CN103825761A - On-board router simulation method of delay-tolerant network - Google Patents

Abstract

The present invention proposes an on-board router emulation method of a delay tolerant network, comprising: 1. Combining CCSDS and DTNRG to construct an on-board router model, and verifying feasibility and effectiveness in OPNET; 2. Realizing virtualization based on the SITL interface of OPNET Conversion of data streams and real data streams to verify gateway processing capabilities; 3. Use STK to obtain interstellar and star-to-earth link data in the network topology, and import simulation software to simulate interstellar networks; 4. Delay and error codes for network damage meters rate and other related settings to simulate space links. Therefore, the present invention has the following advantages: the integration of the existing space communication protocol system has proposed the router model on the star, and some preparatory work has been done for the system-level realization of my country's DTN protocol system and even the key technology research of the space information network; the combination of software and hardware is more real The environment of the satellite network is simulated, which is helpful to verify the applicability of the on-board router to the space environment.

Description

A Simulation Method of On-Star Router for Delay Tolerant Network

the

technical field

The invention belongs to the two fields of satellite network technology and computer technology, and specifically relates to a method for simulating an on-board router of a delay-tolerant network.

Background technique

At present, my country's aerospace industry has entered a period of rapid development. The number and types of space vehicles launched by my country have increased to a large extent compared with the past. The rapid development of aerospace technology and space applications has put forward an urgent need for space network research. The great success of the terrestrial Internet makes people naturally think of moving the working mode of the terrestrial Internet to the space network. The advantages of the terrestrial Internet IP protocol system are high technical maturity, greatly reduced space cost, and easy upgrade. In 2001, the Goddard Space Center of the United States launched a research project named OMNI [1], which mainly studies the realization of space communication schemes using the ground commercial IP protocol. Although the space IP protocol system can basically meet the information transmission between the ground and low-Earth orbit spacecraft, the TCP protocol is based on end-to-end retransmission protocol, which needs to assume that the transmission delay is very small, which is inconsistent with deep space communication. Furthermore, ground-based routing protocols implemented in a hierarchical manner are not suitable for the operational environment of deep space communications.

Established in 1982, the Consultative Committee on Space Data Systems (CCSDS) has issued a series of recommendations for space links from the physical layer to the application layer. According to the characteristics of the space environment, CCSDS improves the ground standard TCP/IP protocol accordingly, and develops a set of space communication protocol specifications (SCPS) [2] covering the network layer to the application layer, which comprehensively solves the problem of space information transmission. According to the characteristics of deep space communication, the CCSDS protocol also proposes corresponding solutions, such as the CCSDS file transfer protocol CFDP [3]. However, SCPS does not propose a specific routing algorithm for the deep space communication environment where communication resources are very precious; reliable transmission still adopts the mode of establishing a connection first and then transmitting data; the selective retransmission mechanism is still based on end-to-end retransmission. The CFDP protocol is limited to file transfer applications, the solution is not thorough enough, lacks more perfect application services, and the system structure considered from the perspective of interstellar interconnection is not complete.

Vint Cerf, one of the engineers and network pioneers at NASA's Jet Propulsion Laboratory (JPL) in 1998, started research on the Interplanetary Network (IPN), the basic idea of which is to make the earth and spaceships far away Data communication between nodes can be simplified as it happens between two nodes on the earth. Relevant personnel later developed and established the Internet Society Special Interest Group, also known as IPNSIG[4]. However, IPNSIG encountered a problem that there is no such interstellar network that can be tested, and the establishment of an interstellar network is very expensive, so some people began to study how to embody the concept of IPN in land applications, especially sensor networks. The hypothetical IPN network has many common features, and the sensor network experiment is easy to carry out (there is a material basis for the experiment). For this reason, the IRTF research group (Internet research task f roce) established a new working group to find a more general delay-tolerant network. This working group is called DTNRG (DTN research group) [5], which is the current DTN architecture and It is the main public organization for protocol research, and has submitted multiple drafts and standards including the DTN architecture [6] to the IETF. In order to solve the problem of reliable transmission in the deep space environment, JPL submitted a protocol draft to support DTN network in December 2002. The Lieklider transmission protocol [7] replaces the IP protocol and the TCP protocol. In early 2004, DARPA (Defence Advanced Research Projects Agency) under the US Department of Defense proposed Disruption Tolerant Networking, also referred to as DTN [8]. Therefore, sometimes DTN is used not only to represent a delay-tolerant network, but also to represent a delay-interruption-tolerant network.

In recent years, the idea of combining the CCSDS protocol system with the space IP protocol system has also emerged, that is, the CCSDS proposal can still be used at the data link layer, such as packet telemetry, packet remote control, AOS, Proximity, etc., and the network layer uses IP and its Extended technology, transport layer and application layer use commercial standard protocol or CCSDS protocol. This solution has relatively flexible protocol configuration capabilities, but it does not fundamentally eliminate the inherent defects of the space IP protocol system and the current CCSDS protocol system in deep space communication. The adaptability of the protocol stack is weak, and there are still many challenges .

The three protocol systems of IP, CCSDS, and DTN are not pure technological evolution, but interdependent and coexistent. The mature technology of terrestrial Internet TCP/IP and the application verification of space IP provide a clear direction for the improvement suggested by CCSDS. The protocol system combining CCSDS and IP has become a basic solution to give full play to the huge advantages of TCP/IP and meet the requirements of space communication. way. The DTN protocol architecture can solve the problem of reliable transmission under heterogeneous conditions of space environment by integrating the first two layered protocols and its own protocol. However, the DTN protocol has obvious differences compared with the former two: first, DTN does not assume that there is an end-to-end path between the sender and the receiver, and the package is delivered by store and forward; second, DTN introduces the so-called "package layer ( Bundle Layer)" serves as an overlay layer connecting different constrained networks, and nodes adopting this overlay rely on sending asynchronous messages called "packages" to communicate. The wrapper layer provides similar functionality to an Internet gateway, but it focuses on virtual message forwarding rather than packet switching.

