborative method for data analysis tasks includes:

[0058] Step S1: Production data collected by a data source is distributed to each client through a multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree.

[0059] In this embodiment, the production data multicast network includes data sources, routers, and clients. The routers in the network support multicast, and some routers have data analysis capabilities. This embodiment does not require all routers in the network to have data analysis capabilities; instead, only some routers have data analysis capabilities, thereby reducing overall network costs and making it easier to implement hybrid networking using some common routers.

[0060] Specifically, a data source IPC collects production data (such as a video stream) and distributes it to client PCs through a multicast distribution tree. The video stream is first sent from the data source to the DR router, and then from the DR router to the client PCs. The multicast distribution tree is often represented by a multicast entry (S, G), where S is the source IP address (specifically, the IP address of the data source) and G is the multicast group IP address.

[0061] When the client PC has a data analysis task, it initiates a data analysis request, which carries the S and G information of the video stream distribution and the corresponding data analysis task, and then transmits it back to the data source along the (S, G) path, that is, along the multicast distribution tree. Each multicast router sends the message from the input interface of the (S, G) table entry to the upstream router, and each multicast router on the multicast distribution tree can receive it. The forwarding line between the multicast sender and the receiver constitutes the multicast distribution tree, which is the (S, G) path in this embodiment. The multicast distribution tree contains DR routers and other multicast routers. The multicast distribution tree is a relatively mature technology in this field and will not be described in detail here.

[0062] Step S2: The gateway router connected to the first client records the data analysis request and then transmits it backward along the multicast distribution tree. If each router responsible for transmitting upward has data analysis capabilities, it will embed its own router information in the data analysis request in turn.

[0063] For example, client PC1 initiates a data analysis request carrying data analysis task A. The "data analysis request" message contains two pieces of information (S, G) corresponding to the multicast video stream, where S represents the IP address of the video source and G represents the multicast group address of the video.

[0064] The "data analysis request" message will first reach the gateway router GW1, and the gateway router GW1 will record the "data analysis task A", indicating that the gateway router GW1 will forward the data analysis results of the "data analysis task A". Then the "data analysis request" message will be transmitted back to the data source along the (S, G) path.

[0065] Each router responsible for upstream transmission extracts the data source IP address S and multicast group address G from the "Data Analysis Request" and matches S and G to the (S, G) forwarding table entry on the router. The router then discovers the inbound interface and upstream PIM neighbor router from the (S, G) forwarding table entry and then sends the "Data Analysis Request" out the inbound interface, with the destination IP address being the upstream PIM neighbor router.

[0066] At the same time, each router responsible for forwarding data, if capable of data analysis, will sequentially embed its own router information, which can be either a router ID or an IP address, in the "Data Analysis Request" message. This embedded router information forms a list of routers capable of data analysis. The order of the routers in the list represents their position on the path from the client to the DR router. Obviously, the router closest to client PC1 with data analysis capabilities is embedded first, and the router closest to the data source with data analysis capabilities is embedded last.

[0067] In this way, this embodiment collects router information in the network that can be used for data analysis.

[0068] Step S3: Before the data analysis request reaches the DR router, if the first router it passes through is executing the task corresponding to the data analysis request, the transmission is stopped and the data analysis request is executed by the first router; if there is no router that is executing the task corresponding to the data analysis request in the middle, the DR router finally receives the data analysis request and then sends a data analysis task scheduling message, the destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address.

[0069] In this step, the data analysis request is transmitted reversely along the multicast distribution tree. If each router it passes through has data analysis capabilities, it will also check whether it is executing "data analysis task A". If it is executing "data analysis task A", the router will be called the first router, and the router will perform data analysis and return the data analysis results to the client.

[0070] If no router along the way is currently executing "Data Analysis Task A," the "Data Analysis Request" message continues until it reaches the DR router connected to the data source. The DR router extracts the embedded router information from the message, forming a list of routers capable of data analysis. The order of the routers in the list represents their position on the path from the client to the DR router.

[0071] The DR router sends a "Data Analysis Task Scheduling Message" with the destination IP address being multicast group G and the source IP address being video source S. The message is forwarded from the DR router to the client PC. The "Data Analysis Task Scheduling Message" includes a list of routers.

[0072] Step S4: The router that receives the data analysis task arrangement message determines whether to execute the data analysis request itself based on the embedded router information. If so, it executes the data analysis request; otherwise, it passes it downward.

