CN102006301A - Method for controlling movement of stream control transmission protocol (SCTP) association among multiple terminals - Google Patents

Abstract

The present invention discloses a mobile control method for flow control transmission protocol coupling between multiple terminals. In the method, the first terminal establishes a control coupling with the second terminal of the migration target, and the control coupling and the original communication coupling have a fixed existence. The port matching relationship between the second terminal and the opposite terminal is maintained through DAR reconfiguration. The SCTP coupling migration control method proposed by the present invention is compatible with the existing standard-compliant SCTP implementation system, and enables real-time multimedia applications running online to move from one physical terminal to another in a seamless manner, overcoming network layer mobility and transmission discontinuities caused by application layer mobility. In addition, the SCTP coupled migration control process proposed by the present invention neither limits the number of terminals nor imposes special requirements on the functions of the terminals, so it is applicable to mutual migration within a multi-terminal group and cross-local migration.

Description

A mobile control method for flow control transmission protocol coupling between multiple terminals

technical field

The present invention proposes a coupling migration control method based on the Stream Control Transmission Protocol (SCTP, Stream Control Transmission Protocol) technology, completes the seamless migration of real-time multimedia communication between multiple terminals, and supports applications and sessions between physically separated terminals Smooth movement, the technical field is electronics and communication engineering.

Background technique

The mobility of the Internet, or mobile Internet, is the basis for all kinds of new information and communication applications. The biggest difference between mobile applications and fixed applications is mainly manifested in terminal mobile, user mobile, network mobile and application mobile.

(1) Terminal mobility refers to the migration of a terminal used by a user between different access points, mainly through handover at the existing physical layer and Mobile Internet Protocol (MIP, Mobile Internet Protocol) at the network layer to achieve Uninterrupted access of the terminal to the mobile base station after physical location relocation, and communication continuity during and after the terminal network address change.

(2) User mobility refers to the migration of a user between multiple terminals. The session initiation protocol (SIP, Session Initiation Protocol) of the Internet can be used to support the transfer of sessions between different terminals and realize calls similar to telephone services. Transfer, call forwarding and multi-device simultaneous vibration functions.

(3) Network mobility refers to the collective synchronous migration of multiple terminals in the same subnetwork between different access points, mainly through Internet network mobility (NEMO, Network Mobility) and MIP technology to achieve access to routers or Uninterrupted switching of gateways between different host networks.

(4) Application mobility refers to the information source or sink of an application, which migrates among multiple terminals at the request of one user or multiple users. Service Architecture) technology to achieve uninterrupted switching.

The above four types of mobile, in the current technical implementation methods, involve handover at the physical layer, MIP and NEMO at the network layer, and SIP technology at the application layer. In recent years, people have proposed the Host Identity Protocol (HIP, Host Identity Protocol), which can be used to solve the performance degradation problem of terminal mobility. HIP is the cushion layer between the network layer and the transport layer, and is mainly applicable to the mobility of a single terminal. In short, up to now, the Internet technology lacks a mobility control protocol for the transport layer to support migration of applications between different terminals.

SCTP, located at the transport layer of the TCP/IP protocol suite, establishes an association between two communication endpoints, and provides media stream and control stream bearer services for Internet-based applications. At first, the application goal of SCTP was to carry traditional telecommunication network signaling through the Internet, and according to the multiplexing characteristics of signaling, a multiple streaming control method was proposed, and a multi-homing (multiple streaming) control method was standardized for high reliability requirements. homing) and failover (failover) control process. After years of revisions and supplements, the scope of application of SCTP has been gradually expanded, and now it has functions such as multimedia real-time transmission, heterogeneous network seamless switching, and multi-server concurrent services. In the existing SCTP specification, the main function of coupling is to maintain transmission control between two communication endpoints. So far, there is no public and clear report on applying SCTP to session migration and application migration between multiple terminals.