At present, the communication equipment on the satellite generally only realizes the function of forwarding but not the function of routing. Satellite transponders are divided into transparent transponders and regenerative transponders. The transparent transponder has functions such as signal amplification and frequency down conversion; the regenerative transponder has functions such as radio frequency beam switching, demodulation-remodulation, baseband switching, multiple access mode conversion, etc., to reduce transmission error rate, improve efficiency, eliminate interference, Reduce transmission delay and improve switching performance. OMNI[1] was the first to put routers on satellites. It directly selects and combines existing commercial communication protocols to adapt to the operating environment and communication requirements of space missions, and reduces the cost of space missions by using off-the-shelf software and hardware resources. It only needs to add a channel error correction coding/decoding module behind the radio frequency equipment to connect with commercial routers, and directly transmit IP packets carrying user data on the network. In 2002, the U.S. intelligence agency, the Department of Defense, and NASA jointly launched the Transformational Communications Study (TCS). Eighteen months later, the joint research team formulated the Transformational Communications Architecture (TCA) Baseline 1.0 [9], the most famous of which is Mo Due to the TSAT plan, it is expected to use all laser cross links and TSAT routers to build a space backbone network. Although the project was cancelled, the overall structure of TCA has basically remained unchanged [10-11]. In 2007, the U.S. Department of Defense included the "Space Internet Router" (IRIS) program in the financial budget for joint capability technology demonstration and verification. IRIS JCTD further opened up the operation of this new network-centric capability through a router carried on a geosynchronous communication satellite. Concepts, tactics, techniques and methods, and experimental results can better point out the capabilities of this next-generation satellite communication network [12-15]. The biggest advantage of IRIS is to simplify communication between satellites and reduce the delay of audio transmission. Under the traditional satellite transmission method, even if the flagship and the destroyer are only a few miles apart, the communication between them needs to be passed several times between different satellites and ground networks to reach the destination. The IRIS program can send IP data packets between satellites in space, and the working principle is similar to the Internet on the ground, which reduces the delay of data transmission, saves satellite communication bandwidth, and provides higher network flexibility. Because the technology can improve the efficiency and reduce the cost of military satellite communications, less equipped and smaller forces will be able to use the functions of the central network in a meaningful sense for the first time. IRIS promises better adaptability to traffic, lower cost per byte, and higher throughput for poorly equipped troops. the

It can be seen that the router on the star is still in the research stage and has not become a mature load on the star. The goal of on-board router simulation for delay-tolerant networks is to verify the applicability of on-board routers to the space environment. The difficulty lies not in the router itself, but in the authenticity of the satellite network environment simulation. At present, the simulation methods of satellite network and its network equipment at home and abroad mainly include:

(1), software-based network simulation. Use network simulation software to model and analyze network equipment and network environment. Commonly used are OPNET and NS2 software. OPNET supports accurate modeling of devices, links, and protocols at all levels of the network, and can accurately analyze the performance and behavior of complex networks, and can truly reflect the network environment. It is often used in engineering and commercial applications. The simulation of NS2 is relatively abstract, which is suitable for the research of protocol algorithms, and is generally widely used in the field of scientific research. In addition, there is ONE software specifically for delay-tolerant networks. ONE is open source software. From the perspective of network simulation, ONE can simulate motion models and DTN routing. From a functional point of view, ONE integrates a large number of specific implementations of motion models and routing algorithms, and is easy to expand, but it is difficult for ONE to simulate devices in detail.

(2), semi-physical simulation. Use software (such as OPNET) to simulate real routers, switches and other network equipment. The network environment is built with software and hardware, that is, use PC+ simulation software to simulate the router on the star, and use other software and hardware forms to simulate the network environment. (including source and sink) for simulation. This method is more complicated to build a platform and requires the simulation software to have a corresponding conversion interface, but it can more realistically simulate the space network.

(3) On-star router PC simulation. Under the PC platform, the communication software with the same protocol stack structure as the on-board communication equipment is developed and operated. On the ground, the software and hardware are used to build a network simulation environment and the router under the PC platform is tested. This method can simulate the space network more realistically, but the simulation cycle is very long, and it is close to the verification test system on the ground. Although the equipment development platform is on the PC, the overall framework of the entire communication software does not change during the transplantation process.

Contents of the invention

Above-mentioned technical problem of the present invention is mainly solved by following technical scheme:

An on-board router emulation method of a delay tolerant network, characterized in that it comprises:

The steps of on-board router modeling and simulation: establish an on-board router model; combine the network hierarchy defined by CCSDS and DTNRG to build a protocol model suitable for on-board routers, and use software simulation to verify the feasibility and effectiveness of the model Verification; the on-star router modeling and software simulation are carried out based on OPNET, including node domain modeling, process domain modeling, and external system domain modeling;

Steps of conversion of simulation data and real data: convert the verified virtual data flow in the software simulation environment into real data flow, so as to verify the processing ability of the gateway to the real flow. Since the gateway is still implemented in the form of software, the specific The implementation process includes the conversion interface from simulation data to real data and the conversion interface from real data to simulation data; this step is carried out based on the OPNET SITL interface, including the sub-steps of time synchronization, data packet interception and data packet conversion;

