CN114721828A - Distributed reconstruction resource management method based on cloud equipment - Google Patents

Abstract

The invention relates to a distributed reconstruction resource management method based on cloud equipment, which is applied to a distributed reconstruction system for reconstructing medical images in medical image equipment, wherein a plurality of nodes for realizing data reconstruction programs are configured in the distributed reconstruction system, all the nodes are positioned in a broadcast domain of the distributed reconstruction system, and available resource information is periodically broadcast in a broadcast mode, and the method comprises the following steps: the first node and each node in the broadcast domain update respective available resource information in real time; if the first node in the broadcast domain receives the local reconstruction task, performing data segmentation on the local reconstruction task and creating a thread in segmented data according to the resource allocation strategy, the available resource information of the current first node and the available resource information of other nodes in the broadcast domain so as to complete the local reconstruction task in a distributed manner. The method improves the data processing speed and solves the problem of unbalanced computing resources.

Description

A distributed reconstruction resource management method based on cloud equipment

technical field

The present invention relates to the technical field of distributed reconstruction, and in particular, to a method for managing distributed reconstruction resources based on cloud devices.

Background technique

PET-CT is a nuclear medicine imaging device that combines PET and CT. Among them, PET (Positron Emission Tomography, positron emission tomography) is responsible for collecting PET sequences with functional imaging function; CT (X-rayComputed Tomography, X-ray tomography) is responsible for collecting CT sequences with structural imaging function. After the PET data collection is completed, the algorithm module uses the PET raw data for reconstruction to generate a PET image. In the reconstruction process, CT images are also used to perform attenuation correction for PET image reconstruction.

The raw data obtained during PET scanning will be transmitted to the PET reconstruction workstation in real time, which is responsible for storing the PET raw data and various system files required for reconstruction, as well as reconstructing the PET raw data into images. Take a more common hardware configuration as an example, such as IntelXeonW series 12-core CPU, 32GB memory, under this hardware configuration, usually the amount of data obtained by torso scanning in 2 minutes, using OSEM (Ordered Subset Maximum Expected Value Method) + TOF ( time of flight) reconstruction method, 2 iterations, 10 subsets, reconstruction takes ~2 min to complete. Moreover, due to the limited hardware resources of the PET reconstruction workstation, only one reconstruction can be processed at the same time. Once the scan contains multiple reconstructions, or the user needs to perform multiple post-reconstructions on the existing data during scanning, the reconstruction that cannot be started can only be Entering the queue to wait, this kind of waiting often lasts for a long time. In recent years, with the continuous improvement of reconstruction algorithms, various advanced algorithms have gradually emerged, such as image optimization algorithms using deep learning technology. These advanced algorithms, due to the complexity of the internal calculation, need to consume more time under the same hardware configuration.

In order to shorten the reconstruction time and improve the reconstruction efficiency, how to use cloud computing to realize remote reconstruction has become a technical problem that needs to be solved urgently.

SUMMARY OF THE INVENTION

(1) Technical problems to be solved

In view of the above shortcomings and deficiencies of the prior art, the present invention provides a distributed reconstruction resource management method based on cloud equipment, which is used to shorten reconstruction time and improve reconstruction efficiency in cloud reconstruction, and solve the technical problem of unbalanced computing resources.

(2) Technical solutions

In order to achieve the above-mentioned purpose, the main technical scheme adopted in the present invention includes:

In a first aspect, an embodiment of the present invention provides a distributed reconstruction resource management method based on a cloud device. The distributed reconstruction resource management method is applied to a distributed reconstruction system for medical image reconstruction in medical imaging equipment. The distributed reconstruction system is configured with a plurality of nodes that implement data reconstruction procedures, all nodes are located in the broadcast domain of a distributed reconstruction system, and the available resource information is broadcast periodically by broadcasting, and the method includes:

S10, the first node periodically acquires the available resource information in the first node and broadcasts it to other nodes in the broadcast domain, and receives the available resource information broadcast by each node in the broadcast domain;

S20, the first node judges whether the local reconstruction task sent by the imaging device connected to the first node is received;

S30. If the first node receives the local reconstruction task, according to the resource allocation policy, the available resource information of the current first node, and the available resource information of other nodes in the broadcast domain, perform data segmentation on the local reconstruction task and create a partitioned post-data thread to complete the local reconstruction task in a distributed manner;

The resource allocation strategy is used to determine one or more nodes capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and to assign all nodes according to the selected node and the processing efficiency of the node and the available resource information of the node. segmentation based on the local reconstruction task described above.

