Distributed network simulation engine supporting digital twinning

Abstract

The invention provides a distributed network simulation engine supporting digital twinning, which relates to the technical field of digital twinning simulation and comprises a master control node and a plurality of distributed simulation nodes, the plurality of distributed simulation nodes are respectively connected with the master control node, and the master control node and the plurality of distributed simulation nodes are respectively connected to the switch through the network card; a virtual switch OVS and an LXC / Docker container are arranged in the distributed simulation node; the distributed network simulation engine performs simulation by adopting data driving and model driving; and the virtual switch OVS is used for constructing a tunnel interface for communication, implementing a data plane and implementing a control plane. The digital twinning simulation method has the advantage that more accurate and reliable simulation is carried out for digital twinning implementation.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twin simulation, and in particular to a distributed network simulation engine supporting digital twins. Background Art

[0002] A digital twin is essentially a simulation model of a physical object that can evolve in real time by receiving data from the physical object, thereby remaining consistent with the physical object throughout its life cycle.

[0003] As the best way to achieve interactive integration of physical space and virtual space, digital twins have received great attention in the field. Existing network simulation engines such as OPNET and Exata cannot support network twin functions well, for example, there will be defects such as large errors.

[0004] Therefore, there is an urgent need to design and optimize distributed simulation engines so that they can better implement more accurate and reliable simulations for digital twins. Summary of the invention

[0005] The purpose of the present invention is to provide a distributed network simulation engine supporting digital twins, which can better realize more accurate and reliable simulation for digital twins.

[0006] The present invention is achieved through the following technical solutions:

[0007] A distributed network simulation engine supporting digital twins, comprising a master control node and multiple distributed simulation nodes; the multiple distributed simulation nodes are respectively connected to the master control node, and the master control node and the multiple distributed simulation nodes are respectively connected to a switch through a network card;

[0008] The distributed simulation node is provided with a virtual switch OVS and an LXC / Docker container;

[0009] The distributed network simulation engine uses data-driven and model-driven simulation;

[0010] The virtual switch OVS is used to build tunnel interfaces for communication, implement a data plane, and implement a control plane.

[0011] Preferably, the method of simulating using data drive includes the following steps which are performed cyclically:

[0012] Conduct data collection;

[0013] Performing data modeling to obtain a data model, and establishing a storage solution for the data model;

[0014] Conduct simulation and perform data analysis based on the simulation results;

[0015] Provide data feedback and optimize the simulation framework based on the feedback results.

[0016] Preferably, when performing the data collection, the collected content includes data type, data owner and data source.

[0017] Preferably, the storage scheme includes whether it is appendable, whether it is modifiable, access type and whether it needs to be deleted.

[0018] Preferably, the method for the virtual switch OVS to construct a tunnel interface for communication includes:

[0019] Divide the LAN to isolate the LAN between virtual machines and between virtual machines across hosts;

[0020] A tunnel is built through the virtual switch OVS of the edge device, where the tunnel is a GRE tunnel or a VXLAN tunnel, and the VTEP node in the VXLAN tunnel is implemented through the virtual switch OVS;

[0021] Realize data forwarding of hosts under different LAN network topologies and realize routing forwarding.

[0022] Preferably, the method for implementing LAN isolation between virtual machines is that the virtual switch OVS of the edge device adds the virtual machines in the host to different LANs respectively;

[0023] The method for implementing LAN isolation between virtual machines across hosts is to assign the virtual switches OVS of edge devices to different LANs respectively.

[0024] Preferably, the communication method between the virtual machines in different local area networks is:

[0025] Send the ARP message in the local area network to the controller so that the master node can perceive the network nodes in the Openflow network;

[0026] Get network device information in openf l ow;

[0027] The corresponding ARP flow table is sent to the virtual switch OVS of the edge device in the networking.

[0028] Preferably, the distributed network simulation engine realizes data communication between various servers through SSH secure session and data tunnel technology, and the method includes:

[0029] Create a key pair in the master server;

[0030] Put the public key on each slave server;

[0031] When the master server sends a connection request to the slave server, verification is performed using the master server's key and the slave server's shared key.

