A GPON+IMS fusion-based railway telephone private network migration system

Abstract

The application belongs to the technical field of communication, and specifically discloses a railway telephone private network migration system based on GPON+IMS fusion. The system extracts call data from a V5.2 link, realizes bidirectional conversion between V5.2 signaling and SIP signaling by using a dual-mode protocol conversion, so that existing OLT, ONU and other access network devices supporting only V5.2 protocol can be directly connected to an IMS core network based on SIP protocol without hardware modification, thereby solving the problem of incompatible device protocols. At the same time, the system determines the candidate SBC type according to the called number, selects a target SBC node according to the physical topology identification of the calling ONU, and realizes accurate service shunting. Finally, the system performs a three-stage gradual switching operation in the parallel period, the bypass monitoring period and the switching period, thereby avoiding the signaling storm caused by large-scale ONU registration to IMS at the same time, and solving the problem of service interruption in traditional switching.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and specifically discloses a railway telephone private network migration system based on GPON+IMS convergence. Background Technology

[0002] The railway telephone network is a dedicated voice communication network used within the railway system for train dispatching and administrative operations. Early networks commonly used program-controlled switching technology, connecting access network equipment at stations and sections along the line to the program-controlled switches via V5.2 interfaces. With the development of communication technology, all-IP networks centered on IMS have become the mainstream evolution direction. To achieve resource integration, the entire railway telephone network needs to be migrated to the IMS network.

[0003] However, the migration of the existing railway telephone network faces the following technical problems: First, equipment protocols are incompatible: the existing access network equipment only supports the V5.2 protocol and cannot directly connect to the IMS core network based on the SIP protocol. Replacing all the equipment would be extremely costly and would result in a large amount of usable old equipment being idle and wasted.

[0004] Second, the service routing cannot be intelligently distributed: Railway telephone network services are complex. Regular landline users should connect to the nearest provincial mobile fixed-line network SBC, while transfer-type voice lines need to connect to the provincial mobile I-SBC where the railway bureau is located. Existing equipment typically uses static configuration or random selection when choosing SBC nodes, making it difficult to select nodes specifically based on the called number's service type. This easily leads to transfer-type services being incorrectly sent to the provincial network SBC, resulting in call failure, or regular landlines being sent to a distant I-SBC, increasing transmission latency and severely impacting call quality and service reliability.

[0005] Third, the cutover process carries the risk of service interruption: to achieve service takeover from the program-controlled exchange to the IMS core network, existing equipment needs to be cut over. Traditional cutover solutions typically adopt a disconnect-then-reconnect model, which requires interrupting existing services. However, railway communications involve critical services such as train dispatching and emergency rescue, and have high requirements for communication continuity. Any interruption of communication may affect train operation safety. Summary of the Invention

[0006] To address or at least partially address the aforementioned technical problems, this invention provides a railway telephone private network migration system based on GPON+IMS convergence. Through a dual-mode access architecture and intelligent routing mechanism, it enables the reuse of old equipment, precise service diversion, and zero-interruption cutover, thereby reducing transformation costs, improving migration efficiency, and ensuring the continuity and reliability of critical railway communication services.

[0007] The objective of this invention can be achieved through the following technical solution: a railway telephone private network migration system based on GPON+IMS convergence, comprising: a dual-mode protocol conversion module, which extracts call data from the V5.2 link, converts V5.2 signaling into SIP signaling according to the SIP routing header returned by the intelligent routing decision module, and encapsulates voice into an RTP stream.

[0008] The intelligent routing decision module determines the candidate SBC type based on the called number, selects the target SBC node based on the physical topology identifier of the calling ONU, and generates routing header information containing the SBC address.

[0009] The zero-interruption cutover control module executes a three-stage cutover process: during the parallel phase, it registers with both the PBX and IMS simultaneously and locks the main path of new calls to the PBX; during the bypass monitoring phase, it modifies the routing priority so that new calls take priority through IMS and back to the PBX in case of failure; during the switching phase, it issues mode switching instructions to ONUs in batches according to the active call status and the principle of distribution, and monitors the registration request rate to execute backoff or alarm.

[0010] The QoS policy control module identifies RTP and SIP streams sent from the dual-mode protocol conversion module, marks their priority, and allocates transmission resources higher than those for ordinary data streams.