The identification parameters of the SCTP association include the network layer address and the transport layer port number of the two communication endpoints. IETF (Internet Engineering Task Force) RFC4960 (Document No. 4960) stipulates that both ends of an SCTP coupling can have a multi-homed configuration, that is, there can be multiple network layer addresses, but there can only be one transmission at any end. layer port value. Within an endpoint, different SCTP associations must be assigned different port values; a pair of endpoints can only establish one association at the same time. SCTP's Dynamic Address Reconfiguration (DAR, Dynamic Address Reconfiguration) function (see IETF RFC5061), allows the endpoint to dynamically configure the network layer address without interrupting the application and session, and control the message block through ASCONF (Address CONFiguration) Notify the peer. DAR provides an autonomous mobility management capability within the endpoint at the transport layer, which can be combined with mobility technologies at the physical layer, network layer, and application layer to provide support for the migration of a terminal between multiple interfaces.

With the development and popularization of smart phone terminals, Internet TV, tablet PCs and mobile Internet, the number of communication terminals that users can use at the same time or alternately generally exceeds one, and the functions and performance of terminals present diversified and coordinated development requirements. Multi-terminal aggregation and reconfiguration has become one of the key issues in terminal communication technology. Based on SCTP, the present invention is oriented to real-time multimedia sessions and applications, and proposes a method associated with communication coupling and control coupling and a coupling migration control flow for supporting session and application mobility among multiple terminals.

The invention provides a mobile control method of the IP network at the transmission layer, which can construct a complete mobile Internet application environment together with the mobile technology of the application layer, network layer and physical channel layer.

The mobility support method of the application layer mainly includes SIP technology. A SIP network consists of user agents and web servers. The network server mainly provides services such as routing, authentication, authentication, and registration for the user agent. The SIP terminal has a unique SIP URI (Universal Resource Identifier). When the terminal is assigned a new IP address, the binding between the new IP address and the SIP URL is established through the registration process, so no matter how the terminal moves, its corresponding location information can be updated, and the established sessions can be obtained maintain. SIP's support for mobility does not depend on any low-layer protocol technology, and has great applicability. However, its switching from SIP will cause a large switching delay, and its applicability to real-time communication is low.

The mobility support methods at the network layer mainly include MIP and HIP. MIP designs a home network for the terminal, and assigns a permanent home address to the mobile terminal. In addition, MIP introduces home agent and foreign agent. When a MIP mobile terminal leaves its home network and enters a foreign network, it obtains a temporary care-of address from the foreign agent, and sends a registration request message to the home agent to notify its new address. All the packets sent from the corresponding end to the mobile terminal are intercepted by the home agent when they reach the home network and forwarded to the foreign agent managing the mobile terminal through the tunnel, and then the foreign agent forwards the message to the mobile terminal. MIP has the problem of triangular routing, and at the same time, the work efficiency is not high, and the switching delay is large, which is not conducive to the mobility management of real-time applications.

HIP adds a new protocol layer between the network layer and the transport layer to separate the identification function of the IP address and provides support for mobility. HIP introduces a new host identifier name field, and the connection of the transport layer is bound to the host identifier instead of the IP address. HIP can provide a secure and seamless communication mechanism, but it needs to restructure the existing Internet architecture, making it difficult to deploy on a large scale.

Physical layer mobility is mainly used in handover and is a key technology in the mobile communication system. When a user moves from the area covered by a cell base station to another cell, a new link must be established between the user and the new cell, and the link between the user and the original cell must be deleted and released, so that The user's call can continue. At present, the handover control of the Global System for Mobile Communications (GSM, Global System for Mobile Communications) system and the Code Division Multiple Access (CDMA, Code Division Multiple Access) system is based on the mobile station-assisted handover method, which requires the mobile station to measure the surrounding ports. The signal strength is reported to the old port, and then the network determines whether to switch and which port to switch to. The main difference between GSM and CDMA is that the way to switch to a new port is unreasonable. The mobile station in the GSM system adopts a hard handover method. The mobile station must first disconnect the connection with the old port and then establish a connection with the new port; With soft handover, the user can establish a connection to a new port while remaining connected to the old port, and then disconnect from the old port.