Satellite topology simulation steps: Generate constellation network operation data according to the orbital parameters of each satellite, obtain interstellar and satellite-ground link on-off time, link transmission delay and other data, and analyze the network topology change data; in the simulation software Input these data in, construct the topological model of the interstellar network, and on this basis, simulate the satellite-ground and inter-satellite communication protocols in the interstellar network; this step is based on STK, and different satellites are determined by defining the parameters of the satellite in STK, So as to simulate the satellite network topology;

The simulation steps of the space link: simulate the satellite-ground and inter-satellite links, provide a simulation interface for parameters such as delay and bit error rate, and observe the impact of different links on network performance through the configuration of these parameters; specifically: use iTrinegy network damage The instrument simulates the space link; the computers that simulate satellite routers are connected to the network damage instrument through Ethernet, and in the .xml configuration file of the network damage instrument, different delay and bit error rate parameters are set to simulate different Space link parameters and link continuity.

In the above-mentioned on-board router simulation method of a delay-tolerant network, in the step 1, the specific implementation process of the three modeling methods is as follows:

Modeling 1: node domain modeling;

Define the addition of an additional Bundle layer between the transport layer and the application layer, the transport layer adaptively selects the SCPS-TP or UDP protocol, and the link layer uses the IP OVER CCSDS AOS architecture

Modeling 2: process domain modeling; according to the node domain model, the process models that need to be described and implemented include Bundle, SCPS-TP and IP OVER CCSDS AOS;

Bundle process model: Bundle initial transmission, forwarding procedure, receiving procedure, scheduling procedure,

Forwarding restricted procedures, forwarding failure procedures, deletion procedures and local delivery procedures, and processing actions including consent to custody, custody release and custody failure;

SCPS-TP process model: Similar to the implementation of the TCP protocol, the implementation of the SCPS-TP protocol establishes two process models, one is the management process, and the other is the processing sub-process. The management process is the interface process with the upper and lower layers, and the sub-process is The actual processing process of the SCPS-TP protocol; the SCPS-TP management state process is responsible for receiving all commands sent from the application layer and the lower layer, and enters different states according to the commands; the SCPS-TP protocol processing sub-process is responsible for the SCPS-TP protocol from establishing a connection From the whole process of closing the connection, how to deal with the received data segments at different times, use the connection state to represent these different events, and use the state transition diagram to illustrate how SCPS-TP handles various data in various states how the segment is handled;

IP OVER CCSDS AOS process model: The IP OVER CCSDS AOS management process first enters the initialization state to assign state variables, and then enters the idle state to start processing various events; after entering the idle state, the process begins to wait for events; when the event is a stream interruption, The process first reads the inflow port; if it is from the upper layer IP process, the process enters the from_upper state to perform AOS frame encapsulation on the IP data; if it is from the lower MAC process, the process enters the from_lower state to decapsulate the AOS frame and reassemble the IP data; When Polling_MPDU is self-interrupted, the process enters the from_upper state to judge the frame length of the AOS frame currently undergoing encapsulation operation. If it is less than the specified fixed length, then generate idle data and insert it into the MPDU packet area to ensure that the AOS frame is a fixed length, and then Call the IP OVER CCSDS AOS periodic sending subprocess to send the AOS frame;

Modeling 3: external system domain modeling; external system modules are mainly used for co-simulation; co-simulation is a simulation run jointly by multiple simulation tools, which consists of Modeler model, co-simulation code and external code; Modeler model contains a or multiple external system definitions, and the external code can be a complete emulator or a communication mechanism connected to other operating system processes. The first is to write a custom SITL function, and then select the external file containing the SITL conversion function in the SITL gateway module; what the SITL part needs to implement is the support for Bundle, SCPS-TP and IP OVER CCSDS AOS protocols.

In the above-mentioned on-star router emulation method of a delay tolerant network, in the step 1, the specific establishment method of the Bundle process model is as follows:

At the beginning of the process, it first passes through the pre-initialization state until other processes are ready, and then performs some necessary initialization work, mainly initializing the static routing table and other state variables and registering the Bundle node; after the initialization is completed, it enters the idle state, which is mainly Determine the type of interruption; if it is a flow interruption FROM_LOW, it means that the data is received from the lower layer process. According to the port number of the transport layer, it is determined whether it is application data or Bundle data. If the port number of the transport layer is 550, it means that it is Bundle data. Send the storage success signal, and then wait for the arrival of the delivery interrupt signal; otherwise, it is the application data. At this time, the corresponding transport layer connection needs to be opened first. From the previously designed comprehensive gateway protocol stack model, here is mainly a UDP or SCPS-TP connection ; After the connection is established, Bundle packaging is performed, and then inserted into the storage queue to wait for storage and forwarding;

When the source endpoint of the bundle is far away from the destination endpoint and there is only intermittent connection, and the duration of the connection is not enough to transmit the entire bundle, at this time, as long as the bundle processing flag field of the basic block does not indicate that the bundle cannot be segmented, the bundle can be processed Segmentation processing, each fragment is encapsulated into a new Bundle for transmission; each fragment has the same source endpoint ID and creation timestamp, indicating that they come from the same Bundle; specific segmentation forms include active segmentation and Passive segmentation, the method of active segmentation is adopted here, that is, the Bundle process actively divides the application data into many smaller blocks, and then encapsulates each block into a Bundle for transmission; at this time, the destination node needs to receive The application data of these small pieces is extracted from the Bundle and reassembled to obtain the complete application data; the reorganization of the Bundle is based on the indication of the segment offset and payload length, and all received segments with the same endpoint ID and creation timestamp as the current segment If the length of the formed byte array is equal to the total length of the application data unit indicated in the basic block, then the application data unit reassembled with this byte array replaces the current segment load, and the load of each segment is the reassembled application data unit a subset of data units;

If it is a self-interruption FROM_ROUTER, call the routing module to calculate the next hop node at this time;

If it is a self-interrupted STATE_IND, at this time, the received Bundle data packet is decapsulated and reassembled, and then delivered to the destination node;

If it is a self-interrupted RETRANS, take out the Bundle data packet in the storage queue and forward it to the next hop node.

