Hospital multi-activity data center network management system based on software defined network

Abstract

The invention discloses a hospital multi-activity data center network management system based on a software defined network, and relates to the technical field of data management, the network management system is controlled based on an SDN controller, and the network management system comprises a flow collection module, a prediction module, a scheduling module, a multi-activity data module and a security module; the flow collection module is used for collecting historical people flow data of a hospital, analyzing and managing the acquired people flow data to obtain flow load historical data of each device of the hospital, and storing the data; and the prediction module predicts the future hospital equipment flow load based on the historical data acquired in the flow collection module, and uploads the prediction result to the scheduling module. According to the method, the future hospital equipment traffic load can be predicted, the network traffic is reasonably distributed to each data center and the corresponding server according to the prediction results, the network configuration is flexible, the network resource distribution is quickly adjusted, and the medical experience of a patient is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and more particularly to a network management system for a hospital multi-active data center based on software-defined networking. Background Art

[0002] A multi-active data center refers to building multiple data centers that are all in an active state at the same time. They are distributed in different geographical locations and are connected to each other through a high-speed network. They can synchronize and interact data in real time or near real time, and can all provide services externally. Different from the traditional primary and standby data center mode, in the primary and standby mode, the standby data center is mostly idle usually, and is only enabled when the primary data center fails; while each center in the multi-active data center undertakes the task of business processing in daily operations and jointly provides services for users. With the continuous advancement of hospital informatization construction, a large amount of medical data has been accumulated within the hospital, such as patient electronic medical records, medical imaging materials, and test results. And many key business systems rely on a stable and reliable data center to operate. The traditional data center network management method has many deficiencies when facing the requirements of a hospital multi-active data center. The network configuration lacks flexibility and it is difficult to quickly adjust the network resource allocation according to different business scenarios, resulting in slow response of some systems during the business peak period, affecting the patient's medical experience and the hospital's operation efficiency. In view of this, we propose a network management system for a hospital multi-active data center based on software-defined networking. Summary of the Invention

[0003] To solve the above technical problems, a network management system for a hospital multi-active data center based on software-defined networking is provided, and this technical solution solves the problem of slow system response during the above business peak period.

[0004] To achieve the above object, the technical solution adopted by the present invention is: A network management system for a hospital multi-active data center based on software-defined networking, the network management system is controlled based on an SDN controller, and includes: a traffic collection module, a prediction module, a scheduling module, a multi-active data module, and a security module;

[0005] The traffic collection module is used to collect the historical pedestrian flow data of the hospital, analyze and manage the obtained pedestrian flow data, obtain the historical data of the traffic load of each device in the hospital, and store the data;

[0006] The prediction module predicts the future traffic load of hospital devices based on the historical data obtained in the traffic collection module, and uploads the predicted results to the scheduling module;

[0007] The scheduling module distributes the network traffic from each business terminal of the hospital to each data center and the corresponding servers based on the prediction results of the prediction module, the traffic scheduling algorithm, and the sensitivity of the service, to share the load;

[0008] The multi - live data module includes multiple hospital data centers distributed in different geographical locations. Each data center is equipped with a server cluster for storing and processing medical data, and at the same time, network access devices are provided for connecting to the external network. The data centers are connected by high - speed dedicated network links for synchronization and mutual backup and assistance of services. When a service device fails, it can seamlessly take over the corresponding service;

[0009] The security module continuously monitors the data access requirements in the network, intercepts illegal access and attack behaviors, and ensures network security.

[0010] Preferably, the traffic collection module uses a variety of different types of monitoring devices. At each entrance and exit of the hospital, infrared induction counters are installed, and the infrared induction devices record. Cameras are deployed at each floor passage and department entrance inside the hospital to capture patient flow data, and based on the medical record reports of doctors, data is recorded to obtain patient flow data of different departments;

[0011] The prediction module obtains the stored historical data for analysis. After pre - processing the data, through time - series analysis methods, it analyzes the change trend of the number of people flow in different time periods, finds out the peak and trough periods and periodic laws of the hospital's people flow, and uses association analysis algorithms to analyze the internal relationship between the number of people flow and the use of different departments and different devices in the hospital to obtain the historical usage flow load data of hospital equipment.

[0012] Preferably, if the analysis range is n days and there are m observation time points per day, then the average number of people flow P on the i - th day i is expressed as:

[0013]

[0014] where P(i, j) is the number of people flow at the j - th observation time point on the i - th day, and j represents the observation time point.