The SCTP specification formulated by IETF includes: protocol data unit structure, function and coding; establishment, management and teardown control of a single association; flow control and error control of message transmission; dynamic address reconstruction, first address designation; fault management and path manage. Currently, multiple operating systems have been equipped with protocol stacks following the SCTP specification.

R.Stewart (Stewart), P.Lei (Ray) and M.Tuexen (Tucson) proposed the SCTP stream reconfiguration (Stream Reconfiguration) draft (drafi-ietf-tsvwg-sctp-strrst), with the enterprise as the upper layer protocol ( ULP, Up Layer Protocol) and applications provide stream reuse capabilities. The flow reconfiguration draft designs the SCTP flow reset protocol, which allows either end of the coupling to send instructions to the other end, requiring the other end to reset the flow control state parameters to specified values. The flow control status parameters that can be reset by the protocol include the transmission sequence number (TSN, Transmission Sequence Number) and stream sequence number (SSN, StreamSequence Number) of the receiving and sending ends in both input and output directions.

There are mainly 6 published patents in the same field as the present invention, which are described as follows.

(1) Public patent "A method for recovering protocol message transmission in flow control transmission protocol coupling (CN1512714)".

This patent discloses a method for recovering the transmission of protocol messages in the flow control transmission protocol coupling. The method includes the following steps: A. The upper layer protocol of the faulty end device switches its standby part to the active part, and disconnects itself B. According to the association information on the current active part, the device at the faulty end sends an association request to the SCTP endpoint at the opposite end, and re-establishes the connection between the two SCTP endpoints corresponding to the two devices with the opposite end device. Coupling; C. The SCTP endpoint of the opposite end device sends a restart request to the upper layer protocol of the opposite end, and restores the upper layer service environment of the failed end device and the opposite end device through restarting. This method can quickly restore the connection and rebuild the coupling and upper-layer transmission business environment in the event of a device failure at one end, avoiding that after one end fails, the other end does not know and still thinks that the coupling between the two ends is normal, resulting in a large number of Loss of protocol messages.

(2) Public patent "multi-address selection method based on flow control transmission protocol (CN1534951)"

This patent realizes that when retransmitting data for an SCTP endpoint with multiple addresses, all peers and local addresses are selected in sequence; in this way, if the preferred path used fails, other available paths can be found, and in the selected In the process of addressing, it can be ensured that each address existing in the SCTP endpoint can be selected; at the same time, the present invention can still use the primary address for data transfer when the primary local address or peer address has been invalidated and then restored. Transmission to meet the transmission needs of users.

(3) Public patent "Method for Protecting Coupling Based on Flow Control Transmission Protocol (CN1533100)"

This patent relates to a method for protecting SCTP coupling between processors in an IP network. It includes the following steps: the primary and backup machines at the local end establish an SCTP coupling with the peer endpoint, the primary machine enters the coupling establishment state, and the standby machine enters the coupling establishment backup state; the coupling established on the main machine executes the complete SCTP protocol process, and the backup The machine enters the establishment of backup state; when the coupling on the main machine is terminated or closed, the main machine and the standby machine switch and switch roles to work normally.

(4) Patent No. "A method for realizing the multi-homing function of the flow control transmission protocol (CN101094240)"

This patent provides a method for realizing the SCTP multi-homing function, including the following processing: maintaining preset parameters in the data structure of each communication path coupled by the flow control transmission protocol, that is, SCTP; determining the number of communication paths coupled by SCTP, And determine the address correspondence between the source address and the destination address of each communication path, and determine the main path; when it is necessary to switch the main path, poll the other communication paths of the SCTP coupling except the main path, select The appropriate communication path is used as the new primary path.