In the above-mentioned on-board router emulation method of a delay tolerant network, in the step of converting the simulated data and real data, the specific methods of time synchronization, data packet interception and data packet conversion are as follows:

Since the SITL module adds two special models, the gateway node and the link model, it is necessary to import the data packets running in the real protocol stack into the virtual simulation environment, including:

Step 1: Synchronization of time;

Synchronize the simulation time with the system time;

Step 2: interception of data packets;

Use the WinPcap auxiliary packet capture tool to capture the original data packet, and then analyze the protocol type according to the data packet header, and then realize packet filtering;

Step 3: conversion of data packets;

When the SITL module receives a data packet, whether it is a real packet from a physical system or a virtual packet from a simulation system, it first tries to judge the packet format and the protocol it belongs to, and finds the appropriate conversion by matching with the conversion function function, and call it to convert the data packet;

When the simulation starts, first SITL calls the initialization function to load all conversion/detection function pairs; if a data packet arrives, the WinPcap packet capture tool intercepts it and sends it to the SITL gateway node, and SITL calls the conversion entry function to convert the data packet, usually In this case, the conversion function of the basic packet format is first called; if the data packet is nested with other types of data packets, then the conversion function will call the conversion function of the data field to convert the next layer of data packets until the conversion to the set The topmost protocol specified.

In the above-mentioned on-board router simulation method of a delay tolerant network, the simulation step of the satellite topology is to determine different satellites by defining the parameters of the satellite in the STK, thereby performing the simulation of the satellite network topology, including

Setting 1: Basic property setting of the satellite;

Set the basic attributes of the satellite: orbit, attitude, trajectory breakpoint, quality, description, graphics, attributes, display time, constraint time;

Setting 2: Define satellite orbit settings;

Determine the size, shape, orbital orientation and position of the satellite orbit by setting orbital parameters, mainly including: semi-major axis, apogee radius, apogee height, return period, daily orbital circles, eccentricity, perigee radius, Orbital inclination, right ascension of ascending node, argument of perigee, true anomaly, mean anomaly, partial perigee, ascending angular distance, ascending node time, and perigee time.

The on-board router emulation method of the above-mentioned delay tolerant network also includes a step of testing the network performance, randomly selecting the iperf test module or the VLC_media_player test module to carry out:

Random test 1: Use the iperf test module for network performance testing, the steps are:

Step 6.11, iperf software configuration: client configuration: send a 1KByte UDP packet to port 5001 of the server at a rate of 2Mb/s for a duration of 200s, and configure the server to listen to port 5001 of UDP;

Step 6.12, configuration of the OPNET simulation environment: enable the DTN router function, set the fragment size to 1470 bytes, set the packet lifetime to 300s, set port 5001 to monitor UDP on the two routers, and set the aggregation layer of the bundle to UDPCL ;

Step 6.13, start the iperf server on the device in sequence, as well as opnet software, iperf client, and save the experimental results;

Random test 2: use VLC_media_player software for actual business test, the steps are:

Step 6.21, run two VLC_media_player processes on both sides of the communication (this software does not support two-way communication, that is, one process can only be used as a client or server), as a client and a server for two-way communication respectively, and the communication ports of the two servers are respectively Configured as 10000 and 10001, open the corresponding port of the DTN gateway in the OPNET environment;

Step 6.22, configure the network damage tester, set the delay to 100ms, 1s and 10s respectively;

Step 6.23, test whether the VLC_media_player software can be used for two-way communication in the case of a high-delay link between the two communication parties;

Step 6.24, unplug the Ethernet cable, and then record into the microphone of the communication device; then plug in the Ethernet cable again, and test whether the recorded content can be heard on the receiving device.

Therefore, the present invention has the following advantages: integrate the existing space communication protocol system to design the on-board router, and do some preparatory work for the system-level realization of the DTN protocol system and even the key technology research of the space information network in my country; software and hardware Combined with a more realistic simulated satellite network environment, it is helpful to verify the applicability of the on-board router to the space environment.

Description of drawings

Accompanying drawing 1 is the protocol system of the integrated gateway of the embodiment in the present invention.

Accompanying drawing 2 is the basic processing procedure of the Bundle protocol of the embodiment in the present invention.

Accompanying drawing 3 is the data message format of the SCPS-TP agreement of the embodiment among the present invention.

Accompanying drawing 4 is the idle state control flow of the embodiment in the present invention.

Accompanying drawing 5 is DTN router test result among the present invention.

Accompanying drawing 6 is a schematic flow chart of the method of the present invention.

Detailed ways

The technical solutions of the present invention will be further specifically described below through the embodiments and in conjunction with the accompanying drawings.

Example:

The technical solution of the present invention will be described in detail below in conjunction with the drawings and embodiments.

Step 1. On-board router modeling and simulation based on OPNET.

(1) Node domain modeling.

The job of the node field is to define the internal structure of the node. The protocol architecture shown in Figure 1 is adopted in OPNET. An additional Bundle layer is added between the transport layer and the application layer. The transport layer adaptively selects the SCPS-TP or UDP protocol, and the link layer uses the IP OVER CCSDS AOS architecture. When communicating in space, there is generally no need to consider the complex situation of the application layer gateway.