[0073] During the process of data analysis task arrangement message transmission, the router that receives the data analysis task arrangement message in the middle will determine whether it is the router closest to the client based on the router list. If it is, it will execute the data analysis request, otherwise it will pass it down.

[0074] Therefore, the message will eventually be delivered to the router closest to the client in the router list, and this router will be responsible for executing "data analysis task A". It will not be passed down further because there will be no routers with data analysis capabilities downstream.

[0075] In the above embodiment, two scenarios are described. The first scenario is that a client sends a data analysis request. If, during transmission, the data analysis request encounters a router that is currently executing the task corresponding to the data analysis request, the router directly executes the data analysis request and provides feedback on the analysis results. The second scenario is that a client sends a data analysis request. If, during transmission, the data analysis request does not encounter a router that is currently executing the task corresponding to the data analysis request, the data analysis request reaches a DR router, which sends a data analysis task scheduling message. Ultimately, the data analysis request is executed by the router closest to the client that has data analysis capabilities and provides feedback on the analysis results.

[0076] For example, client PC1 needs to perform data analysis and sends a data analysis request, carrying data analysis task A. The data analysis request is transmitted in the reverse direction along the (S, G) path. If the (S, G) path of production data is: DR-R1-R2-R3-R4-R5-GW1, then the reverse path of the data analysis request is: GW1-R5-R4-R3-R2-R1-DR.

[0077] Assume that only routers R2 and R4 have data analysis capabilities, and neither is executing data analysis task A. When passing through GW1, GW1 records data analysis task A and sends it to router R5. Since R5 does not have data analysis capabilities, the request is passed upward to router R4. R4, which has data analysis capabilities, embeds R4's ID (i.e., R4) into the data analysis request and then passes it to router R3. Since R3 does not have data analysis capabilities, the request is passed upward to router R2. R2, which has data analysis capabilities, embeds R2's ID (i.e., R2) into the data analysis request and then passes it to router R1. Since R1 does not have data analysis capabilities, the request is passed upward to the DR router. At this point, the list of routers in the data analysis request is R2-R4, with R4 being the router closest to client PC1 that has data analysis capabilities.

[0078] The DR router generates a data analysis task scheduling message, carrying the aforementioned router list, and passes it down the (S, G) path. When it reaches R1 or R3, since neither is in the router list, it's passed directly down. When it reaches R2, R2 has data analysis capabilities, but since R2 isn't closest to client PC1, it's passed directly down. When it reaches R4, R4 has data analysis capabilities and is closest to client PC1, so R4 executes data analysis task A.

[0079] For another example, suppose only routers R2 and R4 have data analysis capabilities, and R2 is executing data analysis task A. When the data analysis request reaches R4, R4 is embedded in the request and passed onward. When it reaches R2, it detects that R2 is already executing data analysis task A. The request is no longer passed onward, but instead executes data analysis task A on R2 and reports the analysis results.

[0080] The above method takes into account local analysis, distribution of analysis pressure and avoidance of duplicate arrangements, fully utilizes the data analysis capabilities of all routers with data analysis capabilities, reduces the overhead of data transmission in the network, and can quickly locate the router performing task analysis.

[0081] In another embodiment, the gateway router connected to the first client records the data analysis request, including:

[0082] The gateway router connected to the first client records the task corresponding to the data analysis request;

[0083] The collaborative method for data analysis tasks further includes:

[0084] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the transmission of the data analysis request of the second client is stopped, and the gateway router forwards the data analysis result to the second client.

[0085] For example, if client PC1 initiates a data analysis request for Data Analysis Task A, it first reaches gateway router GW1. If GW1 doesn't have "Data Analysis Task A" logged, the request will be forwarded back to the data source along the (S, G) path. Using the aforementioned method, it will find the router executing Data Analysis Task A, such as R4. R4 will then perform the data analysis and return the results to client PC1.

[0086] When client PC2 also initiates a data analysis request corresponding to data analysis task A, the data analysis request first reaches the gateway router GW1. Since "data analysis task A" is recorded on the gateway router GW1, the transmission of the data analysis request is stopped, and R4 directly performs data analysis and feeds back the data analysis results to client PC2.

[0087] In this embodiment, when the gateway router does not support data analysis, it will also be recorded on the gateway router connected to the first client. When the gateway router with the data analysis task recorded receives the same task request from the second client, it only needs to send a copy of the data analysis result to the second client, avoiding repeated message transmission and analysis router selection operations, and speeding up the implementation of data analysis.

