CN-121727630-B - High-speed processing and forwarding simulation method for space network

CN121727630BCN 121727630 BCN121727630 BCN 121727630BCN-121727630-B

Abstract

The invention discloses a high-speed processing and forwarding simulation method of a space network, and belongs to the technical field of network simulation. The method comprises the steps of constructing a space network simulation node aiming at a space network scene to be simulated, configuring and initializing a vector data packet processing component for each simulation node to realize the scale deployment of the simulation nodes, configuring the vector data packet processing components in the simulation nodes at two ends of a simulation link to construct the simulation link according to the link communication requirement and the scale deployment requirement, constructing a simulation link attribute control mechanism, carrying out dynamic configuration of simulation link attributes in each time slot, carrying out simulation link switching based on a smooth switching mechanism and matching with the vector data packet processing components, and capturing and analyzing target data in a simulation process by adopting an automatic simulation data acquisition method.

Inventors

LI WENFENG
JIANG CHEN
YE YANFEI
ZHAO KANGLIAN
LIU JUN

Assignees

南京大学

Dates

Publication Date: 20260512
Application Date: 20260213

Claims (6)

1. A high-speed processing and forwarding simulation method of a space network is characterized by comprising the following steps: s1, initializing a simulation environment, namely constructing a plurality of space network simulation nodes aiming at a space network scene to be simulated, configuring and initializing a vector data packet processing assembly for each simulation node, and realizing the large-scale deployment of the simulation nodes; S2, constructing a hybrid simulation link, namely constructing a space network simulation topology among simulation nodes, constructing a hybrid link architecture based on service attributes of the space links, dividing all simulation links into two types of conventional simulation links and high-speed simulation links according to expected communication loads and performance requirements of the simulation links, and configuring vector data packet processing components with corresponding working modes in the simulation nodes at two ends of the simulation links; s3, the dynamic control of the link attribute driven by the discrete time slot is that the space network simulation process is divided into discrete time slots, the link attribute of each simulation link is updated according to the preset scene parameter in each discrete time slot, and the dynamic configuration and updating of the link attribute of each discrete time slot are executed by loading a link attribute configuration plug-in on a vector data packet processing component in a simulation node; S4, simulation link switching, namely executing simulation link switching based on a smooth switching mechanism, and realizing switching of the simulation links by matching vector data packet processing components in simulation nodes at two ends of the simulation links according to a preset link switching time slot plan and simulation link types; And S5, automatically acquiring data, namely loading a data acquisition plug-in for a vector data packet processing component of the data acquisition plug-in a designated simulation node by adopting a data automatic acquisition method, and acquiring and analyzing target data, wherein the data automatic acquisition method is to automatically export the acquired data for analysis by following a preset data filtering rule and acquisition termination conditions.
2. The method for high-speed processing and forwarding simulation of a space network according to claim 1, wherein the conventional simulation link in step S2 refers to a simulation link without quantization performance constraint for transmission bandwidth and delay jitter change in space network simulation, a simulation scene is configured by default by using such links to form a simulation network topology, and vector data packet processing components in simulation nodes at both ends of the conventional simulation link are configured by using host-interface mode to create a link, and the virtual switch is accessed to perform topology management.
3. The method for high-speed processing and forwarding simulation of a spatial network according to claim 1, wherein the high-speed simulation link in step S2 is a simulation link with a preset bandwidth requirement higher than a preset bandwidth differentiating standard and/or a preset delay jitter upper limit not higher than a preset delay jitter differentiating standard, and the vector data packet processing components in the simulation nodes at both ends of the high-speed simulation link are configured based on memif modes of a shared memory to create the link.
4. The method for high-speed processing and forwarding simulation of a spatial network according to claim 1, wherein the discrete time slot driven link attribute dynamic control in step S3 specifically comprises dynamic control of the following attributes: S3.1, in each simulation time slot, calculating the propagation delay of the link according to the dynamic distance of the simulation link in a space scene, and dynamically configuring delay parameters of vector data packet processing components in simulation nodes at two ends of the simulation link through a link attribute configuration plug-in; And S3.2, controlling the packet loss rate of the simulation link, namely dynamically configuring the packet loss rate parameters of vector data packet processing components in the simulation nodes at two ends of the simulation link through a link attribute configuration plug-in each simulation time slot according to a packet loss rate plan determined by a simulation script or a dynamic model.
5. The method for high-speed processing and forwarding simulation of a spatial network according to claim 1, wherein in step S4, when implementing the simulated link switching based on the smooth switching mechanism, different operation flows are adopted according to different types of the simulated links: s4.1, the switching flow of the conventional simulation link is as follows: (1) Before the switching moment is planned, logic configuration of the new link is finished in the simulation nodes at the two ends of the new simulation link in advance through the vector data packet processing component, but the new link is not connected into the virtual switch; (2) When the simulation time reaches the planned switching time, enabling the logic ports of the simulation nodes at the two ends of the new link on the virtual switch to realize the access of the new link; (3) Disabling logical ports of the simulation nodes at two ends of the old link on the virtual switch, realizing the dismantling of the old link and completing smooth switching; S4.2, the switching flow of the high-speed simulation link is as follows: (1) Before the switching moment is planned, creating and configuring a new link in advance in simulation nodes at two ends of the new simulation link by adopting memif modes through a vector data packet processing component; (2) When the simulation time reaches the planned switching time, activating a corresponding memif interface in the simulation nodes at two ends of the new simulation link through the vector data packet processing component, and starting the new link; (3) And deleting the corresponding memif interfaces in the simulation nodes at the two ends of the old simulation link through the vector data packet processing component, removing the old link, and finishing smooth switching.
6. The method for high-speed processing and forwarding simulation of a spatial network according to claim 1, wherein the automatic data collection method in step S5 is implemented by loading a data capture plug-in for a vector data packet processing component and designating a capture interface in a target simulation node, setting a data packet filtering rule to screen target data, setting a collection termination condition, and when the captured data reaches the termination condition, the plug-in automatically exports the data to a designated file, and obtaining a network simulation result by analyzing the file.