In the node model, the interconnected blocks that make up the node model become modules, and each module contains a set of inputs and outputs, a state memory, and a function that computes an output based on the inputs and state memory. The input and output of the module are divided into packet flow and statistics. The module can output data packets from its output packet flow or receive data packets from the input packet flow, and can also output values from output statistics or input from input statistics. value.

(2) Process domain modeling.

According to the node domain model, the process models that need to be described and implemented include Bundle, SCPS-TP and IP OVER CCSDS AOS.

1. Bundle process model.

The basic processing flow of the Bundle protocol is shown in Figure 2, mainly including Bundle initial transmission, forwarding procedures, receiving procedures, scheduling procedures, forwarding restricted procedures, forwarding failure procedures, deletion procedures, and local delivery procedures. In addition, there are some processing actions, such as consent to safekeeping, safekeeping release, and safekeeping failure. For the specific processing procedures of various procedures, please refer to the protocol specification.

At the beginning of the process, it first passes through the pre-initialization state until other processes are ready, and then performs some necessary initialization work, mainly initializing the static routing table and other state variables and registering the Bundle node. After the initialization is completed, it enters the idle state, which is mainly to judge the type of interrupt. If it is a flow interruption FROM_LOW, it means that data is received from the lower layer process, and it is determined whether it is application data or Bundle data according to the port number of the transport layer. If the port number of the transport layer is 550, it means that it is Bundle data. It needs to send a storage success signal to the storage node, and then wait for the arrival of the delivery interruption signal; otherwise, it is application data. At this time, the corresponding transport layer connection needs to be opened first. According to the previous design It can be known from the integrated gateway protocol stack model that here is mainly a UDP or SCPS-TP connection. After the connection is established, the Bundle is encapsulated, and then inserted into the storage queue to wait for storage and forwarding.

When the source endpoint of the bundle is far away from the destination endpoint and there is only intermittent connection, and the duration of the connection is not enough to transmit the entire bundle, at this time, as long as the bundle processing flag field of the basic block does not indicate that the bundle cannot be segmented, the bundle can be processed Fragmentation processing, each fragment is encapsulated into a new Bundle for transmission. Fragments have the same source endpoint ID and creation timestamp, indicating that they originate from the same bundle. The specific segmentation forms include active segmentation and passive segmentation. Here, the active segmentation method is adopted, that is, the Bundle process actively divides the application data into many smaller blocks, and then encapsulates each block into a Bundle for transmission. . At this point, the destination node should extract these small pieces of application data from the received Bundle and reassemble to obtain complete application data. The reassembly of Bundle is to connect all received segment payloads with the same endpoint ID and creation timestamp as the current segment according to segment offset and payload length. If the length of the composed byte array is equal to the total length of the application data unit indicated in the basic block, then the application data unit reassembled with this byte array replaces the payload of the current segment, and the payload of each segment is a subsection of the reassembled application data unit. set.

If it is a self-interrupt FROM_ROUTER, call the routing module to calculate the next hop node at this time.

If it is a self-interrupted STATE_IND, at this time, the received Bundle data packet is decapsulated and reassembled, and then delivered to the destination node.

If it is a self-interrupted RETRANS, take out the Bundle data packet in the storage queue and forward it to the next hop node.

2. SCPS-TP process model.

Similar to the implementation of the TCP protocol, the implementation of the SCPS-TP protocol establishes two process models, one is the management process, and the other is the processing sub-process. The management process is the interface process with the upper and lower layers, and the sub-process is the actual process of the SCPS-TP protocol. process. The SCPS-TP management state process is responsible for receiving all commands sent from the application layer and the lower layer, and enters different states according to the commands:

1) OPEN: open a SCPS-TP connection;

2) SEND: Send a data packet to the established SCPS-TP connection;

3) RECEIVE: SCPS-TP requests the upper layer to inject data packets;

4) CLOSE: After sending the data, end the SCPS-TP connection;

5) ABORT: abnormally interrupt the SCPS-TP connection;

6) SEG_REV: receive data packets;

7) STATUS: SCPS-TP status indication.

The SCPS-TP protocol processing sub-process is responsible for the entire process of the SCPS-TP protocol from establishing a connection to closing the connection. What kind of processing is used to receive data segments at different times? These different events can be represented by the state of the connection, and the state can be used Transition diagram to illustrate how SCPS-TP handles various data segments in various states:

1) LISTEN: Listen for connection requests from non-local SCPS-TP ports;

2) SYN-SENT: Wait for the corresponding connection request after sending the connection request;

3) SYN-RECEIVED: After receiving and sending a connection request, wait for the other party to confirm the connection request;

4) ESTABLISHED: Indicates that a connection has been established;

5) FIN-WAIT-1: Wait for the remote SCPS-TP connection interruption request, or the confirmation of the previous connection interruption request;

6) FIN-WAIT-2: Waiting for connection interruption request from remote SCPS-TP;

7) CLOSE-WAIT: Waiting for a connection interruption request from a local user;

8) CLOSING: Waiting for the remote SCPS-TP to confirm the connection interruption;

9) LAST-ACK: Waiting for the confirmation of the original connection interruption request sent to the remote SCPS-TP;

10) TIME-WAIT: Wait for enough time to ensure that the remote SCPS-TP receives an acknowledgment of the connection interruption request;

11) CLOSED: Indicates that there is no current connection status.