[0011] Combining all the above technical solutions, the positive effects of this invention are as follows: 1. This invention extracts call data from the V5.2 link and uses a dual-mode protocol conversion module to achieve bidirectional conversion between V5.2 signaling and SIP signaling, as well as mutual encapsulation / decapsulation of voice media streams and RTP packets. This allows existing access network devices such as OLTs and ONUs that only support the V5.2 protocol to directly connect to the IMS core network based on the SIP protocol without hardware modification, solving the problem of device protocol incompatibility, avoiding the high costs brought about by large-scale equipment replacement, and realizing the full utilization of existing equipment.

[0012] 2. This invention determines the candidate SBC type based on the called number and selects the target SBC node based on the physical topology identifier of the calling ONU. This enables transfer-type services to be automatically routed to the railway bureau's I-SBC, while ordinary landlines are connected to the nearest provincial network SBC. This greatly reduces the incidence of transfer-type services being mistakenly sent to the provincial network SBC and resulting in call failure, and ordinary landlines being sent to the I-SBC over long distances and increasing transmission latency. As a result, it achieves accurate service diversion, reduces end-to-end latency, and improves the call quality and reliability of the railway private network.

[0013] 3. This invention performs a three-stage progressive cutover operation during the parallel phase, the bypass monitoring phase, and the switching phase. During the parallel phase and the bypass monitoring phase, a dual registration and main path locking mechanism is used to ensure uninterrupted service. During the switching phase, mode switching instructions are issued in batches according to priority and distribution principles, and the registration request rate is monitored in real time to perform backoff or alarm. On the one hand, this avoids the signaling storm caused by a large number of ONUs registering to IMS at the same time. On the other hand, it retains the hot standby of the PBX and supports automatic fallback in case of failure, solving the problem of service interruption required by traditional cutover and achieving the effect of zero call interruption during the cutover process. Attached Figure Description

[0014] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0015] Figure 1 This is a system module connection diagram of the present invention.

[0016] Figure 2 This is a schematic diagram of the architecture of the dual-mode protocol conversion module in this invention.

[0017] Figure 3 This is a flowchart illustrating the implementation process of dividing ONUs into batches according to their active call status and the principle of distribution in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In railway telephone networks, due to the wide distribution of stations and nodes along railway lines, it is necessary to unify the access of scattered terminal equipment to the network and enable calls between any two points. For this purpose, program-controlled exchanges are used to provide voice communication through circuit switching technology. However, program-controlled exchanges are closed equipment with weak service expansion capabilities and high maintenance costs, making them unsuitable for the current needs of network convergence and rapid service deployment.

[0020] With the development of communication technology, IMS, as an all-IP core network architecture based on the SIP protocol, supports the converged access of multimedia services such as voice, data, and video, and has become the mainstream evolution direction of communication networks. To reduce network operation and maintenance costs and enable open service capabilities, it is necessary to migrate the entire railway telephone network to the IMS network. However, existing access network equipment only supports the V5.2 protocol and cannot directly interface with IMS, and the migration process must ensure zero service interruption.

[0021] To this end, this invention proposes a railway telephone private network migration system based on GPON+IMS convergence, including a dual-mode protocol conversion module, an intelligent routing decision module, a zero-interruption cutover control module, and a QoS policy control module. The connection relationships between these modules are described in [reference needed]. Figure 1 As shown.

[0022] The dual-mode protocol conversion module is deployed between the existing OLT and the new GPON OLT, connecting to the V5.2 link, the output of the intelligent routing decision module, and the input of the QoS policy control module. It is used to extract call data from the V5.2 link and send it to the intelligent routing decision module. Based on the SIP routing header returned by the intelligent routing decision module, it converts the V5.2 signaling into SIP signaling, encapsulates the media stream into RTP packets, and outputs them to the QoS policy control module, realizing bidirectional lossless conversion between traditional 2M / E1 signals and IP voice streams.