[0015] Preferably, the formula for calculating the overall average number of people flow is:

[0016]

[0017] where P represents the overall average number of people flow within the entire analysis time period;

[0018] Furthermore, calculate the fluctuation index of the number of people flow:

[0019]

[0020] where αp is the standard deviation of the number of people flow.

[0021] Preferably, the peak and off-peak periods are further determined by setting a relative threshold. The period with a flow rate higher than the overall average plus a certain multiple of the standard deviation is defined as the peak period, and the period with a flow rate lower than the overall average minus a certain multiple of the standard deviation is defined as the off-peak period. Let the multiple of the threshold be k. Then the judgment for the peak period is expressed as:

[0022] P(t) > P + kαp

[0023] where P(t) represents the flow rate at time t;

[0024] The judgment condition for the off-peak period is:

[0025] P(t) < P - kαp

[0026] The peak and off-peak periods of the hospital are analyzed through calculation.

[0027] Preferably, there is a linear relationship between the historical usage flow load data of hospital equipment and the usage of different departments and different equipment in the hospital. The analysis formula is:

[0028] Q = β 0 + β 1 P(t) + ∈

[0029] where Q is the predicted output value, β 0 is the intercept term, β 1 is the regression coefficient, and ∈ is the error term; the intercept term and the regression coefficient are calculated based on the least squares method.

[0030] Preferably, the scheduling module distributes the network traffic of each service terminal to the data center and the server. Guided by the prediction results of the prediction module, the traffic distribution mechanism is started. The traffic distribution formula is:

[0031]

[0032] where F represents the result of the calculated actual network traffic value allocated to the data center server from the server side, Ft represents the size of the network traffic generated by the service terminal at time t, At is the final weight coefficient for allocating the network traffic of the service terminal to the server within the data center, which comprehensively considers multiple factors such as service sensitivity, link bandwidth utilization, server load status, and network latency, and comprehensively reflects the suitability of allocating network traffic to a specific server within a specific data center. d ∈ D represents a specific data center, and s ∈ Sd is one of the servers in data center d; the above formula means that according to the proportion of the final weight coefficient of each data center and server in the total weight coefficient, the network traffic generated by the service terminal is allocated, so as to achieve the goal of distributing the network traffic from each service terminal in the hospital to each data center and the corresponding server based on comprehensive consideration of multiple factors and sharing the network load.

[0033] Preferably, the multi-active data module is jointly composed of multiple hospital data centers, including a server cluster and network access devices. The server cluster is used for data management, and the network access devices are used to connect to the external network. Each data center is tightly connected through a high-speed dedicated network link, and the link uses optical fiber transmission media for transmission. After the server cluster of a certain data center receives a new patient electronic medical record data entry request, the data center will send the newly added data to other data centers through the network link. After receiving the synchronized data, other data centers will perform data update operations in the local server cluster.

[0034] Preferably, after a server fails in the multi-active data module, and other normal servers sense it, the registration requests originally sent to the failed server will be transferred over, and the internal server cluster will continue to process them.

[0035] Preferably, the security module controls the data in the network from different dimensions, captures network nodes, analyzes the header information and payload content of data packets to identify the source and intention of access requests, and insight into the legality of data access.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0037] Through the prediction module, the present invention can predict the future hospital equipment traffic load by means of the rich historical data obtained by the traffic collection module. Based on these prediction results, the network traffic is reasonably allocated to each data center and the corresponding servers, avoiding the performance degradation of the servers due to traffic overload, and at the same time reducing the situation of idle resources of other servers, so that the server resources in the entire network can be fully and appropriately utilized, maximizing their computing and storage capabilities, improving the overall utilization efficiency of server resources, with flexible network configuration, quickly adjusting network resource allocation, and enhancing the patient's medical experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a framework diagram of the network management system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0039] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations.

[0040] Refer to Figure 1 As shown, a hospital multi-active data center network management system based on software-defined network, the network management system is controlled based on an SDN controller, including: a traffic collection module, a prediction module, a scheduling module, a multi-active data module and a security module;

[0041] The traffic collection module is used to collect the historical pedestrian flow data of the hospital, analyze and manage the obtained pedestrian flow data, obtain the historical data of the traffic load of each device in the hospital, and store the data;

[0042] Based on the historical data obtained in the traffic collection module, the prediction module predicts the future traffic load of hospital devices and uploads the prediction results to the scheduling module;

[0043] Based on the prediction results of the prediction module and considering the traffic scheduling algorithm and business sensitivity, the scheduling module distributes the network traffic from each business terminal of the hospital to each data center and the corresponding servers to share the load;

[0044] The multi-active data module includes multiple hospital data centers distributed in different geographical locations. Each data center is equipped with a server cluster for storing and processing medical data, and is also equipped with network access devices for connecting to the external network. The data centers are connected by high-speed dedicated network links for synchronization and mutual backup and assistance of services. When a business device fails, it can seamlessly take over the corresponding business;

[0045] The security module continuously monitors the data access requirements in the network, intercepts illegal access and attack behaviors, and ensures network security.