The data packet format of the SCPS-TP protocol is shown in Figure 3:

The ields field is the 20-byte standard TCP packet header; the SACK-Permitted Option field and the SACK Option field are the SACK permission option field and the SACK option field used by the SACK mechanism; the SCPS Capabilities Option field and the SNACK Option field are related to the SNACK mechanism field.

3. IP OVER CCSDS AOS process model

The IP OVER CCSDS AOS management process first enters the initialization state to assign state variables, and then enters the idle state to start processing various events, as shown in Figure 4. When entering the idle state, the process begins to wait for events. When the event is a stream interruption, the process first reads the inflow port. If it is from the upper layer IP process, the process enters the from_upper state to perform AOS frame encapsulation on the IP data; if it is from the lower layer MAC process, the process enters the from_lower state to decapsulate the AOS frame and reassemble the IP data. When the event is self-interrupting Polling_MPDU, the process enters the from_upper state to judge the frame length of the AOS frame currently undergoing encapsulation operation. If it is less than the specified fixed length, then generate idle data and insert it into the MPDU packet area to ensure that the AOS frame is a fixed length , and then call the IP OVER CCSDS AOS periodic sending subprocess to send the AOS frame.

(3) External system domain modeling.

External system blocks are mainly used for co-simulation. Co-simulation is a simulation that is run jointly by multiple simulation tools, and is composed of Modeler model, co-simulation code and external code. The Modeler model contains one or more external system definitions, and the external code can be a complete simulator or a communication mechanism connecting other operating system processes, and the two are connected as one through the co-simulation code. The realization of the SITL part in the embodiment of the present invention belongs to the modeling of the external system domain. Firstly, a self-defined SITL function is written, and then an external file containing the SITL conversion function is selected in the SITL gateway module. What SITL needs to implement is the support for Bundle, SCPS-TP and IP OVER CCSDS AOS protocols.

Step 2. Conversion of simulation data and real data based on OPNET SITL interface.

The SITL module adds two special models, the gateway node and the link model. In the real network, we use the TCP/IP protocol stack to forward and communicate data streams, while the protocol stack in the simulated network is constructed virtually. Different data structures make it impossible for real devices and simulated networks to communicate directly. To import the data packets running in the real protocol stack into the virtual simulation environment, three problems need to be solved: time synchronization, packet interception and data packet conversion.

(1) Synchronization of time.

Generally, the time synchronization of SITL simulation is ensured by synchronizing the simulation time with the system time, which can be completed through simple settings.

(2) Interception of data packets.

The WinPcap auxiliary packet capture tool is used to capture the original data packet, and on this basis, the type of the protocol is analyzed according to the data packet header, and then the packet filtering is realized.

(3) Conversion of data packets.

When the SITL module receives a data packet, whether it is a real packet from a physical system or a virtual packet from a simulation system, it first tries to judge the packet format and the protocol it belongs to, and finds the appropriate conversion by matching with the conversion function function, and call it to convert the packet.

When the simulation starts, first SITL calls the initialization function to load all the conversion/detection function pairs. If a data packet arrives, the WinPcap packet capture tool intercepts it and sends it to the SITL gateway node. SITL calls the conversion entry function to convert the data packet. Usually, the conversion function of the basic packet format is called first. If the data packet is nested with other types of data packets, then the conversion function will call the conversion function of the data field to convert the next layer of data packets until it is converted to the set top-level protocol.

Step 3, simulation of satellite network topology based on STK.

By defining the parameters of the satellite in STK to determine different satellites, the simulation of the satellite network topology is carried out.

(1) The basic attribute setting of the satellite.

Set the basic properties of the satellite: orbit, attitude, track breakpoint, quality, description, graphics, attributes, display time, constraint time, etc.

(2) Define satellite orbit settings.

Determine the size, shape, orbital orientation and position of the satellite orbit by setting orbital parameters, mainly including: semi-major axis, apogee radius, apogee height, return period, daily orbital circles, eccentricity, perigee radius, Orbital inclination, right ascension of ascending node, argument of perigee. True anomaly, mean anomaly, partial perigee, ascending angular distance, time of passing ascending node, time of passing perigee.

Step 4, the simulation of the space link based on the network damage instrument.

The space link is simulated by iTrinegy network damage instrument. Connect the computer that simulates the satellite router to the network damage tester with Ethernet, and set different delay and bit error rate parameters in the .xml configuration file of the network damage tester to simulate different space link parameters and link on and off. the

Step five, network performance test.

Adopt two methods to test network performance in the embodiment of the present invention: adopt

(1) Use iperf software for network performance testing, the steps are:

1. iperf software configuration. Client configuration: Send a 1KByte UDP packet to port 5001 of the server at a rate of 2Mb/s for a duration of 200s. The server is configured to listen on UDP port 5001.

2. Configuration of OPNET simulation environment. Turn on the DTN router function, set the fragment size to 1470 bytes, set the packet lifetime to 300s, set the port 5001 to monitor UDP on the two routers, and set the aggregation layer of the bundle to UDPCL.

3. Start the iperf server, opnet software, and iperf client on the device in sequence. Save the experiment results. Figure 5 shows the test results of the DTN router. The results show that although there is no end-to-end connection, the data can still be kept and forwarded, and finally delivered to the server.

(2) Use VLC_media_player software for actual business testing, the steps are:

1. Run two VLC_media_player processes on both sides of the communication (the software does not support two-way communication, that is, one process can only be used as a client or server), as the client and server of two-way communication respectively. The communication ports of the two servers are configured as 10000 and 10001 respectively. Open the corresponding port of the DTN gateway in the OPNET environment.