[0023] Specifically, the dual-mode protocol conversion module adopts a three-layer architecture, see [link to documentation]. Figure 2 As shown, the system includes a physical layer, a protocol adaptation layer, and a session mapping layer. The protocol adaptation layer includes a SIP protocol generation unit and a media gateway control unit, which can separate physical signal processing, protocol parsing and conversion, and session state maintenance. Specifically, it performs the following conversion process: First, it receives the 2M / E1 signal carrying call data from the V5.2 link. 2M / E1 is a digital trunk signal, and each frame contains 32 time slots, where a time slot is a fixed time interval, and each time slot can transmit one channel of digital voice or signaling information. The physical layer extracts two types of data from each time slot in each frame: signaling data (i.e., V5.2 protocol messages, used for call establishment, release, and other control) and voice data (analog voice signals encoded by PCM). Among them, the signaling data is sent to the protocol adaptation layer, and the voice data is temporarily stored in the time slot buffer.

[0024] Then, the V5.2 signaling parsing unit in the protocol adaptation layer parses the incoming signaling data, i.e., the V5.2 message frames, frame by frame, extracting the calling number, called number, time slot allocation, and service type. This information is used to construct subsequent SIP requests: the calling and called numbers determine the From / To header of the SIP message, the time slot allocation is used to associate the voice channel, and the service type affects the routing policy. The parsed information is then passed to the SIP protocol generation unit.

[0025] Secondly, the SIP protocol generation unit constructs a SIP INVITE request or response message based on the above information and the routing header returned by the intelligent routing decision module. The routing header is a field in the SIP message that specifies the next routing node the message must traverse, i.e., the selected SBC address. SIP INVITE is a request message used in the SIP protocol to initiate a session, while the response message indicates acceptance or rejection of the session. The constructed SIP message is sent to the IP network through the GE / 10GE optical interface of the physical layer; simultaneously, the protocol adaptation layer encapsulates the corresponding voice data in the time slot buffer into RTP packets frame by frame and sends them to the IP network through the physical layer. This converts circuit-slot voice into an IP-domain voice stream, enabling media streaming.

[0026] Next, for SIP messages from the IP network, such as responses or requests from the peer, after parsing by the SIP protocol generation unit, the H.248 media gateway control unit initiates a V5.2 establishment request to the PBX side, notifying the PBX to allocate the corresponding circuit resources. Simultaneously, RTP packets received from the IP network are decapsulated by the protocol adaptation layer, restored to the original PCM voice data, mapped to the corresponding 2M / E1 timeslot, and sent to the V5.2 link through the physical layer. Thus, a bidirectional media channel is established between the IP side and the circuit side, enabling end-to-end communication.

[0027] Finally, the session mapping layer is responsible for maintaining the correspondence between call states and physical resources. When a new call is established, a SIP session identifier is assigned to the call based on the time slot allocation information, and this identifier is bound one-to-one with the V5.2 time slot number occupied by the call. Simultaneously, the corresponding RTP synchronization source identifier is associated to identify RTP packets from the same media stream, and a mapping record is generated and stored in memory. When a call is released, the corresponding mapping record is found and deleted based on the V5.2 time slot number to release resources, avoid leaving useless states in memory, and thus support the normal establishment of subsequent new calls.

[0028] Through the above layered processing, the dual-mode protocol conversion module achieves bidirectional lossless conversion between V5.2 and SIP protocols, and supports real-time status synchronization between primary and backup gateways, thus solving the problem of device protocol incompatibility.

[0029] After resolving the equipment protocol incompatibility issues, existing V5.2 access network equipment was able to achieve basic communication with the IMS core network. However, the railway telephone private network has complex service types, and there are fundamental differences in routing strategies between ordinary landline users and call-transfer services: ordinary landlines should connect to the nearest provincial mobile fixed-line network SBC to reduce transmission latency; call-transfer services must connect to the provincial mobile I-SBC where the railway bureau is located to meet call routing management and security supervision requirements. Existing equipment typically uses static configuration or random selection when choosing SBC nodes, making it difficult to perform targeted routing based on the called number's service type, severely impacting call quality and service reliability.

[0030] To address the aforementioned issues, this invention provides an intelligent routing decision module. Its input is connected to the call data output of the dual-mode protocol conversion module, and its output is connected to the routing header input of the dual-mode protocol conversion module. This module determines the candidate SBC type based on the called number, selects the target SBC node from the candidate SBC types based on the calling ONU physical topology identifier, generates routing header information containing the SBC address, and returns it to the dual-mode protocol conversion module. This allows the dual-mode protocol conversion module to directly reuse the routing header when constructing subsequent SIP messages, achieving accurate service routing and optimal path selection.