Optionally, the method further includes:

After all nodes processing the local reconstruction task complete the processing of the data allocated to them, update the available resource information inside the node and broadcast it;

Wherein, when the first node determines more than one node capable of processing the local reconstruction task, the priority of the first node is the highest priority.

Optionally, each node maintains a data structure of available resource information of each node in the broadcast domain and a node efficiency information table of each node;

The information in the node efficiency information table of each node stored in the first node is the processing efficiency of each node calculated by the first node based on the data packets processed by each node within a specified time period.

Optionally, the SOCKET communication method is used for communication between any two nodes;

The available resource information includes one or more of the following:

The total number of CPU cores, the number of currently available cores; the memory information, free memory information, response speed, and supported algorithms of the node.

In a second aspect, an embodiment of the present invention further provides a server, where the server belongs to a node of a distributed reconstruction system for medical image reconstruction in a medical imaging device, the server is located in a broadcast domain, and broadcasts to Other nodes in the distributed reconstruction system broadcast available resource information, and the server includes:

The distributed computing processing module is used to communicate with other nodes in the distributed reconstruction system and receive local reconstruction tasks transmitted by the video equipment connected to the server, as well as other nodes in the broadcast domain based on the resource allocation strategy and the available resource information of the current node. the available resource information, perform data segmentation on the local reconstruction task and create threads and/or processes in the segmented data;

A reconstruction module, configured to call the thread and/or process of at least one node determined by the distributed computing processing module according to the distributed service interface in the distributed reconstruction system to process the divided local reconstruction task to complete the reconstruction;

The resource allocation strategy is used to determine one or more nodes capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and to assign all nodes according to the selected node and the processing efficiency of the node and the available resource information of the node. segmentation based on the local reconstruction task described above.

Optionally, the distributed computing processing module includes:

a network communication unit for communicating with other nodes in the distributed reconstruction system;

The computing resource management unit is used to periodically obtain the available resource information of other nodes in the distributed reconstruction system, and maintain the data structure of the available resource information of each node in the broadcast domain and the node efficiency information table of each node stored in the server;

A computing resource dynamic utilization unit, configured to determine one or more nodes capable of processing the local reconstruction task based on the resource allocation strategy, available resource information of the current node, and available resource information of other nodes in the broadcast domain;

The data splitting and merging unit is used for splitting the local reconstruction task according to all the selected nodes, the processing efficiency and available resource information of each node in the all nodes, and distributes the local reconstruction tasks to the selected nodes for reconstruction processing.

Optionally, the network communication unit is used to communicate in a SOCKET communication mode; or, communicate in an HTTPSOCKET communication mode;

The computing resource management unit is used to send the available resource information and reconstruction task information of the node to the broadcast domain in the form of broadcast; and receive broadcast packets sent by other nodes in the broadcast domain;

The reconstruction task information includes information of local reconstruction tasks or information of reconstruction tasks distributed by other nodes;

The computing resource dynamic utilization unit is used to obtain the available resource information that can be used by this node after receiving the local reconstruction task and determining the reconstruction algorithm; and the available computing resource information of other nodes in the broadcast domain, and according to the local reconstruction concurrency limit and Intra-broadcast domain reconstruction concurrency limit determines the available resource information of the local server and the available resource information of other nodes in the broadcast domain, selects all nodes processing local reconstruction tasks, and creates multiple threads and/or processes for processing local reconstruction tasks;

The data splitting and merging unit is used to receive the threads and/or processes sent by the computing resource dynamic utilization unit, the available resource information of all the selected nodes, and the processing efficiency of each node in all the selected nodes, and to split and merge the local reconstruction task. Determine the size of data allocated to each node selected for distribution;

The threads and/or processes that process the segmented data run partly on the local node and partly on other nodes in the broadcast domain.

Optionally, also include:

The computing resource dynamic utilization unit creates a maximum number of threads for processing local reconstruction tasks based on available resource information of all selected nodes.

In a third aspect, an embodiment of the present invention further provides a distributed reconstruction system based on a cloud device, including a plurality of servers described in any one of the second aspect, each server serving as a node to form a decentralized broadcast domain, and each server Corresponds to more than one imaging equipment located in the hospital.