(5) Public patent "Transfer transfer method between network hosts (US 20060164974A1)"

This patent provides an error tolerance method based on SCTP, and the application scenarios include the main host, the backup host and the peer host. The primary host establishes an SCTP connection with the peer host; the standby host backs up the state of the primary host and its SCTP coupling information through a redundancy mechanism, and detects whether the primary host is available through a heartbeat mechanism. When the primary host is unavailable, the standby host sends SCTP packets to the peer host to update its association information, and establishes its SCTP association with the peer host in a non-standard way through the response information of the peer host. During the transfer process, the message sent by the peer host will be lost, which makes this method unsuitable for real-time services, and it should be mainly used for the transfer of BGP routes.

(6) Public patent "Method and system for providing high-availability SCTP applications (US20100218034A1)"

A high-availability server system is proposed, which includes a primary server and a backup server. The active server and the backup server communicate through the established control channel. When the active server and the client establish SCTP coupling, it sends the IP addresses of the active server and the backup server to the client, and at the same time maps the status of the SCTP stack and upper-layer applications to the backup server through the control channel. When the primary server is detected to be unavailable, the backup server continues to provide services to clients. This method is mainly aimed at the mobility of the server-side network.

Contents of the invention

Technical problem: the object of the present invention is to utilize the control function of SCTP to dynamically manage physically separated multiple communication terminals, to cooperate to complete the uninterrupted migration of SCTP coupling between these terminals, and to support real-time multimedia at the session level and above Applications move seamlessly and smoothly between terminals, overcoming the problem of real-time multimedia transmission interruption caused by network layer mobility and application layer mobility.

Technical solution: The present invention discloses a mobile control method for flow control transmission protocol coupling between multiple terminals. The first terminal establishes a control coupling with the second terminal of the migration target, and there is a fixed gap between the control coupling and the original communication coupling. The port matching relationship maintains the continuity of the interactive transmission between the second terminal and the peer terminal through DAR reconfiguration. The specific implementation method is:

1) The first terminal uses a port value different from that of the peer terminal when establishing the communication coupling;

2) When the first terminal establishes the control coupling, the ports on both sides match the communication coupling, that is, the service port value of the second terminal uses the port value of the communication coupling on the first terminal side, and the first terminal uses the communication coupling port value on the peer side;

3) The first terminal initiates modification of the coupled address table of the opposite end and the second terminal, and ensures a matching relationship, that is, the target address maintained by the opposite end includes the reachable address of the second terminal, and the target address maintained by the second terminal includes the address of the opposite end. reachable address;

4) The first terminal controls to complete the state synchronization between the second terminal and the first terminal;

5) After the above control is completed, the control coupling is transformed into communication coupling; the second terminal replaces the original first terminal, and initiates coupling migration control to other terminals including the original first terminal again;

6) During the SCTP coupling migration process, the SCTP coupling maintained by the peer end is not interrupted, and the message and data transmission between the peer ULP and the application is not interrupted.

The state synchronization control described in step 4) not only includes the state of ULP, but also includes the state related to the SCTP flow, wherein the state parameters related to the SCTP flow at least include the transmission sequence number (TSN), flow label (SID), Stream Sequence Number (SSN).

The synchronization control of the SCTP flow is completed by the ULP through the service primitive or in the transport layer by using the SCTP flow reconfiguration protocol.

Beneficial effects: the SCTP coupling migration control method proposed by the present invention is compatible with the existing standard-compliant SCTP implementation system, and enables online real-time multimedia applications to move from one physical terminal to another in a seamless manner, overcoming Transmission discontinuity caused by network layer mobility and application layer mobility. In addition, the SCTP coupled migration control process proposed by the present invention neither limits the number of terminals nor imposes special requirements on the functions of the terminals, so it is applicable to mutual migration within a multi-terminal group and cross-local migration.

Description of drawings

Figure 1 is a typical application scenario of SCTP-coupled migration.

Fig. 2 is a control message flow of SCTP coupled migration.

Figure 3 is a state transition diagram for SCTP-coupled migration.