[0032] Preferably, when the distributed network simulation engine dynamically adds nodes or process services, it schedules and allocates server resources, and the allocation method is:

[0033] The system resources are virtualized into a resource pool. When the system needs computing, storage, and network resources, the system searches the resource pool and determines whether the remaining resources can meet the needs. If they can, the resources are dynamically allocated.

[0034] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0035] The present invention supports local deployment and network deployment, and the supported user scale and usage methods are much larger than those of traditional simulation engines;

[0036] The present invention can separate the model from the data and integrate the main functions into the simulation layer, which is beneficial to function expansion and modularization.

[0037] The present invention uses a combination of model-driven and data-driven methods to achieve simulation, which is beneficial to improving the accuracy of the model. The model can also be driven in real time using data;

[0038] The present invention has reasonable design and simple structure, is easy to apply to various simulation requirements and simulation scenarios, and has high scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of the structure of a distributed network simulation engine supporting digital twins provided in Example 1 of the present invention;

[0040] Figure 2 A schematic diagram of the principle of a distributed application case provided in Example 1 of the present invention;

[0041] Figure 3 A distributed design framework provided in Example 1 of the present invention;

[0042] Figure 4 This is an event-driven simulation principle diagram provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0044] Example 1

[0045] This embodiment provides a distributed network simulation engine that supports digital twins. Figure 1 , including a master control node and multiple distributed simulation nodes; the multiple distributed simulation nodes are respectively connected to the master control node, and the master control node and the multiple distributed simulation nodes are respectively connected to the switch through a network card, namely eth0 in the figure;

[0046] The distributed simulation node is provided with a virtual switch OVS and an LXC / Docker container;

[0047] The distributed network simulation engine uses data-driven and model-driven simulation;

[0048] The virtual switch OVS is used to build tunnel interfaces for communication, implement a data plane, and implement a control plane.

[0049] As the nodes of future networks gradually become complex and heterogeneous, such as the number of satellite nodes in the Starlink system reaching 42,000, the number of network nodes in different systems such as the Internet of Vehicles, Internet of Things, and 6G has reached an astonishing number. Currently, mature engines cannot support such a large network scale, and since the computing power, storage and network capabilities of a single server are limited, distribution has become an inevitable choice. Because in the construction of a network digital twin system, the scale of engine support and data authenticity are very important. The network scale supported by its engine is required to be large. In view of the problem that traditional virtualization technology cannot respond quickly to user needs and has low resource utilization, this embodiment uses container technology to simulate real nodes in the simulation network. Virtualization container technology can give full play to the characteristics of the operating system and meet the requirements of lightweight network node construction.

[0050] Based on the structure of this embodiment, LXC (Li nux Containers) is a lightweight virtualization technology that allows the creation and operation of multiple independent virtual containers on a single kernel host. Compared with traditional virtualization technologies, LXC has smaller resource usage and higher performance, making it widely used in cloud computing and containerization fields; Docker is based on the cgroup, namespace, and Union FS of the AUFS class of the Li nux kernel to encapsulate and isolate processes. Since the isolated process is independent of the host and other isolated processes, it is also called a container; the virtual switch OVS supports multi-layer data forwarding and is mainly deployed on servers. Compared with traditional switches, it has good programming scalability and has the network isolation and data forwarding functions implemented by traditional switches. It runs on each physical machine that implements virtualization and provides remote management.

[0051] The distributed network simulation engine of this embodiment uses RPC technologies such as SSH to achieve command synchronization between multiple hosts. In the distributed network simulation engine simulation, different simulation nodes of a simulation scene are deployed in different hosts in a distributed environment, and a business is split into multiple sub-businesses. In this way, parallel computing simulations cooperate with each other to achieve the same common goal. For a distributed application case, please refer to Figure 2 .

[0052] In order to cope with the complex and changeable problems of large parallel tasks in large simulation scenarios, the distributed network simulation engine of this embodiment can also be deployed on multiple simulation servers, that is, multiple servers in the figure, and controlled by a single controller. The controller representing the entire topology scenario can be run on one of the simulation servers or a separate machine. Its distributed design framework can be referred to in Figure 3 .