(3) Beneficial effects

In the method of the embodiment of the present invention, each node in the broadcast domain broadcasts the available resource information in real time or periodically, so that when a node has a local reconstruction task, it can be jointly processed by more than one node, which effectively improves the local reconstruction task. At the same time, the node that receives the local reconstruction task is selected for processing, and the resources are distributed reasonably, and the idle resources in the distributed reconstruction system are rationally utilized.

In addition, in the embodiment of the present invention, the service of each node in the broadcast domain grasps the dynamic computing resource information of all nodes in the broadcast domain. At this time, the domain composed of all nodes is decentralized, and the nodes in the domain can be flexible Dynamically increase and decrease without affecting the functionality of the overall distributed computing.

Description of drawings

1 is a schematic flowchart of a method for managing distributed reconstruction resources based on a cloud device according to an embodiment of the present invention;

2 is a schematic flowchart of a cloud device-based distributed reconstruction resource management method provided by another embodiment of the present invention;

3 is a schematic structural diagram of a cloud device-based distributed reconstruction system provided by an embodiment of the present invention;

4 is a schematic diagram of a time slice used by a node in a distributed reconstruction system according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of dividing a task to be reconstructed into data packets according to an embodiment of the present invention.

Detailed ways

In order to better explain the present invention and facilitate understanding, the present invention will be described in detail below with reference to the accompanying drawings and through specific embodiments.

With the development of cloud computing for many years and the gradual popularization of the new generation of communication technologies represented by 5G, the use of powerful cloud computing resources and communication networks to achieve remote reconstruction has become an achievable technical solution.

That is to say, many medical imaging equipment such as PET-CT and PET will no longer be configured with local workstations, but will use proprietary workstations in the cloud. The image data are all done using cloud computing. At this time, the problem of unbalanced utilization of computing resources is brought about.

For example, if there is Company A, which sells and installs 100 devices, there will be 100 cloud workstations correspondingly. Since these 100 devices are distributed in different hospitals in different regions, the number of patients and the busy time are different. During a certain period of time, among the 100 cloud workstations, some nodes (ie cloud workstations) must be busy and some nodes are idle. Even some nodes, due to the small number of patients, are idle most of the time. In this case, the computing resources of idle nodes can be fully utilized to share some computing tasks for busy nodes, improve data processing speed, return results to imaging devices faster, and improve data processing performance.

The names involved in the embodiments of the present invention are described as follows:

Node: A service in the cloud is used for data reconstruction of imaging equipment, that is, a cloud computing workstation for image reconstruction calculations. All such workstations in the embodiments of the present invention are referred to as nodes (ie, cloud devices).

Broadcast number: It is a unique identifier unique to each node. The service running on each node for processing image reconstruction tasks uses this number to identify itself, and is used to be identified by services of other nodes.

Broadcast domain: A broadcast domain is a collection of a group or a class of broadcast numbers. The service of each node can use broadcast to send information to one or more broadcast domains.

Local rebuild concurrency limit: is a set of values that specifies the maximum number of CPU cores that the service can utilize locally on the node and the corresponding maximum number of concurrency.

Intra-domain reconstruction concurrency limit: It is a set of values used to specify the maximum number of CPU cores and the corresponding maximum number of concurrency that can be used by the service in its broadcast domain in addition to the local node.

Service: a program running mode, the functions described in this embodiment run on the nodes in the form of services. The services mentioned in the following embodiments may all refer to functional categories running on nodes in the form of services.

Time slice: It is the time interval for adjusting the assignment of node tasks when performing reconstruction task planning, as shown in Figure 5. Within a time slice, the usage of a specific node (the number of CPU cores occupied, the size of the data sent) remains unchanged. Since the times when each node starts and ends the reconstruction task are not synchronized, the time slice in the embodiment of the present invention is planned independently for each node. If the time slice is too short, the thread task will not have enough time to complete, and the algorithm consumes a lot of time on the dynamic programming task. If the time slice is too long, node resources may be occupied by non-local reconstruction tasks for a long time, resulting in limited use of local node computing resources by local tasks. In the embodiment of the present invention, the time slice information of each node is determined based on the node processing efficiency of each node.