[0031] In a specific implementation of the present invention, the processing of the intelligent routing decision module is divided into the following two steps: Step 1: Determine the candidate SBC type based on the called number. This is to distinguish between ordinary landline and transfer services, and to provide a basis for subsequent routing selection. Specifically, it includes the following sub-steps: S11: Obtain the called number string from the call data sent by the dual-mode protocol conversion module. This string is the complete number dialed by the user.

[0032] S12. Export all number segment configurations from the railway communication management system, extract the prefix of each number segment as the key, and use the service type to which the number segment belongs as the value to form multiple mapping records; store all mapping records in key-value pair format to build a user feature database, which stores different number prefixes and their corresponding service types; transfer services include but are not limited to: railway customer service center numbers (such as 12306), dispatch console numbers (such as those starting with 0931-9), private network emergency calls (such as 110), and trunk number segments between specific railway bureaus; the above number segments are uniformly configured by the railway communication management system and synchronized to the user feature database.

[0033] As an example, the mapping records are shown in Table 1.

[0034] Table 1 Mapping Records

[0035]

[0036] S13. Compare the obtained called number string with each number prefix in the database one by one: Take out the number prefix of each record in turn, check whether the starting part of the called number string, that is, the substring with a length equal to the prefix length starting from the first character, is completely the same as the prefix; if they are the same, it means that the prefix matches the head of the called number, and the record is determined to be a successful match, and the character length of the number prefix is ​​used as the matching length for this match.

[0037] S14. Among all successfully matched records, select the record with the longest matching length as the optimal matching record. This is because there is an inclusion relationship between number prefixes. A longer prefix corresponds to a more specific number segment, which has a higher priority and can avoid ambiguity.

[0038] For example, the called number is "0935-900123", starting with "0935-". If a record also has "0935-" as its prefix, matching the beginning of the called number, this record will be a successful match with a length of 5. Another record has "0935-9" as its prefix, matching the beginning of the called number, and this record will also be a successful match with a length of 6. The latter is longer and will be selected.

[0039] S15. Determine the type of service based on the best matching record: If the service type is a transfer type, then the candidate SBC type is determined to be the railway bureau I-SBC; if the service type is a registration type ordinary landline or no record is matched, then the candidate SBC type is determined to be the provincial network SBC.

[0040] The second step is to select the target SBC node based on the physical topology identifier of the calling ONU. Based on the determined candidate SBC types, the optimal node is selected from multiple SBC nodes under that type. Specifically, the following sub-steps are included: S21. Obtain the physical topology identifier of the calling ONU from the call data. The physical topology identifier consists of a triplet of OLT device number, PON port number and ONU serial number.

[0041] S22. The maintenance personnel pre-configure the equipment location mapping table according to the equipment deployment location along the railway line, and record the station or section base station information where each OLT device and PON port is located. The OLT device number and PON port number are used to filter the geographical area where the ONU is located by querying the mapping table. The purpose is to know the physical location of the calling user so as to select the nearest SBC node in the future.

[0042] S23. Obtain all available SBC nodes under the candidate SBC type, and extract the number of transmission hops from the deployment area to the geographical area label of each SBC node from the node information table that stores the number of transmission hops between the deployment area of ​​each SBC node and each railway geographical area. This node information table is pre-configured by the transmission network management system. The number of transmission hops reflects the network path length. The smaller the number of hops, the lower the transmission latency.

[0043] S24. Arrange all available nodes in ascending order of transmission hop count to form a node priority list, and select the SBC node with the smallest transmission hop count as the target SBC node to ensure the lowest latency.

[0044] S25. During the call setup process, the dual-mode protocol conversion module sends a SIP request to the selected target SBC node and starts a timer.

[0045] S26. If no response is received within the time window set by the timer, it is determined that the response has timed out; if the status code in the received SIP response is a congestion indication, it is determined that there is congestion.

[0046] S27. When the target SBC node experiences a response timeout or congestion indication, select the next node from the node priority list as a replacement, and repeat the timeout / congestion determination until an available node is found.