[0088] In another embodiment, the gateway router connected to the first client records the data analysis request, including:

[0089] The gateway router connected to the first client records the task corresponding to the data analysis request and the IP address of the first client;

[0090] The gateway router unicasts the data analysis result to the first client;

[0091] When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the IP address of the second client is added to the record;

[0092] The gateway router unicasts the data analysis result to the first client and the second client.

[0093] In this embodiment, when the gateway router records a data analysis task, it also records the client's IP address. This allows the gateway router to unicast the data analysis results to that client, avoiding interference with other clients connected to the same gateway router. When a second client requests the same task, the second client's IP address is added to the task name, and a copy of the analysis results is sent to both clients.

[0094] In another embodiment, the collaborative method for data analysis tasks further includes:

[0095] When the second router that is executing the data analysis request needs to stop the data analysis request, it sends a handover query request upward and downward along the multicast distribution tree;

[0096] If the router that receives the handover query request has data analysis capabilities, it will embed its own router information in the handover query request and continue to pass it on;

[0097] The DR router and gateway router that finally receive the handover query request send the router information embedded in the handover query request to the second router. The second router selects a router with data analysis capabilities to execute the data analysis request based on the received router information.

[0098] Specifically, if a router R4 that is executing a data analysis request wants to suspend data analysis, it sends a "handover query" message to the upstream and downstream. The destination IP of the handover query sent upstream is S, and the source is G. The message contains the information of the router that suspends the analysis and is sent from the input interface of the local (SG) table entry. After receiving it, if the routers along the way support data analysis, they insert their own router information and then send it from the input interface of their own (SG) table entry until the DR router. The DR router extracts the list of routers that support data analysis from the handover query message and sends it to the router R4 that suspends data analysis according to the router information of suspending data analysis in the message. The destination IP of the handover query message sent downstream is G, and the source is S. It is embedded with the local IP and is sent from the output interface of (SG). Routers that support data analysis along the way embed their own router information and finally reach the gateway router. The gateway router sends the router list to the router R4 that suspends data analysis.

[0099] Router R4, which has paused analysis, first checks whether the Router List in the Handover Reply message from the DR router is not empty. If not, it hands over the data analysis task to the nearest upstream router, such as R2. Otherwise, it randomly selects a message with a non-empty Router List from multiple "Handover Reply" messages from the gateway router and selects the nearest downstream router for handover.

[0100] This embodiment realizes the handover of the data analysis router, so that the client does not need to re-initiate the data analysis request and repeat the process of searching and selecting a router with data analysis capabilities, which can realize the handover of data analysis more quickly.

[0101] Another embodiment of the present application provides a collaborative system for data analysis tasks, which is applied to a production data multicast network. The production data multicast network includes a data source, a router, and a client. The collaborative system for data analysis tasks includes:

[0102] The production data collected by the data source is distributed to each client through the multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree.

[0103] The gateway router connected to the first client records the data analysis request and then passes it back along the multicast distribution tree. Each router responsible for passing the data back up, if it has data analysis capabilities, embeds its own router information in the data analysis request.

[0104] Before a data analysis request reaches a DR router, if the first router it passes through is currently executing the task corresponding to the data analysis request, the request is stopped and the first router executes the data analysis request. If no router is currently executing the task corresponding to the data analysis request, the DR router eventually receives the data analysis request and then sends a data analysis task scheduling message. The destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address.

[0105] The router that receives the data analysis task arrangement message determines whether to execute the data analysis request itself according to the embedded router information. If so, it executes the data analysis request; otherwise, it passes it downward.

[0106] The specific implementation scheme of the collaborative system for the above-mentioned data analysis task has been described in detail in the previous embodiment of the collaborative method for data analysis task, and will not be repeated here.

[0107] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A collaborative method for data analysis tasks, applied to a production data multicast network, wherein the production data multicast network includes a data source, a router, and a client, characterized in that: The collaborative method for the data analysis task includes: The production data collected by the data source is distributed to each client through the multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree. The gateway router connected to the first client records the data analysis request and then passes it back along the multicast distribution tree. Each router responsible for passing the data back up, if it has data analysis capabilities, embeds its own router information in the data analysis request. Before a data analysis request reaches a DR router, if the first router it passes through is currently executing the task corresponding to the data analysis request, the request is stopped and the first router executes the data analysis request. If no router is currently executing the task corresponding to the data analysis request, the DR router eventually receives the data analysis request and then sends a data analysis task scheduling message. The destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address. The router that receives the data analysis task arrangement message determines whether to execute the data analysis request by itself according to the embedded router information. If so, it executes the data analysis request; otherwise, it passes it down. Wherein, if each router responsible for transmitting upward has data analysis capability, it will embed its own router information in the data analysis request in sequence, including: The router information embedded in the data analysis request is arranged in the embedding order to form a router list; The determining, based on the embedded router information, whether to execute the data analysis request by itself includes: If the router itself is a router with data analysis capability that is closest to the first client in the router list, the data analysis request is executed.