Example 1

As shown in FIG. 1 and FIG. 2 , an embodiment of the present invention provides a distributed reconstruction resource management method based on a cloud device. In this embodiment, the distributed reconstruction resource management method is applied to medical imaging equipment (such as PET-CT, A distributed reconstruction system for medical image reconstruction in PET, CT), the distributed reconstruction system is configured with a plurality of nodes (ie cloud devices) that implement data reconstruction procedures, and all nodes are located in the broadcast domain of a distributed reconstruction system and broadcast the available resource information periodically in a broadcast manner, the method includes:

S10. The first node periodically acquires available resource information in the first node and broadcasts it to other nodes in the broadcast domain, and receives available resource information broadcast by each node in the broadcast domain.

Specifically, the service on each node may broadcast available resource information to the broadcast domain according to a predetermined time interval (for example, 1s/2s/3s, etc.). The available resource information may include: computing resource information of the node, such as the number of cores and load of the CPU, and memory occupation.

It should be noted that the first node in the above steps will also receive available resource information broadcast by other nodes in the broadcast domain. By broadcasting available resource information, the service of each node in the broadcast domain can obtain the dynamic computing resource information of all nodes in the broadcast domain in real time. As a result, all nodes in the broadcast domain are decentralized, and nodes in the broadcast domain can be dynamically increased and decreased without affecting the overall distributed computing function.

S20. The first node periodically checks whether a local reconstruction task is received.

Generally, each node in the broadcast domain corresponds to one or more imaging equipment installed locally in the hospital, which is used for receiving data uploaded by the imaging equipment for realizing reconstruction. The service in each node is used to reconstruct the received data. In this embodiment, the service in the node realizes the reconstruction of a received task to be reconstructed based on the resources in the broadcast domain.

S30. If the first node receives the local reconstruction task, according to the resource allocation policy, the available resource information of the current first node, and the available resource information of other nodes in the broadcast domain, perform data segmentation on the local reconstruction task and create a partitioned post-data thread/or process to complete the local reconstruction task in a distributed manner;

The resource allocation strategy is used to determine one or more nodes capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and according to the selected node, the processing efficiency of the node, and the available resource information of the node. Local reconstruction task for segmentation.

When a set of data, its reconstruction method, and reconstruction parameters are waiting to be reconstructed from the hospital's local imaging device to the corresponding node, the service will send the data locally to the node and other nodes in the broadcast domain according to the local reconstruction concurrency limit and the intra-domain reconstruction concurrency limit. Get information about the available resources for this rebuild task. If all available resource information in the broadcast domain cannot reach the upper limit set by the intra-domain reconstruction concurrency limit, obtain the maximum computing resources (that is, available resource information) that can be provided in the broadcast domain; A service within the first node of the reconstruction task (may be referred to as a raw data reconstruction service).

For example, if the first node is providing computing resources to other nodes in the broadcast domain and the upper limit of the local reconstruction concurrency limit is not reached, the largest computing resource in the first node is obtained. That is, the computing resources in the first node are preferentially provided to the local reconstruction task received by the first node. If all computing resources in the broadcast domain cannot reach the upper limit set by the intra-domain reconstruction concurrency limit, obtain the maximum available resource information that can be provided in the broadcast domain, and provide the available resource information to the original data reconstruction service in the first node.

At this time, the original data reconstruction service in the first node creates threads and/or processes according to the obtained available resource information, and processes the original data to be reconstructed in parallel.

In this embodiment, the maximum number of threads that can be created by the original data reconstruction service matches the local reconstruction concurrency limit plus the intra-domain reconstruction concurrency limit. The maximum number of threads can be determined by the available resource information in the broadcast domain that can be obtained by the original data re-addition service in the first node.

Based on the above description, it can be understood that the method of this embodiment may further include the following step S40 not shown in the figure:

S40: After the selected node completes the processing of the local reconstruction task, update the available resource information inside the node and broadcast it;

Wherein, when the first node determines more than one node capable of processing the local reconstruction task, the priority of the first node is the highest priority.

In this embodiment, each node comprehensively stores the available resource information of each node in the broadcast domain and the node processing efficiency; the node processing efficiency refers to the amount of data processed by one thread in one time slice. That is, the original data reconstruction service, based on the obtained available resource information, creates threads and processes the data to be reconstructed in parallel. The maximum number of threads that the raw data reconstruction service can create.