2. Configure the network damage tester, and set the delay to 100ms, 1s and 10s respectively.

3. Test whether the VLC_media_player software can be used for two-way communication in the case of a high-latency link between the two communicating parties.

4. Unplug the Ethernet cable, and then record into the microphone of the communication device. Then plug in the Ethernet cable again and test whether you can hear what you just recorded on the receiving device.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention belongs can make various modifications or supplements to the described specific embodiments or adopt similar methods to replace them, but they will not deviate from the spirit of the present invention or go beyond the definition of the appended claims range.

Claims (6)

1. A method for emulating the on-star router of a delay tolerant network, characterized in that it comprises: The steps of on-board router modeling and simulation: establish an on-board router model; combine the network hierarchy defined by CCSDS and DTNRG to build a protocol model suitable for on-board routers, and use software simulation to verify the feasibility and effectiveness of the model Verification; the on-star router modeling and software simulation are carried out based on OPNET, including node domain modeling, process domain modeling, and external system domain modeling; Steps of conversion of simulation data and real data: convert the verified virtual data flow in the software simulation environment into real data flow, so as to verify the processing ability of the gateway to the real flow. Since the gateway is still implemented in the form of software, the specific The implementation process includes the conversion interface from simulation data to real data and the conversion interface from real data to simulation data; this step is carried out based on the OPNET SITL interface, including the sub-steps of time synchronization, data packet interception and data packet conversion; Satellite topology simulation steps: Generate constellation network operation data according to the orbital parameters of each satellite, obtain interstellar and satellite-ground link on-off time, link transmission delay and other data, and analyze the network topology change data; in the simulation software Input these data in, construct the interstellar network topology model, and on this basis, simulate the satellite-ground and inter-satellite communication protocols in the interstellar network; this step is based on STK, and determine different satellites by defining satellite parameters in STK , so as to simulate the satellite network topology; The simulation steps of the space link: simulate the satellite-ground and inter-satellite links, provide a simulation interface for parameters such as delay and bit error rate, and observe the impact of different links on network performance through the configuration of these parameters; specifically: use iTrinegy network damage The instrument simulates the space link; the computers that simulate satellite routers are connected to the network damage instrument through Ethernet, and in the .xml configuration file of the network damage instrument, different delay and bit error rate parameters are set to simulate different Space link parameters and link continuity. 2. the on-star router emulation method of a kind of delay tolerant network according to claim 1, is characterized in that, in described step 1, three modeling methods specific implementation processes are as follows: Modeling 1: node domain modeling; Define the addition of an additional Bundle layer between the transport layer and the application layer, the transport layer adaptively selects the SCPS-TP or UDP protocol, and the link layer uses the IP OVER CCSDS AOS architecture Modeling 2: process domain modeling; according to the node domain model, the process models that need to be described and implemented include Bundle, SCPS-TP and IP OVER CCSDS AOS; Bundle process model: Bundle initial transmission, forwarding procedure, receiving procedure, scheduling procedure, Forwarding restricted procedures, forwarding failure procedures, deletion procedures and local delivery procedures, and processing actions including consent to custody, custody release and custody failure; SCPS-TP process model: Similar to the implementation of the TCP protocol, the implementation of the SCPS-TP protocol establishes two process models, one is the management process, and the other is the processing sub-process. The management process is the interface process with the upper and lower layers, and the sub-process is The actual processing process of the SCPS-TP protocol; the SCPS-TP management state process is responsible for receiving all commands sent from the application layer and the lower layer, and enters different states according to the commands; the SCPS-TP protocol processing sub-process is responsible for the SCPS-TP protocol from establishing a connection From the whole process of closing the connection, how to deal with the received data segments at different times, use the connection state to represent these different events, and use the state transition diagram to illustrate how SCPS-TP handles various data in various states how the segment is handled; IP OVER CCSDS AOS process model: The IP OVER CCSDS AOS management process first enters the initialization state to assign state variables, and then enters the idle state to start processing various events; after entering the idle state, the process begins to wait for events; when the event is a stream interruption, The process first reads the inflow port; if it is from the upper layer IP process, the process enters the from_upper state to perform AOS frame encapsulation on the IP data; if it is from the lower MAC process, the process enters the from_lower state to decapsulate the AOS frame and reassemble the IP data; When Polling_MPDU is self-interrupted, the process enters the from_upper state to judge the frame length of the AOS frame currently undergoing encapsulation operation. If it is less than the specified fixed length, then generate idle data and insert it into the MPDU packet area to ensure that the AOS frame is a fixed length, and then Call the IP OVER CCSDS AOS periodic sending subprocess to send the AOS frame; Modeling 3: external system domain modeling; external system modules are mainly used for co-simulation; co-simulation is a simulation run jointly by multiple simulation tools, which consists of Modeler model, co-simulation code and external code; Modeler model contains a or multiple external system definitions, and the external code can be a complete emulator or a communication mechanism connected to other operating system processes. The first is to write a custom SITL function, and then select the external file containing the SITL conversion function in the SITL gateway module; what the SITL part needs to implement is the support for Bundle, SCPS-TP and IP OVER CCSDS AOS protocols. 