2. The collaborative method for data analysis tasks according to claim 1, characterized in that: The gateway router connected to the first client records the data analysis request, including: The gateway router connected to the first client records the task corresponding to the data analysis request; The collaborative method for data analysis tasks further includes: When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the transmission of the data analysis request of the second client is stopped, and the gateway router forwards the data analysis result to the second client.

3. The collaborative method for data analysis tasks according to claim 1, characterized in that: The gateway router connected to the first client records the data analysis request, including: The gateway router connected to the first client records the task corresponding to the data analysis request and the IP address of the first client; The gateway router unicasts the data analysis result to the first client; When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the IP address of the second client is added to the record; The gateway router unicasts the data analysis result to the first client and the second client.

4. The collaborative method for data analysis tasks according to claim 1, characterized in that: The collaborative method for data analysis tasks further includes: When the second router that is executing the data analysis request needs to stop the data analysis request, it sends a handover query request upward and downward along the multicast distribution tree; If the router that receives the handover query request has data analysis capabilities, it will embed its own router information in the handover query request and continue to pass it on; The DR router and gateway router that finally receive the handover query request send the router information embedded in the handover query request to the second router. The second router selects a router with data analysis capability to execute the data analysis request based on the received router information.

5. A collaborative system for data analysis tasks, applied to a production data multicast network, the production data multicast network comprising a data source, a router, and a client, characterized in that: The collaborative system for the data analysis task includes: The production data collected by the data source is distributed to each client through the multicast distribution tree. The first client that needs to perform data analysis initiates a data analysis request, which carries the source IP address and multicast group IP address of the multicast distribution tree. The gateway router connected to the first client records the data analysis request and then passes it back along the multicast distribution tree. Each router responsible for passing the data back up, if it has data analysis capabilities, embeds its own router information in the data analysis request. Before a data analysis request reaches a DR router, if the first router it passes through is currently executing the task corresponding to the data analysis request, the request is stopped and the first router executes the data analysis request. If no router is currently executing the task corresponding to the data analysis request, the DR router eventually receives the data analysis request and then sends a data analysis task scheduling message. The destination IP address of the data analysis task scheduling message is the multicast group IP address, and the source IP address is the data source IP address. The router that receives the data analysis task arrangement message determines whether to execute the data analysis request by itself according to the embedded router information. If so, it executes the data analysis request; otherwise, it passes it down. Wherein, if each router responsible for transmitting upward has data analysis capability, it will embed its own router information in the data analysis request in sequence, including: The router information embedded in the data analysis request is arranged in the embedding order to form a router list; The determining, based on the embedded router information, whether to execute the data analysis request by itself includes: If the router itself is a router with data analysis capability that is closest to the first client in the router list, the data analysis request is executed.

6. The collaborative system for data analysis tasks according to claim 5, characterized in that: The gateway router connected to the first client records the data analysis request, including: The gateway router connected to the first client records the task corresponding to the data analysis request; The collaborative method for data analysis tasks further includes: When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the transmission of the data analysis request of the second client is stopped, and the gateway router forwards the data analysis result to the second client.

7. The collaborative system for data analysis tasks according to claim 5, characterized in that: The gateway router connected to the first client records the data analysis request, including: The gateway router connected to the first client records the task corresponding to the data analysis request and the IP address of the first client; The gateway router unicasts the data analysis result to the first client; When the gateway router connected to the first client receives the data analysis request of the second client, if it is found that the task corresponding to the data analysis request of the second client has been recorded on the gateway router connected to the first client, the IP address of the second client is added to the record; The gateway router unicasts the data analysis result to the first client and the second client.

8. The collaborative system for data analysis tasks according to claim 5, characterized in that: The collaborative system for the data analysis task further includes: When the second router that is executing the data analysis request needs to stop the data analysis request, it sends a handover query request upward and downward along the multicast distribution tree; If the router that receives the handover query request has data analysis capabilities, it will embed its own router information in the handover query request and continue to pass it on; The DR router and gateway router that ultimately receive the handover query request send the router information embedded in the handover query request to the second router. The second router selects a router with data analysis capabilities based on the received router information to execute the data analysis request.

Images