In this embodiment, the SOCKET communication method is used for communication between any two nodes in the broadcast domain; or, the HTTPSOCKET communication method is used for communication between any two nodes;

The available resource information (may be referred to as computing resource/computing resource information) includes one or more of the following: the total number of CPU cores, the number of currently available cores; the memory information of the node, free memory information, response speed, supported algorithm.

It is understandable that, in the process of reconstructing data by the service in each node, the service also needs to determine the division granularity of the data to be reconstructed (that is, the original data of the local reconstruction task received by the first node). The granularity here refers to the length of each data segment after the original data is divided.

In this embodiment, the segmentation algorithm of the first node may determine the segmentation granularity of the local reconstruction task based on segmentation, merging efficiency, and single-task processing time. If the granularity is too small, the services in each node will consume a lot of resources on raw data segmentation and calculation results integration. If the granularity is too large, the processing time of a single data will be too long, so that the computing resources of the nodes in the broadcast domain are occupied by other nodes for a long time, and the local computing resources cannot be obtained in time when the local reconstruction task occurs on the node.

The reconstruction resource management service in this embodiment can be used as a middleware and exists between the post-rebuild program and the operating system. For the post-reconstruction procedure, it is not clear whether the threads executing in parallel are executed on the CPU of the local node or the CPU of other nodes in other cloud devices. The service integrates cloud nodes through a unique distributed computing algorithm, makes full use of the computing resources of idle nodes, and improves post-reconstruction speed.

Embodiment 2

In this embodiment, when the cloud computing workstation is used to perform the local PET reconstruction task/PET-CT reconstruction task in the hospital, the computing resources of the cloud workstation are fully utilized to achieve load balance and improve the reconstruction speed.

The server in this embodiment, that is, the cloud device, belongs to a node in the distributed reconstruction system for medical image reconstruction in the medical imaging device. The service (that is, the program) in the node is used to realize the reconstruction of PET-CT data, and the server is located in a In the broadcast domain, the available resource information is broadcast to other nodes in the distributed reconstruction system in a broadcast manner. All nodes constitute the distributed reconstruction system in this embodiment.

With reference to FIG. 3 , the server in this embodiment may include: a distributed computing processing module A10 and a reconstruction module A20.

Specifically, the distributed computing processing module A10 is used to communicate with other nodes in the distributed reconstruction system, and periodically check the local reconstruction tasks, and broadcast the available resources of other nodes in the domain based on the resource allocation strategy, the available resource information of the current node, and the information, perform data segmentation on the local reconstruction task and create threads, processes, etc. in the segmented data;

The reconstruction module A20 is configured to call the thread, process, etc. of at least one node determined by the distributed computing processing module according to the distributed service interface in the distributed reconstruction system to process the divided local reconstruction task to complete the reconstruction.

It should be noted that the reconstruction module A20 can learn from the function of the reconstruction module in the existing cloud workstation, and realize the local reconstruction task by calling the process, thread and other interfaces created by the distributed computing processing module A10 corresponding to distributed services of computing resources. reconstruction.

The resource allocation strategy in this embodiment is used to determine one or more nodes capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and pair the selected node with the processing efficiency of the node and the available resource information of the node according to the selected node. The local reconstruction task is segmented.

In a possible implementation manner, the above-mentioned distributed computing processing module A10 may include: a network communication unit A11, a computing resource management unit A12, a computing resource dynamic utilization unit A13, and a data division and merging unit A14.

Wherein, the network communication unit A11 is used to communicate with other nodes in the distributed reconstruction system;

The computing resource management unit A12 is used to periodically obtain the available resource information of other nodes in the distributed reconstruction system; the computing resource management unit A12 in this embodiment maintains multiple types of tables, for example, the available resources of each node updated in real time Hash table of resource information, real-time updated record table of node processing efficiency of each node, data packet information summary table of the number of data packets currently processed by each node, etc. In other embodiments, the table maintained by the computing resource management unit A12 may be a data packet information table and a node efficiency information table corresponding to the task to be rebuilt, which is not limited in this embodiment, and can be adjusted according to actual needs.

For example, the data packet information table includes various information contained in the data packet header, and the table records the number of data packets. There are multiple data packet information tables, and each data processing node corresponds to a data packet information table, which is used to analyze the processing efficiency of each node. The data packet information table in this embodiment may be a linked list. The characteristic of the linked list is that it is extremely efficient to delete and add nodes at the head and tail. It is suitable for situations where the tabular data changes in real time.