[0046] This application is based on software-defined, which refers to an idea and method of flexibly controlling, managing, and allocating underlying resources through software as reflected in the related technical concepts of software-defined network and software-defined storage. In this application document, it is the SDN controller;

[0047] With the help of the rich historical data obtained by the traffic collection module, the prediction module can accurately predict the future traffic load of hospital devices. Based on these prediction results, the scheduling module combines the traffic scheduling algorithm and the consideration of business sensitivity to reasonably allocate network traffic to each data center and the corresponding servers. In this way, it avoids the performance degradation of some servers due to traffic overload and reduces the situation of idle resources of other servers, enabling the server resources in the entire network to be fully and appropriately utilized, maximizing their computing and storage capabilities, and improving the overall utilization efficiency of server resources; during the traffic scheduling process, the bandwidth utilization of the network link is considered. Based on the prediction results, the scheduling module can know in advance which services will generate large traffic during a specific period, and then reasonably arrange the traffic direction to avoid the unbalanced state where some links are congested while other links have idle bandwidth;

[0048] By comparing the actual traffic distribution with the predicted results, it is found that the traffic fluctuation of a certain conventional service is greater than expected during a specific period, resulting in uneven loads on some servers. Then, the management personnel can adjust the relevant parameters in the prediction module or the traffic distribution weights in the scheduling module to optimize the subsequent traffic scheduling work and make the network operation more stable and efficient.

[0049] The traffic collection module uses a variety of different types of monitoring devices. At each entrance and exit of the hospital, infrared induction counters are installed and recorded by the infrared induction device. Cameras are deployed at each floor passage and department entrance inside the hospital to capture patient traffic data, and based on the medical record reports of physicians to record data, the patient traffic data of different departments is obtained.

[0050] This application uses a variety of monitoring devices. Installing infrared induction counters at the hospital entrances and exits can accurately record the situation of people entering and leaving the hospital, which provides basic data for mastering the overall inflow, outflow and traffic change trends at different times of the hospital.

[0051] The prediction module obtains and analyzes the stored historical data. After preprocessing the data, through time series analysis methods, it analyzes the change trends of the number of people at different times, finds out the peak and trough periods and periodic laws of the hospital's people flow, and uses association analysis algorithms to analyze the internal relationship between the number of people and the use of different departments and different devices in the hospital to obtain the historical traffic load data of hospital equipment.

[0052] Through the historical traffic load data of hospital equipment obtained by this application's analysis, hospital managers can clearly understand the actual usage and load levels of different departments and different devices at each time period.

[0053] Suppose the analysis range is n days, and there are m observation time points every day, then the average number of people P on the i-th day i is expressed as:

[0054]

[0055] where P(i, j) is the number of people at the j-th observation time point on the i-th day, and j represents the observation time point.

[0056] The formula for calculating the overall average number of people flow is:

[0057]

[0058] where P represents the overall average number of people flow during the entire analysis time period;

[0059] Furthermore, calculate the fluctuation index of the number of people flow:

[0060]

[0061] where αp is the standard deviation of the pedestrian flow.

[0062] Further determine the peak and trough periods, set a relative threshold. The period higher than the overall average pedestrian flow plus a certain multiple of the standard deviation is the peak period, and the period lower than the overall average pedestrian flow minus a certain multiple of the standard deviation is the trough period. Let the threshold multiple be k, then the judgment of the peak period is expressed as:

[0063] P(t) > P + kαp

[0064] where P(t) represents the pedestrian flow at time t;

[0065] The judgment condition for the trough period is:

[0066] P(t) < P - kαp

[0067] Analyze and obtain the peak and trough time periods of the hospital through calculation.

[0068] There is a linear relationship between the historical usage flow load data of hospital equipment and the usage of different departments and different equipment in the hospital. The analysis formula is:

[0069] Q = β 0 + β 1 P(t) + ∈

[0070] where Q is the predicted output value, β 0 is the intercept term, β 1 is the regression coefficient, and ∈ is the error term; Calculate the intercept term and the regression coefficient based on the least squares method.