[0047] S28. Write the final selected SBC node IP address into the Route header of the SIP message and return to the dual-mode protocol conversion module.

[0048] Through the above two-step decision-making process, the intelligent routing decision module ensures the correct matching of service type and SBC type, while realizing proximity routing based on geographical location, effectively avoiding routing errors and increased transmission latency.

[0049] After the intelligent routing decision module completes the route selection, a cutover operation needs to be performed on the existing equipment to achieve service takeover from the program-controlled exchange to the IMS core network. Traditional cutovers usually adopt a disconnect-then-reconnect mode, which requires interrupting existing services. However, railway communications involve critical services such as train dispatching and emergency rescue, and have extremely high requirements for communication continuity. Any interruption of communication may affect train operation safety.

[0050] To this end, this invention proposes a zero-interruption cutover control module. This module connects to the ONU management interface of the new GPON OLT and the IMS registration server respectively, and is used to execute a three-stage cutover process: parallel period, bypass monitoring period, and switching period. The specific implementation methods of each stage are as follows: (1) Parallel period: Under the premise of maintaining the existing services without interruption, verify the registration and basic communication capabilities of the dual-mode protocol conversion module with the IMS core network. This stage generally lasts for 7 days, and specifically performs the following operations: control the dual-mode protocol conversion module to simultaneously maintain the V5.2 registration status with the program-controlled exchange and the SIP registration status with the IMS core network, i.e., dual registration. The purpose of doing so is to enable the gateway to have access capabilities of two types of networks at the same time, so as to switch at any time.

[0051] The primary path for all new calls—that is, the network path through which the call is actually connected—is forcibly locked to the PBX (Private Branch Exchange). In other words, all new calls are still handled by the original PBX, and the IMS only receives mirrored traffic for testing purposes, without carrying real business traffic. This ensures that actual communication is not affected during the verification period.

[0052] (2) Bypass Monitoring Period: Real services are gradually redirected to the IMS core network to test the stability of IMS carrying actual calls, while the PBX is retained as a hot standby path. This phase typically lasts 3 days and involves the following operations: Modifying the routing priority configuration of the dual-mode protocol conversion module to prioritize new calls by attempting to establish a connection through the IMS core network. When IMS path establishment fails (e.g., receiving a 4xx / 5xx error response or timeout with no response), the call is automatically switched to the PBX path within a set time window to ensure uninterrupted service.

[0053] Inside the dual-mode protocol conversion module, two PCM voice streams are recorded simultaneously for the same call: one from the PBX side (via the V5.2 interface), and the other from the IMS side (obtained by decapsulating the received RTP media stream). The purpose of this is to objectively compare the quality differences between the two voice streams.

[0054] The two PCM voice streams are aligned by time frame, and the amplitude difference between the two frames is compared frame by frame. When the absolute value of the amplitude difference of a certain frame exceeds the distortion threshold, the distortion threshold can be set according to the recommendations of standards such as ITU-T P.862. For example, if the relative amplitude deviation is greater than 5%, it is marked as a distorted frame. Then, the proportion of distorted frames is counted as the recording difference rate. The lower the recording difference rate, the closer the call quality on the IMS side is to the original program-controlled exchange.

[0055] If the recording difference rate does not exceed the preset allowable difference rate within several consecutive natural days, such as 3 days, it means that the voice quality of the IMS core network has met the requirements and can be switched from the bypass monitoring period to the handover period. The allowable difference rate can be set according to the railway communication quality requirements, and it is generally recommended to be below 1%.

[0056] Switching Period: Complete the final switch of ONU operating mode, migrating all traffic from the PBX to the IMS core network, while avoiding signaling storms and service interruptions. See [link / reference] Figure 3 As shown, the specific implementation includes the following: (31) Batch principle: Since a large number of ONUs registering to IMS at the same time will cause a registration request storm, it is necessary to batch the ONUs according to whether they have active calls and the principle of distribution. The specific batch process is as follows: (311) Before the start of the switching period, collect the device attributes and current active call status of all ONUs from the management interface of the OLT; among which the device attributes include the PON port to which they belong and the geographical area where they are located, and the active call status includes whether they are currently in a call.