The node efficiency information table includes but is not limited to the following information: node name, node broadcast number, data processing time of the node in the past period, data fragment size processed by the node in the past period, maximum data processing speed of the node, Minimum data processing speed. The "past period of time" here refers to a period of time calculated forward with the current time as the end point, which can be set according to the actual situation, generally 60 seconds.

The computing resource management unit A12 can calculate the processing efficiency related information of the corresponding node in the corresponding time period based on the information in the internally maintained table, and update the information to the node efficiency information table in real time.

In a possible implementation manner, the above-mentioned node processing efficiency may also be calculated by the computing resource management unit according to the following formula (1) and formula (2) and updated in real time.

The node processing efficiency of each node in this embodiment includes: the average processing time and average processing speed of the node for one data packet;

Formula (1) is the average processing time,

Formula (2) is the average processing speed;

A is the node representation, n is the total number of data packets processed, T _k is the processing time of the k-th data packet; D _k is the number of bytes of the k-th data packet.

A computing resource dynamic utilization unit A13, configured to determine one or more nodes capable of processing the local reconstruction task based on the resource allocation strategy, available resource information of the current node, and available resource information of other nodes in the broadcast domain;

The data dividing and merging unit A14 is configured to divide the local reconstruction task according to the selected node, the processing efficiency of the node, and the available resource information of the node, and distribute it to the selected node for processing.

Embodiment 3

In order to better understand the functions of each unit in the distributed computing processing module A10 in the second embodiment, each unit will be described in detail with reference to FIG. 3 and FIG. 4 .

The network communication unit A11 is used for communication between nodes. Since all nodes are in the cloud, virtual subnets can be formed between nodes. When the virtual subnet works normally, SOCKET communication can be used. When the virtual subnet fails, it can automatically switch to HTTPSOCKET, that is, the network communication unit A11 supports two communication methods based on HTTPSOCKET and SOCKET. The communication method of this embodiment can ensure high availability and expandability of the network communication unit.

The computing resource management unit A12 is configured to manage available computing resource information (that is, available resource information) of all computing nodes of the distributed reconstruction system.

In practical applications, the available computing resource information includes one or more of the following: node identification, node broadcast number, node IP, node service port number, CPU information of the node, response speed of the node, supported algorithms version etc.

The CPU information of the node includes one or more of the following: the total number of CPU cores, the number of currently available cores; the memory information of the node, etc.;

The memory information of the node includes one or more of the following: total memory size, current free memory size, etc.

The computing resource management unit A12 maintains a table/data structure/hash table containing the above-mentioned available computing resource information for each node, so that the distributed computing processing module A10 can use it.

Usually, the computing resource management unit A12 dynamically maintains the above table, for example, refreshes it at a fixed time interval (usually less than 1 second) to ensure that the real-time status of each node is maintained. In specific processing, the above table can be implemented using a hash table structure to improve the speed of element access and modification.

The data splitting and merging unit A14 is responsible for splitting the data to which the local reconstruction task belongs, and then distributes it to the node that calls and processes the split data, so that the node can process it. After dividing the data, the data dividing and merging unit A14 forms a data packet with the divided data, and the structure of the data packet is divided into two parts: the data packet header (PackageHead) and the data (PackageData). The data packet header contains but is not limited to the following information: data size (in bytes, how many bytes are the data length), data index (the position of the data in the original data after the entire reconstruction), sending node, receiving node, data Sending time (sending node), data recovery time (sending node), data receiving time (receiving node), data sending time (receiving node), data processing time (empty when sending, filled by the node that receives and processes the data) . After the packet header, is the data, as shown in Figure 4.

In this embodiment, the distributed computing processing module A10 will appropriately fine-tune the amount of data and the size of data packets allocated to the node according to the processing efficiency of different nodes, so as to improve the overall processing efficiency.

Embodiment 4

With respect to the distributed reconstruction system described in the above embodiments, the following describes the use process of the distributed reconstruction system in detail with reference to step 01 to step 07 .

Step 01: Each node in the distributed reconstruction system transmits the computing resource information and/or data to be reconstructed of the node to the broadcast domain in the form of broadcast at a specified frequency.

At the same time, each node is also used to receive broadcast packets with the same information sent by other nodes in the distributed reconstruction system.

Each node reads the information in the broadcast packet, and updates the corresponding information table (such as the above hash table) maintained locally.