[0071] The scheduling module distributes the network traffic of each business terminal to the data center and the server. Guided by the prediction results of the prediction module, start the traffic distribution mechanism. The traffic distribution formula is:

[0072]

[0073] Among them, F represents the result of calculating the actual network traffic value allocated to the data center server from the server side, Ft represents the magnitude of the network traffic generated by the service terminal at time t, and At is the final weight coefficient for allocating the network traffic of the service terminal to the servers within the data center. It synthesizes multiple factors such as service sensitivity, link bandwidth utilization, server load status, and network latency, and comprehensively reflects the suitability of allocating network traffic to a specific server within a specific data center. d ∈ D represents a specific data center, and s ∈ Sd is one of the servers in data center d; the above formula means that according to the proportion of the final weight coefficients of each data center and server in the total weight coefficient, the network traffic generated by the service terminal is allocated, so as to achieve the goal of allocating the network traffic from each service terminal of the hospital to each data center and the corresponding server based on comprehensive consideration of multiple factors and sharing the network load.

[0074] This application determines the weight coefficient by comprehensively considering the factor of service sensitivity, and can accurately distinguish the urgency and importance of different services' demands for network resources. As part of the weight coefficient, the link bandwidth utilization prompts that in the process of traffic allocation, the data center and server connected by a link with relatively abundant bandwidth and reasonable utilization rate will be preferentially selected to carry network traffic.

[0075] The multi-active data module is jointly composed of multiple hospital data centers, including a server cluster and network access devices. The server cluster is used for data management, and the network access devices are used to connect to the external network. Each data center is closely connected by a high-speed dedicated network link, and the link uses optical fiber transmission medium for transmission. After the server cluster of a certain data center receives a new request for entering patient electronic medical record data, the data center will send the newly added data to other data centers through the network link. After receiving the synchronized data, other data centers will perform data update operations in their local server clusters.

[0076] This application jointly constitutes a multi-active data module by multiple hospital data centers, which means that there are multiple copies of data stored in different geographical locations. Even if a certain data center encounters force majeure factors resulting in data damage or loss, other data centers still store complete data copies and can provide services at any time.

[0077] After a server in the multi-active data module fails, and other normal servers sense it, the registration requests originally sent to the failed server will be transferred over, and the internal server cluster will continue to process them.

[0078] During the peak visiting hours of this application, a large number of patients are operating the self-service registration machines to register. If a server responsible for processing registration requests suddenly fails, through the automatic transfer mechanism of the multi-active data module, these ongoing registration operations will be quickly redirected to other normal servers, and patients can still complete the registration smoothly. The subsequent visiting procedures can also proceed in an orderly manner, maintaining the smoothness of the overall business process of the hospital.

[0079] The security module controls the data in the network from different dimensions, captures network nodes, analyzes the header information and payload content of data packets to identify the source and intention of access requests, and insight into the legitimacy of data access.

[0080] In the hospital network of this application, there is a vast amount of extremely sensitive medical data stored, including patients' personal basic information, detailed medical histories, diagnosis results, and various medical imaging materials. By capturing network nodes and analyzing the content of data packets to insight into the legitimacy of data access, the security module can accurately intercept access attempts from external illegal sources or unauthorized internal personnel, preventing these sensitive data from being stolen and leaked.

[0081] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A hospital multi-active data center network management system based on software defined network, characterized in that: The network management system is controlled based on the SDN controller, including: traffic collection module, prediction module, scheduling module, multi-active data module and security module; The traffic collection module is used to collect the historical flow data of the hospital, analyze and manage the acquired flow data, obtain the historical flow load data of each device in the hospital, and store the data; The prediction module predicts the future hospital equipment traffic load based on the historical data obtained in the traffic collection module, and uploads the prediction results to the scheduling module; The scheduling module distributes the network traffic from each business terminal of the hospital to each data center and the corresponding server based on the prediction results of the prediction module, the traffic scheduling algorithm and the sensitivity of the business, so as to share the load; The multi-active data module includes multiple hospital data centers distributed in different geographical locations. Each data center is equipped with a server cluster for storing and processing medical data, and a network access device for connecting to the external network. The data centers are connected through a high-speed dedicated network link for synchronization and mutual backup and support of business. When a business device fails, the corresponding business can be seamlessly taken over. The security module continuously monitors data access requirements in the network, intercepts illegal access and attack behaviors, and ensures network security.