Description

High-speed processing and forwarding simulation method for space network Technical Field The invention relates to the technical field of network simulation, in particular to a high-speed processing and forwarding simulation method of a space network. Background With the development of low-orbit satellite constellation, ground-air integrated communication system and space internet, space network is gradually becoming an important form of new generation global information infrastructure. The space network has the characteristics of strong topology dynamic property, wide node distribution during link propagation, and the like, which puts strict requirements on communication protocols, resource scheduling and network performance. In the theoretical design and deployment verification stage, constructing a set of efficient simulation platform is an important means for promoting the space network research and engineering practice. Particularly, in typical scenes involving large-flow forwarding, high-frequency path switching and the like, the simulation platform not only needs to support flexible topology modeling, but also has to have the capability of processing high-speed communication flows so as to truly reflect the operation characteristics of the system in a complex space environment. However, currently mainstream space network simulation tools (such as CORE, mininet, etc.) and single-machine network simulation methods based on virtualization technologies which are relatively commonly used at present have complete functions in terms of protocol logic modeling and control flow verification, but most of them are based on traditional kernel protocol stacks, which are based on a design concept of single-packet processing, and each processing of a data packet usually needs to trigger a system call once, enter a kernel mode from a user mode, and return to a user space, and this process involves frequent context switching and system scheduling, thereby significantly increasing processing delay. And secondly, the data packet needs to be subjected to multiple memory copying and interrupt notification between the processing logic received by the network card and the user mode application, so that the memory bandwidth is further wasted. For example, the connection mode of virtual network links provided by the Veth device between virtual satellite nodes adopted in the link dynamic regulation method for the large-scale satellite network simulation platform with the application number of CN202411175141.7, and the Overlay network implemented based on VXLAN encapsulation in the virtual network data communication interaction method and system based on the container technology with the application number of CN201911241576.6 are all the forwarding modes, and the data packet undergoes multiple conversions between the user mode and the kernel mode, which negatively affects both processing delay and data packet throughput. In the current space network communication, a laser communication link is gradually becoming a core bearing means for inter-satellite and inter-satellite communication. Laser communication represents a significant advantage over conventional radio frequency links in terms of high capacity low latency transmission. Currently, the transmission rate of the mainstream low-orbit satellite laser link has been increased from hundreds of Mbps to several Gbps, and part of new generation inter-satellite communication systems have evolved even toward the 10 Gbps or even 100Gbps level. Taking the most representative starlink (star chain) at present as an example, the total number of the in-orbit satellites reaches over 9000, and meanwhile, the transmission rate reaches 100Gbps by adopting a laser inter-satellite link technology. If such a satellite communication constellation with large scale and extremely high transmission rate is simulated, higher data packet processing requirements are also put forward for space network simulation, and as the current mainstream space network simulation tool relies on the traditional kernel protocol stack, the efficiency of processing small packet flows and high-frequency throughput is extremely low, the processing capacity is often less than millions of Packets Per Second (PPS), and the requirements for laser link simulation cannot be met far. I.e., using a kernel protocol stack, has significant limitations in terms of packet forwarding performance, link dynamic response capability, and large-scale node deployment. The processing mechanism depends on frequent system call and memory copy, which is very easy to form performance bottleneck especially in the scenes of simulating inter-satellite large bandwidth links or ground data set transfer and the like. Vector packet processing (Vector Packet Processing, VPP) enables vectorized batch processing of data packets, thereby significantly improving packet processing capacity and reducing latency. However, if it is to be applied to the space n