Figure 4 is a schematic diagram of the topology of the 3-node network simulation experiment. Among them, the first terminal A is composed of a routing-controlled core node (core1), an interface node (intf1), and two SCTP entities (sctp1 and sctp11); the second terminal B is composed of a routing-controlled core node ( core2), an interface node (intf2), and an SCTP entity (sctp2); the peer Z is composed of a routing control core node (core0), an interface node (cn), and an SCTP entity.

Figure 5 shows the relationship between the sequence number of the packet received by the main control endpoint and the simulation time.

Figure 6 shows the relationship between the sequence number of the packet received by the new endpoint and the simulation time after the coupling migration.

Fig. 7 is the relationship between the sequence number of the packet sent by the peer and the simulation time.

Detailed ways

The present invention is based on the transport layer SCTP protocol technology, takes the identification attribute of the SCTP coupling as the control object, establishes the SCTP control coupling between the terminals that need to be migrated, assists the smooth migration of the communication coupling between the terminals, and ensures real-time multimedia applications Real-time and continuity during migration.

The SCTP protocol runs on top of IP and provides connection-oriented end-to-end communication control for two endpoints. For the convenience of description, the technical solution is described below using a formal method.

Without loss of generality, let one endpoint be named A, and the other endpoint be named Z. The SCTP coupling established between A and Z is denoted as S(A, Z), and the coupling between Z and A is denoted as S(Z, A). Another terminal named B gathers with A, joins S(A, Z) through dynamic address configuration, and reconfigures the preferred address.

S(A,Z)→S(B,Z),

S(Z,A)→S(Z,B), (1)

Among them, the symbol → represents transformation. For simplicity, A is called the first terminal, B is the second terminal, and Z is the opposite terminal.

From the perspective of association management and maintenance, identify the attribute parameters of the SCTP association, mainly including network layer address, transport layer port, and transport layer protocol type. Because the transport layer protocol type involved in the present invention is SCTP, therefore, have:

S(A, Z) = (R _{{A, Z}} , P _{{A, Z}} , R _{{Z, A}} , P _{{Z, A})} ,

S(Z, A) = (R _{{Z, A}} , P _{{Z, A}} , R _{{A, Z}} , P _{{A, Z})} , (2)

Among them, P _{{A, Z}} and P _{{Z, A}} are scalars, which are the port values assigned to the coupling by the first terminal A and the opposite terminal Z; R _{{A, Z}} and R _{{Z, A}} are The vector is the address table that A and Z respectively join according to the ULP requirements, and may contain only one address or multiple addresses. The method provided by the present invention requires the first terminal A and the opposite terminal Z to select the same port value differently, namely:

P _{{A, Z}} ≠ P _{{Z, A}} . (3)

Coupling S(A, Z) and S(Z, A) is called a communication coupling. The aggregation and reconfiguration of the second terminal B and the first terminal A are initiated by the ULP, and the coupling S(A, B) and S(B, A) established between A and B:

S(A, B) = (R _{{A, B}} , P _{{A, B}} , R _{{B, A}} , P _{{B, A}} ),

S(B, A) = (R _{{B, A}} , P _{{B, A}} , R _{{A, B}} , P _{{A, B}} ), (4)

Wherein, P _{{A, B}} and P _{{B, A}} are both vectors, which are respectively the address tables added by the first terminal A and the second terminal B according to the ULP requirement. The present invention requires:

P _{{A, B}} = P _{{Z, A}} , and, P _{{B, A}} = P _{{A, Z}} . (5)

Coupling S(A, B) and S(B, A) is called a control coupling. Through DAR control, A sends an address adding message to B, requesting B to add R _{{Z, A}} to the peer address table, namely:

R _{{A, B}} → R _{{AZ, B}} = R _{{A, B}} ∪ R _{{Z, A}} . (6)

Among them, the symbol ∪ means merge. At the same time, through DAR control, A sends an address adding message to Z, requesting Z to add R _{{B, A}} to Z's peer address table, namely:

R _{{A, Z}} → R _{{AB, Z}} = R _{{A, Z}} ∪ R _{{B, A}} . (7)

In this way, the reconfiguration of communication coupling and control coupling is completed, which can be expressed as:

S(Z,A)→S(Z,A∪B),

S(B,A)→S(B,A∪Z).                (8)

Further, through DAR control, A sends address removal messages to B and Z respectively, namely:

R _{{AZ, B}} → R _{{Z, B}} = R _{{A, B}} ∪ R _{{Z, A}} - R _{{A, B}} ,

R _{{AB, Z}} → R _{{B, Z}} = R _{{A, Z}} ∪ R _{{B, A}} - R _{{A, Z}} , (9)

F:

S(A∪B,Z)→S(B,Z),

S(A∪Z,B)→S(Z,B).                (10)

In other words, under the control of the first endpoint A, the communication couplings S(A, Z) and S(Z, A) are continuously transitioned to S(B, Z) and B(Z, B), originally in The message transmission between A and Z can be transferred to B and Z interactively in a continuous manner.

The specific control message flow of coupling migration is given below.

(1) The first terminal establishes a communication coupling with the opposite end

According to the requirement of the ULP, the first terminal initiates to establish an SCTP connection with the opposite terminal. Before the association is established, the local port value selected by the first terminal is required to be different from the port value of the peer terminal in service state. The establishment process of the communication coupling follows the SCTP specification and is completed through a 4-way handshake.

(2) The first terminal conducts a real-time conversation with the opposite end

After any end of the communication coupling detects the path delay through HEARTBEAT (heartbeat), real-time multimedia message transmission is carried out, and flow control and error control are carried out according to the specification requirements.

(3) The first terminal establishes a control coupling with the second terminal

According to the requirement of ULP, it is initiated by the first terminal to establish a control association with the second terminal. Before the coupling is established, the second terminal is in the service state, and the value of its service port is the same as the port on the first terminal side of the communication coupling; the first terminal accesses the second terminal as a client, and its local port must be selected The service port value of the communication coupling on the peer side. Control the establishment process of the coupling, follow the SCTP specification, and complete it through a 4-way handshake.

(4) The first terminal suspends the transmission control process with the opposite terminal

The first terminal suspends all real-time sessions with the opposite terminal. If state synchronization fails, the first terminal dismantles the control coupling with the second terminal, restores the real-time session with the opposite end, and returns to the communication state listed in (2).

(5) State synchronization between the first terminal and the second terminal

According to the requirements of the SCTP specification, after any end of the control coupling is delayed by the HEARTBEAT (heartbeat) detection path, it is controlled by the first terminal, and the second terminal is required to synchronize the ULP state and the SCTP flow control state.

(6) Second terminal address reconfiguration

The first terminal transmits the ASCONF control block to the second terminal through the control coupling according to the DAR specification of the SCTP. The ASCONF control block contains at least two control requests, one of which requires adding the reachable network address of the opposite end, and the other requires that this address be set as the first address. If the DAR control fails, the first terminal dismantles the control coupling with the second terminal, restores the real-time session with the opposite end, and returns to the communication stage listed in (2).

(7) Peer address reconfiguration

The first terminal transmits the ASCONF control block to the peer terminal through communication coupling in compliance with the DAR specification of SCTP. The ASCONF control block contains at least two control requests, one of which requires adding the address of the reachable network of the second terminal, and the other of which needs to set this address as the first address. If the DAR control fails, the first terminal dismantles the control coupling with the second terminal, restores the real-time session with the opposite end, and returns to the communication stage listed in (2).

(8) The second terminal conducts a real-time conversation with the opposite end

Any one of the second terminal and the opposite end continues to transmit the real-time multimedia message after the HEARTBEAT (heartbeat) detection path is delayed, and performs flow control and error control according to the specification requirements.