2. According to the software defined network based hospital multi-active data center network management system according to claim 1, it is characterized in that: The flow collection module uses a variety of different types of monitoring equipment. Infrared sensor counters are installed at each entrance and exit of the hospital. Infrared sensor devices are used for recording. Cameras are deployed at each floor passage and department entrance inside the hospital to capture patient flow data. Based on the data recorded in the physician's medical record report, patient flow data of different departments is obtained. The prediction module obtains the stored historical data for analysis. After preprocessing the data, it uses the time series analysis method to analyze the changing trends of the flow of people in different time periods, find out the peak and trough periods of the hospital's flow of people and the periodic laws, and use the association analysis algorithm to analyze the intrinsic connection between the flow of people and the use of different departments and different equipment in the hospital, and obtain the historical hospital equipment usage flow load data.

3. A hospital multi-active data center network management system based on software defined network according to claim 2, characterized in that: Assume that the analysis range is n days, and there are m observation time points every day, then the average flow of people on the i-th day is P i It is expressed as: Where P(i, j) is the passenger flow at the jth observation time point on the i-th day, and j represents the observation time point.

4. A hospital multi-active data center network management system based on software defined network according to claim 3, characterized in that: The overall average flow formula for calculating the flow of people is: Where P represents the overall average flow of people during the entire analysis period; Further calculation of the fluctuation index of traffic flow: Where αp is the standard deviation of the flow of people.

5. A hospital multi-active data center network management system based on software defined network according to claim 4, characterized in that: To further determine the peak and valley periods, a relative threshold is set. The peak period is higher than the overall average flow plus a certain multiple standard deviation, and the valley period is lower than the overall average flow minus a certain multiple standard deviation. Let the threshold multiple be k, then the peak period judgment is expressed as: P(t)>P+kαp Where P(t) represents the flow of people at time t; The conditions for judging the low period are: P(t)<P-kαp The peak and trough periods of the hospital can be determined through calculation and analysis.

6. A hospital multi-active data center network management system based on software defined network according to claim 2, characterized in that: There is a linear relationship between the historical hospital equipment usage flow load data and the use of different departments and different equipment in the hospital. The analysis formula is: Q=β0+β1P(t)+∈ Where Q is the predicted output value, β0 is the intercept term, β1 is the regression coefficient, and ∈ is the error term; the intercept term and regression coefficient are calculated based on the least squares method.

7. A hospital multi-active data center network management system based on software defined network according to claim 1, characterized in that: The scheduling module distributes the network traffic of each business terminal to the data center and server, and starts the traffic distribution mechanism based on the prediction results of the prediction module. The traffic distribution formula is: Where F represents the actual network traffic value from the server end that is calculated and allocated to the data center server, Ft represents the size of the network traffic generated by the service terminal at time t, At is the final weight coefficient of the network traffic of the service terminal allocated to the server in the data center, which comprehensively reflects the suitability of allocating network traffic to a server in a specific data center by combining multiple factors such as business sensitivity, link bandwidth utilization, server load status and network delay, d∈D represents a specific data center, and s∈Sd is one of the servers in data center d; The above formula indicates that the network traffic generated by the business terminals is distributed according to the proportion of the final weight coefficient of each data center and server to the total weight coefficient, thereby achieving the goal of distributing the network traffic from various business terminals of the hospital to each data center and the corresponding server based on comprehensive consideration of multiple factors, and sharing the network load.

8. A hospital multi-active data center network management system based on software defined network according to claim 1, characterized in that: The multi-active data module is composed of multiple hospital data centers, including server clusters and network access devices. The server cluster is used to manage data, and the network access device is used to connect to the external network. The data centers are closely connected through high-speed dedicated network links, and the links are transmitted using optical fiber transmission media. After the server cluster of a data center receives a new patient electronic medical record data entry request, the data center sends the newly added data to other data centers through the network link. After receiving the synchronized data, other data centers perform data update operations in the local server cluster.

9. A hospital multi-active data center network management system based on software defined network according to claim 1, characterized in that: When a server in the multi-active data module fails, other normal servers will sense it and transfer the registration requests originally sent to the failed server to continue processing them by their own internal server cluster.

10. A hospital multi-active data center network management system based on software defined network according to claim 1, characterized in that: The security module controls the data in the network from different dimensions, captures network nodes, and parses data packet header information and payload content to identify the source and intent of access requests and gain insight into the legitimacy of data access.

Images