[0053] Based on the above scheme, the distributed network simulation engine of this embodiment has the advantage of large scale. In addition, it can minimize the simulation's dependence on physical devices and use virtual network devices to reduce the complexity of deployment, thereby reducing subsequent maintenance and deployment steps for hardware devices. Software automated operation deployment can completely replace this part of the work, and multiple computers can be connected together so that work can be completed on the most suitable computer. The workload can be distributed to available machines in the most effective way to achieve load balancing.

[0054] In this embodiment, the combination of data-driven and model-driven methods can better improve the accuracy of the digital model, make the simulation more reliable and realistic, and achieve the simulation effect of digital twins. In actual data-driven simulation, data is the source of reports and in-depth simulation predictions. The method of simulation using data-driven in this embodiment preferably includes the following steps that are performed cyclically:

[0055] Conduct data collection;

[0056] Performing data modeling to obtain a data model, and establishing a storage solution for the data model;

[0057] Conduct simulation and perform data analysis on the simulation results, such as finding key data and corresponding indicators in the simulation results for analysis;

[0058] Conduct data feedback and optimize the simulation architecture based on the feedback results. In this step, users are accurately grouped, such as knowing user needs and business data, and finally accurately characterizing the simulation architecture. On this basis, targeted improvements can be made to the simulation, forming an automated and refined closed-loop simulation.

[0059] In summary, this embodiment supports automated decision-making based on massive data, automated services, and models, forming a cycle to continuously collect data, complete modeling, and make automated decisions, thus forming a data-driven simulation.

[0060] Furthermore, when performing the data collection, the collected content includes data type, data owner and data source.

[0061] On the other hand, depending on the type of model, the storage solution includes whether it can be appended, whether it can be modified, the access type and whether it needs to be deleted.

[0062] As a preferred solution, the method for the virtual switch OVS to construct a tunnel interface for communication includes:

[0063] Divide the LAN to isolate the LAN between virtual machines and between virtual machines across hosts;

[0064] A tunnel is built through the virtual switch OVS of the edge device. The tunnel is a GRE tunnel or a VXLAN tunnel. The VTEP node in the VXLAN tunnel is realized by the virtual switch OVS. A tunnel refers to a channel through which a message is transmitted using a specific protocol. For example, if a VXLAN tunnel is used in this embodiment, the original message is encapsulated by the upper layer message, reaches the VTEP node, and then is encapsulated by the VXLAN protocol, and then is sent out through the VXLAN tunnel forwarding path. The virtual switch OVS can realize the VXLAN message parsing, encapsulation, forwarding, address learning and other functions that VTEP is responsible for.

[0065] Realize data forwarding of hosts under different LAN network topologies and realize routing forwarding.

[0066] Specifically, the method for implementing LAN isolation between virtual machines is that the virtual switch OVS of the edge device adds the virtual machines in the host to different LANs respectively;

[0067] The method for implementing LAN isolation between virtual machines across hosts is to assign the virtual switches OVS of edge devices to different LANs respectively. This method can hinder the layer 2 communication between the virtual switches OVS of edge devices.

[0068] In addition, the communication method between the virtual machines in different local area networks is:

[0069] Send the ARP message in the local area network to the controller so that the master node can perceive the network nodes in the Openflow network;

[0070] Get network device information in openf l ow;

[0071] The corresponding ARP flow table is sent to the virtual switch OVS of the edge device in the networking.

[0072] In distributed simulation, there is a master server in the DSE system, and the others are slave servers. As a preferred solution of this embodiment, the distributed network simulation engine realizes data communication between various servers through SSH (Secure Shell Protocol Secure Shell) secure session and data tunnel technology, and the method includes:

[0073] Create a key pair in the master server;

[0074] Put the public key on each slave server;

[0075] When the master server sends a connection request to the slave server, verification is performed using the master server's key and the slave server's shared key.

[0076] In the above way, an SSH tunnel can be established, and the data of the master server can flow safely to each slave server.

[0077] Finally, when the distributed network simulation engine dynamically adds nodes or process services, it schedules and allocates server resources, and the allocation method is preferably:

[0078] The system resources are virtualized into a resource pool. When the system needs computing, storage and network resources, the system searches the resource pool and determines whether the remaining resources can meet the needs. If they can, the resources are dynamically allocated. The distributed network simulation engine of this embodiment is mainly data-driven, and the data of the simulation node can also be written into the event table using traditional event-driven. Therefore, the data-driven table needs to schedule and allocate system resources and tasks in the master server according to the specific circumstances of the data or events. This is also the execution unit of the specific implementation of the simulation distributed server deployment scheduling algorithm of the distributed network simulation engine of this embodiment. The schematic diagram can be found in Figure 4 .