It should be pointed out that, for the synchronization control in step (5) of the above control flow, the synchronization content includes not only the status of the ULP, but also the status related to the SCTP flow. The state parameters related to the SCTP stream include at least two directions of transmission sequence number (TSN), stream label (SID), and stream sequence number (SSN). The synchronous control of SCTP flow, its specific implementation, can be completed by ULP through the service primitive, and can also be completed in the transport layer by using the SCTP flow reconfiguration protocol.

In the SCTP coupled migration control process described above, except for state synchronization, the operations of other steps can use existing SCTP protocol software modules that support DAR, such as the open source lksctp, etc.

In order to intuitively display the control process and effect of multi-terminal SCTP coupling migration, the communication system composed of three SCTP nodes is taken as an example, and the implementation method is verified based on the open source NS2 network simulator. The specific process is as follows.

(1) Brief description of the emulator

The open source software package NS2 is a general-purpose network simulator designed with object-oriented methods, and is widely used in the design and verification of communication network protocols and algorithms. Since version 2.19, NS2 has integrated the SCTP protocol emulation function, which has been used for DAR's concurrent multipath transmission (CMT, Concurrent Multiple Transmission) and other SCTP new functions and performance research.

(2) Emulator extension

In the existing NS2 (before version 2.33) software, an SCTP simulation node is composed of one master node and more than one interface node. The interface address added to the SCTP coupling can only be completed through pre-configuration, and the establishment of the coupling depends on The data sending request of the upper layer application. In addition, the existing SCTP emulation module does not have a flow control state parameter reconfiguration interface. Before the simulation experiment, the SCTP simulation function of NS2 needs to be modified and supplemented, mainly including the following 4 points.

Added pause control function for data block flow control;

Modify the original acceptance processing function, and do not respond to the peer in the pause state;

Add the initialization function of the interface address to support the DAR function;

Added functions for reading and reconfiguring flow control state parameters.

(3) Configure the network topology of 3 SCTP nodes

Each SCTP node is composed of two NS2 general nodes, each including a master node and an interface node, and the three interface nodes are interconnected two by two. Its topology is shown in Figure 4. Wherein, the first terminal A is configured with two SCTP protocol emulation modules, and establishes SCTP coupling with the opposite terminal Z and the second terminal B respectively.

In order to observe the change of available bandwidth before and after coupling migration, the bandwidth of the link between the first terminal A and the peer Z is set to 10Mbps, and the delay is set to 2ms; the bandwidth of the link between the second terminal B and the peer Z is set to is 2Mbps, and the delay is set to 20ms; the bandwidth and delay of the link between the first terminal A and the second terminal B do not have a critical impact on coupling migration, and the values are set to 2Mbps and 20ms respectively. The three link settings are bidirectional and symmetrical, and the output buffer of the corresponding physical port is set to 50 packets.

(4) DAR configuration

The first terminal A establishes an SCTP-based FTP connection with the peer Z in advance, constructs a communication coupling, and Z sends data packets continuously. Before the DAR starts, the first terminal A establishes an SCTP-based FTP connection with the second terminal B, and A sends a packet to drive the SCTP emulation module to establish a control coupling.

When DAR starts, the first terminal A suspends the flow control of the communication coupling, and through the emulation configuration command, adds the address of the peer Z on the side of the second terminal B that controls the coupling, and adds B on the Z side of the communication coupling the address of. At the same time, the value of the flow control state parameter of A is obtained, which is used to reconfigure the flow control state parameter of B, and then the SCTP header addresses of B and Z are reset to point to Z and B respectively.

(5) Record all simulation events

In the simulation experiment configuration program, turn on the event tracking function and record it in an external file for analysis of the results after the simulation experiment.

(6) Subsequent processing of SCTP simulation events

From the event tracking record file, all SCTP packets sent from the peer Z, received by the first terminal A, and received by the second terminal B are obtained by filtering, and the SSN and the corresponding simulation time are extracted.