3. the on-star router emulation method of a kind of delay tolerant network according to claim 2, is characterized in that, in described step 1, the concrete establishment method of Bundle process model is as follows: At the beginning of the process, it first passes through the pre-initialization state until other processes are ready, and then performs some necessary initialization work, mainly initializing the static routing table and other state variables and registering the Bundle node; after the initialization is completed, it enters the idle state, which is mainly Determine the type of interruption; if it is a flow interruption FROM_LOW, it means that the data is received from the lower layer process. According to the port number of the transport layer, it is determined whether it is application data or Bundle data. If the port number of the transport layer is 550, it means that it is Bundle data. Send the storage success signal, and then wait for the arrival of the delivery interrupt signal; otherwise, it is the application data. At this time, the corresponding transport layer connection needs to be opened first. From the previously designed comprehensive gateway protocol stack model, here is mainly a UDP or SCPS-TP connection ; After the connection is established, Bundle packaging is performed, and then inserted into the storage queue to wait for storage and forwarding; When the source endpoint of the bundle is far away from the destination endpoint and there is only intermittent connection, and the duration of the connection is not enough to transmit the entire bundle, at this time, as long as the bundle processing flag field of the basic block does not indicate that the bundle cannot be segmented, the bundle can be processed Segmentation processing, each fragment is encapsulated into a new Bundle for transmission; each fragment has the same source endpoint ID and creation timestamp, indicating that they come from the same Bundle; specific segmentation forms include active segmentation and Passive segmentation, the method of active segmentation is adopted here, that is, the Bundle process actively divides the application data into many smaller blocks, and then encapsulates each block into a Bundle for transmission; at this time, the destination node needs to receive The application data of these small pieces is extracted from the Bundle and reassembled to obtain the complete application data; the reorganization of the Bundle is based on the indication of the segment offset and payload length, and all received segments with the same endpoint ID and creation timestamp as the current segment If the length of the formed byte array is equal to the total length of the application data unit indicated in the basic block, then the application data unit reassembled with this byte array replaces the current segment load, and the load of each segment is the reassembled application data unit a subset of data units; If it is a self-interruption FROM_ROUTER, call the routing module to calculate the next hop node at this time; If it is self-interrupted STATE_IND, decapsulate and reassemble the received Bundle data packet at this time, and then deliver it to the destination node; If it is a self-interrupted RETRANS, take out the Bundle data packet in the storage queue and forward it to the next hop node. 4. the on-star router emulation method of a kind of delay tolerant network according to claim 1, is characterized in that, in the step of the conversion of described emulation data and real data, the synchronization of time, the interception of data packet and the interception of data packet The specific method of conversion is as follows: Since the SITL module adds two special models, the gateway node and the link model, it is necessary to import the data packets running in the real protocol stack into the virtual simulation environment, including: Step 1: Synchronization of time; Synchronize the simulation time with the system time; Step 2: interception of data packets; Use the WinPcap auxiliary packet capture tool to capture the original data packet, and then analyze the protocol type according to the data packet header, and then realize packet filtering; Step 3: conversion of data packets; When the SITL module receives a data packet, whether it is a real packet from a physical system or a virtual packet from a simulation system, it first tries to judge the packet format and the protocol it belongs to, and finds the appropriate conversion by matching with the conversion function function, and call it to convert the data packet; When the simulation starts, first SITL calls the initialization function to load all conversion/detection function pairs; if a data packet arrives, the WinPcap packet capture tool intercepts it and sends it to the SITL gateway node, and SITL calls the conversion entry function to convert the data packet, usually In this case, the conversion function of the basic packet format is first called; if the data packet is nested with other types of data packets, then the conversion function will call the conversion function of the data field to convert the next layer of data packets until the conversion to the set The topmost protocol specified. 5. the on-board router simulation method of a kind of delay tolerant network according to claim 1, is characterized in that, the simulation step of described satellite topology is to determine different satellites by defining the parameter of satellite in STK, thereby carries out satellite Simulation of network topologies, including Setting 1: Basic property setting of the satellite; Set the basic attributes of the satellite: orbit, attitude, trajectory breakpoint, quality, description, graphics, attributes, display time, constraint time; Setting 2: Define satellite orbit settings; Determine the size, shape, orbital orientation and position of the satellite orbit by setting orbital parameters, including: semi-major axis, apogee radius, apogee height, return period, daily orbital circles, eccentricity, perigee radius, orbit Inclination, right ascension of ascending node, argument of perigee, true anomaly, mean anomaly, partial perigee, ascending angular distance, ascending node passing time, passing perigee time. 6. the on-star router emulation method of a kind of delay tolerant network according to claim 1, is characterized in that, also comprises the step of the test of a network performance, randomly selects iperf test module or VLC_media_player test module to carry out: Random test 1: Use the iperf test module for network performance testing, the steps are: Step 6.11, iperf software configuration: client configuration: send a 1KByte UDP packet to port 5001 of the server at a rate of 2Mb/s for a duration of 200s, and configure the server to listen to port 5001 of UDP; Step 6.12, configuration of the OPNET simulation environment: enable the DTN router function, set the fragment size to 1470 bytes, set the packet lifetime to 300s, set port 5001 to monitor UDP on the two routers, and set the aggregation layer of the bundle to UDPCL ; Step 6.13, start the iperf server on the device in sequence, as well as opnet software, iperf client, and save the experimental results; Random test 2: use VLC_media_player software for actual business test, the steps are: Step 6.21, run two VLC_media_player processes on both sides of the communication (this software does not support two-way communication, that is, one process can only be used as a client or server), as a client and a server for two-way communication respectively, and the communication ports of the two servers are respectively Configured as 10000 and 10001, open the corresponding port of the DTN gateway in the OPNET environment; Step 6.22, configure the network damage tester, set the delay to 100ms, 1s and 10s respectively; Step 6.23, test whether the VLC_media_player software can be used for two-way communication in the case of a high-delay link between the two communication parties; Step 6.24, unplug the Ethernet cable, and then record into the microphone of the communication device; then plug in the Ethernet cable again, and test whether the recorded content can be heard on the receiving device.

Images