Step 02: If there is information that the distributed computing processing module A10 in a node is called (that is, when a node in the distributed reconstruction system receives the local reconstruction task of the node), obtain the information that the distributed computing processing module A10 can use. computing resources (that is, available resource information). Determine the computing resources currently available on the local node and available computing resources on other nodes in the broadcast domain. For example, the distributed computing processing module A10 of the first node receiving the local reconstruction task determines the computing resources available to the local node and the computing resources available to other nodes in the broadcast domain according to the pre-maintained hash table, and

The distributed computing processing module A10 of the first node determines the actually used local computing resources and computing resources in the broadcast domain according to the local reconstruction concurrency limit and the reconstruction concurrency limit in the broadcast domain. The broadcast domain here is the distributed reconstruction system.

Step 03: The distributed computing processing module A10 of the first node creates a reconstruction processing process including multiple threads.

In this embodiment, when creating a process, the local reconstruction task and the information of available computing resources are handed over to the data splitting and merging unit A14, and the data size allocated to each node is determined by the data splitting and merging unit A14;

Next, the divided data is then handed over to each thread for processing. Some of these threads run on the local node and some on other nodes in the broadcast domain. In this embodiment, based on the broadcasting method, the distributed computing processing module A10 of the first node can acquire the reconstruction progress information of each node in real time.

If the local node is providing computing resource services to other nodes at the time of starting a new local reconstruction task, and the computation has not been completed, only the idle computing resources of the local node are temporarily used.

If other nodes occupy the computing resources of the local node during the reconstruction process, after the remote thread ends, the computing resources occupied by the thread are recovered for local reconstruction use. Other nodes learn through broadcast packets that the computing resources provided by this node are gradually decreasing.

For example, during the reconstruction process, it is learned that other nodes have started local reconstruction tasks through broadcast packets of other nodes, and the number of threads allocated to this node is gradually reduced.

When other nodes start their local reconstruction tasks, they will gradually recycle computing resources for their local use, so the computing resources shared by other nodes will gradually decrease. Therefore, the node will automatically reduce the number of threads allocated to the node based on this information.

The above logic is planned to be executed in a time slice. After the time slice ends, the distributed computing processing module A10 will plan the allocation of reconstruction threads for the next time slice according to the computing resources of the local node and the nodes in the broadcast domain at that time. The length of the time slice can be set according to specific conditions, usually between 500 milliseconds and 5000 milliseconds.

The emphases of the above-mentioned embodiments are different, and there is no contradiction in their integration with each other, and can be selected and set as required in practical applications.

It should be noted that, in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not preclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In the claims enumerating several means, several of these means can be embodied by one and the same item of hardware. The words first, second, third, etc. are used for convenience only and do not imply any order. These words can be understood as part of the part name.

In addition, it should be noted that in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiments", "examples", "specific examples" or "some examples", etc., are Indicates that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, those skilled in the art may combine and combine the different embodiments or examples described in this specification, as well as the features of the different embodiments or examples, without conflicting each other.

Although the preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments will occur to those skilled in the art after learning the basic inventive concepts. Therefore, the claims should be construed to include the preferred embodiment and all changes and modifications that fall within the scope of the present invention.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention. Thus, provided that these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include these modifications and variations.

Claims (9)

1. A distributed reconstruction resource management method based on cloud equipment is characterized in that the distributed reconstruction resource management method is applied to a distributed reconstruction system for medical image reconstruction in medical image equipment, a plurality of nodes for realizing a data reconstruction program are configured in the distributed reconstruction system, all the nodes are located in a broadcast domain of the distributed reconstruction system, and available resource information is periodically broadcast in a broadcast mode, and the method comprises the following steps:

s10, the first node periodically acquires the available resource information in the first node and broadcasts the available resource information to other nodes in the broadcast domain, and receives the available resource information broadcast by each node in the broadcast domain;

s20, the first node judges whether a local reconstruction task sent by the video equipment connected with the first node is received;

s30, if the first node receives the local reconstruction task, according to the resource allocation strategy, the available resource information of the current first node and the available resource information of other nodes in the broadcast domain, performing data segmentation on the local reconstruction task and creating a thread of segmented data to complete the local reconstruction task in a distributed manner;

the resource allocation strategy is used for determining more than one node capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and segmenting the local reconstruction task according to the selected node, the processing efficiency of the node and the available resource information of the node.