(7) Other simulation parameter settings

The path maximum transmission unit (MTU, maximum transmission unit) of SCTP coupling is set to 1500 bytes, and the flow control and error control parameters use the default values of NS2. The communication coupling is established at 0.5s, the control coupling is established at 3.0s, the DAR operation is carried out at 5.0s, and the simulation experiment ends at 10.0s.

(8) Simulation experiment results

Figures 5 to 7 show the relationship between the SSN of the SCTP packet and the simulation time.

It can be seen from FIG. 5 that the first terminal A continuously receives the packet flow from the peer Z during the time between the establishment of the communication coupling and the DAR operation (at 5.0s); after the DAR operation, the Multiple packets from Z are still received within a short period of time.

It can be seen from FIG. 6 that at 5.0s, the second terminal B received the packets from the peer Z within a short time after the DAR operation was completed, but the SSN value of these packets exceeded the SSN waiting to be received. After about 100ms, B continues to receive a stream of packets from Z with the correct SSN.

It can be seen from Figure 7 that the SSN of the sent packet and the time of the peer Z from the establishment of the communication coupling (0.5s) to the DAR operation (5.0s) have a linear relationship, and the slope matches the link bandwidth of 10Mbps . In a short period of time (less than 100ms) after the DAR operation, the relationship between the SSN and time of the packet sent by Z is unstable. After that, the SSN of the packet sent by Z backed off, and after about 500ms, the SSN of the sent packet and time showed a stable linear relationship again, and the slope became smaller, but it matched the link bandwidth of 2Mbps. Before 5.0s and after 5.5s, the change curve of SSN and time, the change of slope, is completely consistent with the topology configuration.

The above SCTP coupling migration verified by simulation can ensure the continuity of ULP data information. The instability duration caused by the out-of-order packet SSN is less than 100ms, and the duration of throughput performance degradation caused by migration is less than 500ms. The available bandwidth can be used steadily.

The SCTP coupling migration method proposed by the invention can meet the performance requirements of real-time multimedia applications.

Claims (3)

1. A mobile control method for flow control transmission protocol coupling between multiple terminals, characterized in that the first terminal establishes a control coupling with the second terminal of the migration target, and there is a fixed port between the control coupling and the original communication coupling Matching relationship, maintain the continuity of the interactive transmission between the second terminal and the peer terminal through DAR reconfiguration, the specific implementation method is: 1) The first terminal uses a port value different from that of the peer terminal when establishing a communication coupling; 2) When the first terminal establishes the control coupling, the ports on both sides match the communication coupling, that is, the service port value of the second terminal uses the port value of the communication coupling on the first terminal side, and the first terminal uses the communication coupling port value on the peer side; 3) The first terminal initiates modification of the coupled address table of the peer and the second terminal, and ensures a matching relationship, that is, the target address maintained by the peer includes the reachable address of the second terminal, and the target address maintained by the second terminal includes the reachable address of the peer. reachable address; 4) The first terminal controls to complete the state synchronization between the second terminal and the first terminal; 5) After the above control is completed, the control coupling is transformed into communication coupling; the second terminal replaces the original first terminal, and initiates coupling migration control to other terminals including the original first terminal again; 6) During the SCTP coupling migration process, the SCTP coupling maintained by the peer is not interrupted, and the message and data transmission between the peer ULP and the application is not interrupted. 2. The mobile control method for flow control transmission protocol coupling between multiple terminals according to claim 1, characterized in that the state synchronization control in step 4) includes not only the state of ULP, but also the state related to SCTP flow. , where the state parameters related to the SCTP stream include at least the transmission sequence number (TSN), stream label (SID), and stream sequence number (SSN) in two directions. 3. the mobile control method of flow control transmission protocol coupling between a kind of multi-terminal according to claim 2, it is characterized in that the synchronous control of described SCTP stream is completed by ULP or adopts SCTP stream to reconfigure The protocol is done within the transport layer.

Images