[0079] Based on the scheme of this embodiment, the distributed network simulation engine of this embodiment is a distributed simulation system designed based on the idea of ​​SDN architecture, i.e., isolation of data plane and control plane. It can cope with large-scale simulation scenarios with high data volume, and use data-driven real-time processing to layer the entire network topology, with high scalability and high availability, and is developed for the physical environment deployment of complex simulation data with low precision. Therefore, its specific deployment can be divided into a foreground subsystem and a background engine subsystem. The foreground subsystem is user-oriented and can provide a simulation control interface, as well as analysis and display of simulation operation results, etc., through C / S or B / S modes. The background engine subsystem realizes the operation of the simulation and monitors the return operation status according to the user's simulation scenario configuration. The foreground subsystem and the background engine subsystem communicate data through a local area network, a wide area network or a cloud, and interact through a database interface.

[0080] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A distributed network simulation engine supporting digital twins, characterized in that: It includes a main control node and multiple distributed simulation nodes; the multiple distributed simulation nodes are respectively connected to the main control node, and the main control node and the multiple distributed simulation nodes are respectively connected to the switch through the network card; The distributed simulation node is provided with a virtual switch OVS and an LXC / Docker container; The distributed network simulation engine uses data-driven and model-driven simulation; The virtual switch OVS is used to build tunnel interfaces for communication, implement a data plane, and implement a control plane.

2. A distributed network simulation engine supporting digital twins according to claim 1, characterized in that: The method for simulating by data drive comprises the following steps which are performed cyclically: Conduct data collection; Performing data modeling to obtain a data model, and establishing a storage solution for the data model; Conduct simulation and perform data analysis based on the simulation results; Provide data feedback and optimize the simulation framework based on the feedback results.

3. A distributed network simulation engine supporting digital twins according to claim 2, characterized in that: When conducting the above-mentioned data collection, the collected content includes data type, data owner and data source.

4. A distributed network simulation engine supporting digital twins according to claim 3, characterized in that: The storage scheme includes whether it can be appended, whether it can be modified, the access type and whether it needs to be deleted.

5. A distributed network simulation engine supporting digital twins according to claim 1, characterized in that: The method for constructing a tunnel interface for communication by the virtual switch OVS includes: Divide the LAN to isolate the LAN between virtual machines and between virtual machines across hosts; A tunnel is built through the virtual switch OVS of the edge device, where the tunnel is a GRE tunnel or a VXLAN tunnel, and the VTEP node in the VXLAN tunnel is implemented through the virtual switch OVS; Realize data forwarding of hosts under different LAN network topologies and realize routing forwarding.

6. A distributed network simulation engine supporting digital twins according to claim 5, characterized in that: The method for implementing LAN isolation between virtual machines is that the virtual switch OVS of the edge device adds the virtual machines in the host to different LANs respectively; The method for implementing LAN isolation between virtual machines across hosts is to assign the virtual switches OVS of edge devices to different LANs respectively.

7. A distributed network simulation engine supporting digital twins according to claim 6, characterized in that: The communication method between the virtual machines in different local area networks is: Send the ARP message in the LAN to the controller so that the master node can perceive the network nodes in the Openflow network; Get network device information in openflow; The corresponding ARP flow table is sent to the virtual switch OVS of the edge device in the networking.

8. A distributed network simulation engine supporting digital twins according to claim 1, characterized in that: The distributed network simulation engine realizes data communication between various servers through SSH secure session and data tunnel technology, and the method includes: Create a key pair in the master server; Put the public key on each slave server; When the master server sends a connection request to the slave server, verification is performed using the master server's key and the slave server's shared key.

9. A distributed network simulation engine supporting digital twins according to claim 1, characterized in that: When the distributed network simulation engine dynamically adds nodes or process services, it schedules and allocates server resources in the following manner: The system resources are virtualized into a resource pool. When the system needs computing, storage, and network resources, the system searches the resource pool and determines whether the remaining resources can meet the needs. If they can, the resources are dynamically allocated.

Images