[0057] (312) All ONUs with no active calls are directly assigned to the first batch because there are no ongoing calls on these ONUs, the switching risk is the lowest, and they can be prioritized.

[0058] (313) Form a set of ONUs with active calls to be assigned, create a new batch, take an ONU from the set to be assigned as the first member of the batch, traverse the remaining ONUs in the set to be assigned, and add ONUs that do not belong to the same PON port and the same geographical area as all existing members in the batch to the current batch. This is to avoid ONUs from the same PON port or the same geographical area being concentrated in the same batch, thereby dispersing the sudden pressure of registration requests. When the size of the current batch reaches the limit of the number of each batch, such as each batch not exceeding 10% of the total number of ONUs, or when no new members can be added after the traversal is completed, stop adding and determine the current batch as the next batch.

[0059] (314) Remove the ONUs that have been added to the next batch from the set to be allocated, and continue to create a new batch until the set to be allocated is empty.

[0060] (32) Issuing mode switching instructions in batches: To avoid a large number of ONUs initiating registration simultaneously, instructions must be issued in batches according to the batch order. The specific operation is as follows: In batch order, network mode switching instructions are issued to the ONUs in the current batch through the OLT's remote management protocol, switching the ONU's working mode from the original V5.2 proxy mode to the IMS local registration mode. After the switch, the ONU will directly initiate a SIP registration request to the IMS registration server, instead of going through the PBX proxy.

[0061] (33) Rate monitoring and backoff / alarm: Considering that large-scale registration may cause instantaneous signaling storms, the registration request rate needs to be monitored in real time after each batch switch, and backoff or alarm should be executed. Specifically, this includes: (331) After the mode switch command for each batch is issued, the arrival rate of registration requests initiated by ONUs in that batch is collected in real time from the IMS registration server.

[0062] (332) When the arrival rate of the collected data exceeds the rate threshold, it is determined to be in a burst registration state. The rate threshold can be set according to the maximum processing capacity of the IMS registration server, such as 80% of the server's processing capacity. Before the start of the next batch, a backoff waiting time window such as 30s is automatically inserted to postpone the issuance of switching instructions for subsequent batches, so as to give the server processing time and avoid overload. At the same time, on the OLT side, the registration requests initiated by the ONU are arranged into a queue with uniform output according to the arrival time, that is, the burst traffic is smoothed into a constant rate outflow, avoiding instantaneous spikes, and duplicate or incorrectly formatted registration requests are discarded to reduce the server's invalid load.

[0063] (333) For each batch, count the total number of registration attempts initiated by all ONUs in the current batch, as well as the number of failures due to authentication failure, timeout without response, or return of error code, and calculate the proportion of the number of failures to the total number of attempts as the registration failure ratio.

[0064] (334) If the registration failure rate exceeds the allowed failure rate, for example, 5%, then stop switching of all subsequent batches, output alarm information containing the current failed batch number, and keep the program-controlled exchange in hot standby state so as to quickly roll back when necessary and ensure that critical communication is not interrupted.

[0065] Through the above three-stage gradual cutover, the railway telephone private network was migrated to the IMS core network with zero interruption and smoothness.

[0066] During the migration of the railway telephone private network to the IMS core network, as voice services are gradually carried on the IP network, the transmission quality of voice streams is highly susceptible to factors such as network congestion and packet jitter. To ensure high-priority transmission of voice services, this invention introduces a QoS policy control module. The input of this module is connected to the IP network exit of the dual-mode protocol conversion module, and the output is connected to the transmission network. It is used to identify RTP streams (carrying voice media) and SIP streams (carrying signaling) originating from the dual-mode protocol conversion module, prioritize them, and allocate transmission resources higher than those for ordinary data streams. The specific implementation is as follows: a) Message identification: At the GE interface of the transmission system, deep packet inspection is used to identify RTP packets with UDP ports within the voice media port range (e.g., 16384–32767), and packets with TCP or UDP ports as SIP signaling ports (e.g., 5060 / 5061). The purpose of this step is to accurately distinguish voice traffic from mixed data streams, laying the foundation for subsequent differentiated services.

[0067] b) Priority Marking: For identified voice packets, an Expedited Forwarding (EF) code point is written to the DSCP field of the IP header, and the highest priority value is written to the EXP field of the MPLS-TP label. These two markings ensure that the packet receives consistent high-priority processing when traversing both the IP network and the MPLS domain.