2. The method of claim 1, further comprising:

after the data processing distributed by all the nodes processed by the local reconstruction task is finished, the available resource information in the node is updated and broadcasted;

when a first node determines more than one node capable of processing the local reconstruction task, the priority of the first node is the highest priority.

3. The method of claim 1, wherein each node maintains a data structure for available resource information of each node in the broadcast domain and a node efficiency information table of each node;

the information in the node efficiency information table of each node stored in the first node is the processing efficiency of each node calculated by the first node based on the data packet processed by each node in the specified time.

4. The method of claim 1, wherein communication between any two nodes is performed in a SOCKET communication manner;

the available resource information includes one or more of:

the total number of cores of the CPU and the number of currently available cores; memory information of the node, idle memory information, response speed and supported algorithm.

5. A server belonging to a node of a distributed reconstruction system for reconstructing medical images in medical imaging equipment, the server being located in a broadcast domain and broadcasting available resource information to other nodes in the distributed reconstruction system in a broadcast manner, the server comprising:

the distributed computing processing module is used for communicating with other nodes in the distributed reconstruction system, receiving a local reconstruction task transmitted by the image equipment connected with the server, and performing data segmentation on the local reconstruction task and creating a thread and/or a process of segmented data based on a resource allocation strategy, available resource information of the current node and available resource information of other nodes in a broadcast domain;

the reconstruction module is used for calling the thread and/or process of at least one node determined by the distributed computing processing module according to a distributed service interface in the distributed reconstruction system to process the segmented local reconstruction task so as to complete reconstruction;

the resource allocation strategy is used for determining more than one node capable of processing the local reconstruction task according to the processing information of the local reconstruction task, and segmenting the local reconstruction task according to the selected node, the processing efficiency of the node and the available resource information of the node.

6. The server according to claim 5, wherein the distributed computing processing module comprises:

the network communication unit is used for communicating with other nodes in the distributed reconstruction system;

the system comprises a calculation resource management unit, a broadcast domain management unit and a server, wherein the calculation resource management unit is used for periodically acquiring available resource information of other nodes in the distributed reconstruction system and maintaining a data structure of the available resource information of each node in the broadcast domain stored in the server and a node efficiency information table of each node;

the computing resource dynamic utilization unit is used for determining more than one node capable of processing the local reconstruction task based on a resource allocation strategy, available resource information of the current node and available resource information of other nodes in a broadcast domain;

and the data segmentation and combination unit is used for segmenting the local reconstruction task according to all the selected nodes, the processing efficiency of each node in all the nodes and the available resource information, and distributing the segmented local reconstruction task to the selected nodes for reconstruction processing.

7. The server according to claim 6, wherein the network communication unit is configured to communicate in a SOCKET communication manner; or, an HTTPSOCKET communication mode is adopted for communication;

the computing resource management unit is used for sending the available resource information and the reconstruction task information of the node to the broadcast domain in a broadcast mode; receiving broadcast packets sent by other nodes in the broadcast domain;

the reconstruction task information comprises information of a local reconstruction task or information of a reconstruction task distributed by other nodes;

the computing resource dynamic utilization unit is used for acquiring available resource information which can be used by the node after receiving the local reconstruction task and determining a reconstruction algorithm; the available computing resource information of other nodes in the broadcast domain, the available resource information of the local server and the available resource information of other nodes in the broadcast domain are determined according to the local reconstruction concurrency limit and the reconstruction concurrency limit in the broadcast domain, all nodes for processing the local reconstruction task are selected, and a plurality of threads and/or processes for processing the local reconstruction task are created;

the data dividing and merging unit is used for receiving the thread and/or the process sent by the dynamic computing resource utilization unit, the available resource information of all the selected nodes and the processing efficiency of each node in all the selected nodes, dividing the local reconstruction task and determining the size of data distributed to each selected node for distribution;

the threads and/or processes that process the partitioned data run partly at the local node and partly at other nodes in the broadcast domain.

8. The server of claim 7, further comprising:

the calculation resource dynamic utilization unit creates a maximum number of threads for processing the local reconstruction task based on the available resource information of all the selected nodes.

9. A distributed reconstruction system based on cloud equipment, comprising a plurality of servers according to any one of claims 5 to 8, wherein each server is used as a node to form a decentralized broadcast domain, and each server corresponds to more than one imaging device located in a hospital.

Images