[0068] c) Queue Allocation and Bandwidth Guarantee: In the egress queue scheduler of the transmission system, a separate priority queue is allocated to voice streams marked with high priority, and a minimum guaranteed bandwidth, such as at least 64kbps per voice stream, is reserved for this queue. This ensures that the voice queue has a fixed transmission opportunity even when other data traffic surges, avoiding packet loss or latency jitter due to queue starvation.

[0069] d) Congestion handling: When congestion occurs in the transmission system, the queue scheduler prioritizes discarding messages in non-voice queues that are not marked with priority, while retaining messages in the voice queue. This can protect the integrity of the voice stream to the greatest extent, ensure the continuity and intelligibility of the call, and thus meet the requirements of railway critical business for high-quality voice communication.

[0070] Through the aforementioned QoS policy control module, this invention provides end-to-end quality of service assurance for railway telephone private network voice services in an IP network environment, effectively avoiding the impact of network congestion on call quality.

[0071] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0072] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0073] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0074] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0075] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A railway telephone private network migration system based on GPON+IMS convergence, characterized in that, include: The dual-mode protocol conversion module extracts call data from the V5.2 link, converts V5.2 signaling into SIP signaling based on the SIP routing header returned by the intelligent routing decision module, and encapsulates voice into an RTP stream. The intelligent routing decision module determines the candidate SBC type based on the called number, selects the target SBC node based on the physical topology identifier of the calling ONU, and generates routing header information containing the SBC address. The zero-interruption cutover control module executes a three-stage cutover process: during the parallel phase, it registers with both the PBX and IMS simultaneously and locks the main path of new calls to the PBX; during the bypass monitoring phase, it modifies the routing priority so that new calls take priority through IMS and falls back to the PBX in case of failure; during the switching phase, it issues mode switching instructions to ONUs in batches according to the active call status and the principle of dispersion and monitors the registration request rate to perform backoff or alarm. The QoS policy control module identifies the output RTP and SIP streams from the dual-mode protocol conversion module, marks their priority, and allocates transmission resources higher than those for ordinary data streams.

2. The railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The dual-mode protocol conversion module adopts a three-layer architecture: a physical layer, a protocol adaptation layer consisting of a SIP protocol generation unit and a media gateway control unit, and a session mapping layer, and performs the following conversion process: The physical layer receives 2M / E1 signals from the V5.2 link, extracts signaling data and voice data frame by frame, outputs the signaling data to the protocol adaptation layer, and temporarily stores the voice data in the time slot buffer. The protocol adaptation layer receives and parses the signaling data frame by frame, extracts the calling number, called number, time slot allocation and service type, and outputs the parsed information to the SIP protocol generation unit. The SIP protocol generation unit constructs a SIPIVITE request or response message based on the received information and the routing header returned by the intelligent routing decision module, and sends the message to the IP network through the physical layer; at the same time, the protocol adaptation layer encapsulates the voice data in the time slot buffer into RTP packets and outputs them to the IP network through the physical layer. After the media gateway control unit passes the SIP messages from the IP network to the protocol adaptation layer for parsing, it initiates a V5.2 establishment request to the PBX based on the parsing result. At the same time, the RTP packets from the IP network are decapsulated into PCM voice data by the protocol adaptation layer, mapped to the 2M / E1 timeslot by the physical layer, and sent to the V5.2 link. When a new call is established, the session mapping layer assigns a SIP session identifier and binds it one-to-one with a V5.2 timeslot number, and associates it with an RTP synchronization source identifier, generating a mapping record and storing it in memory; when the call is released, the corresponding mapping record is deleted according to the V5.2 timeslot number.

3. The railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The content of determining the candidate SBC type based on the called number includes: Retrieve the called number string from the call data; Export all number segment configurations from the railway communication management system, and construct a user characteristic database with the prefix of each number segment and its corresponding business type; The called number string is compared with each number prefix in the database one by one. Records whose number prefix is ​​the same as the beginning part of the called number string are selected as successful matches, and the length of the number prefix is ​​used as the matching length. Among all successfully matched records, the record with the longest match length is selected as the optimal match record; When the service type of the best matching record is a transfer type, the candidate SBC type is determined to be the railway bureau I-SBC; otherwise, it is determined to be the provincial network SBC.

4. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The specific details of selecting the target SBC node include: Obtain the physical topology identifier of the calling ONU from the call data, and extract the OLT device number, PON port number, and ONU serial number; Convert the OLT device number and PON port number into a geographical area label; Get all available SBC nodes under the candidate SBC type, and extract the number of transmission hops from the deployment area to the geographical area label of each SBC node from the node information table that stores the number of transmission hops between the deployment area of ​​each SBC node and each railway geographical area based on the geographical area label. Sort all available nodes in ascending order of transmission hop count, and select the SBC node with the smallest transmission hop count as the target SBC node. Write the target SBC node IP address into the Route header of the SIP message and return it to the dual-mode protocol conversion module.

5. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The parallel period is executed according to the following process: The dual-mode protocol conversion module maintains its registration status with both the PBX and the IMS core network. Lock the main path for all new calls to the PBX.

6. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The bypass monitoring period shall be performed according to the following procedure: Modify the routing priority configuration of the dual-mode protocol conversion module to prioritize the establishment of new calls through the IMS core network. When the IMS path establishment fails, the call will be switched to the PBX path within a set time window. The dual-mode protocol conversion module can simultaneously record two PCM voice streams for the same call: one is the direct PCM voice stream from the PBX side, and the other is the PCM voice stream obtained by decapsulating the RTP media stream from the IMS side. The two PCM audio streams are aligned by time frames, and the amplitude difference between the two frames is compared frame by frame. When the absolute value of the amplitude difference of a certain frame exceeds the distortion threshold, it is marked as a distorted frame. Then, the proportion of distorted frames is counted as the recording difference rate. When the recording difference rate does not exceed the allowable difference rate within several consecutive natural days, the process transitions from the bypass monitoring period to the switching period.

7. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: During the handover period, ONUs will be operated in batches according to their active call status and the principle of distribution, following the procedure below: Before the switchover period begins, the device attributes and current active call status of all ONUs are collected from the OLT's management interface. The device attributes include the PON port to which they belong and the geographical area where they are located. The active call status includes whether they are currently in a call. All ONUs with no current active calls will be included in the first batch. Group the ONUs that currently have active calls into a set to be assigned; Create a new batch, and take an ONU from the set to be assigned as the first member of the new batch; Iterate through the remaining ONUs in the set to be allocated, and add ONUs that do not belong to the same PON port and do not belong to the same geographical region as the existing members in the new batch to the current batch; When the current batch reaches the limit or no new members can be added after the traversal is complete, stop adding members and designate the current batch as the next batch. Remove the ONUs that have been added to the next batch from the set to be allocated, and continue to create new batches until the set to be allocated is empty.

8. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The batch-by-batch distribution of mode switching instructions is executed according to the following process: In batch order, instructions to switch from V5.2 agent mode to IMS local registration mode are sent to the ONUs in the current batch via the OLT's remote management protocol.

9. The railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The specific details of monitoring the registration request rate to execute backoff or alarm include: After the batch switching command is issued, the arrival rate of registration requests initiated by ONUs within the batch is collected; When the arrival rate exceeds the rate threshold, a backoff waiting time window is inserted before the start of the next batch to temporarily suspend the issuance of switching instructions for subsequent batches. Registration requests initiated to ONUs on the OLT side are arranged in chronological order of arrival time, and duplicate or incorrectly formatted registration requests are discarded. The registration failure rate is calculated as the proportion of the number of failed registrations to the total number of registration attempts initiated by all ONUs in the current batch. If the registration failure rate exceeds the allowed failure rate, subsequent batch switching will be stopped, alarm information will be output, and the PBX will be kept in hot standby mode.

10. A railway telephone private network migration system based on GPON+IMS convergence as described in claim 1, characterized in that: The QoS policy control module includes the following: Identify RTP and SIP signaling messages on the GE interface; The identified packets are marked with the highest priority for IP header DSCP accelerated forwarding code points and MPLS-TP labels; Send the tagged messages into a separate priority queue; When transmission congestion occurs, messages in the untagged non-voice queue are dropped first.

Images