VoLTE voice quality assessment method and system
A voice quality and voice technology, applied in the field of communication, can solve problems such as the inability to guarantee the evaluation accuracy, and achieve the effect of convenient query and reliable evaluation results
Inactive Publication Date: 2018-08-21
CHINA MOBILE GRP GUANGDONG CO LTD +1
5 Cites 16 Cited by
AI-Extracted Technical Summary
Problems solved by technology
[0005] Due to the problem that the evaluation accuracy of existing methods cannot be guaranteed, t...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreMethod used
Specifically, described voice and signaling data are analyzed, obtain the media flow of above-mentioned 5 interfaces, by screening, the media flow of Mb interface or Nb interface is screened out, and adopt P.563 algorithm, for all The MOS value scoring of the RTP package can realize efficient and fast query and processing of large amounts of data.
The present embodiment carries out association analysis for the data of multi-interface (comprising Mb, Nb, S11, Gm and Mc interface); Adopt distributed framework, utilize acquisition server, processing server, loader server, hadoop cluster, database server and application The combination of servers can process the data of the above five interfaces in real time (15-minute granularity presentation), massive (2W users in the whole network), and high-efficiency (minute-level query is completed); for the voice quality evaluation of RTP media streams, the ITU standard is adopted P.563 algorithm for evaluation, the evaluation results are reliable and accurate. Compared with the traditional drive test results, the fitting degree of the MOS results based on the P.563 algorithm reaches 0.82, and the average error is less than 0.3, while the average error between the MOS results based on the E-Model model algorithm and the traditional drive test results exceeds 0.45; using hadoop The cluster performs data cleaning, query and processing. It can custo...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreAbstract
Embodiments of the invention disclose a VoLTE voice quality assessment method and system. The method includes the following steps of: collecting voice and signaling data at each interface after lightsplitting and gathering; analyzing the voice and the signaling data, and performing an average subjective opinion MOS value evaluation on the analyzed media stream to obtain voice quality assessment data; adding user information and location information to the voice quality assessment data according to control plane data to generate a voice quality assessment file; and storing the voice quality assessment file to a database server. Through the addition of the user information and the location information in the voice quality assessment data to generate the voice quality assessment file, inquiry is convenient, the VoLTE voice quality assessment method is suitable for all VoLTE periods, and especially for scenes that VoLTE dials VoLTE users and the VoLTE dials traditional 2/3G users, and assessment results are reliable and accurate by performing a MOS value assessment.
Application Domain
Technology Topic
Image
Examples
- Experimental program(1)
Example Embodiment
[0043] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solutions of the present invention more clearly, but not to limit the protection scope of the present invention.
[0044] figure 1 A schematic flow chart of a VoLTE voice quality assessment method provided in this embodiment is shown, including:
[0045] S101. Collect voice and signaling data after optical splitting and aggregation at each interface.
[0046] Wherein, the interfaces include two media stream interfaces (Mb, Nb) and three control plane interfaces (S11, Gm, Mc). Specifically, RTP voice packets can be obtained by collecting two interfaces Mb and Nb; corresponding control plane data can be obtained by collecting three interfaces S11, Gm and Mc.
[0047] The above 5 interfaces are all data converged on the CE side through 2/8 splitting. After splitting, 80% of the optical power is returned to the main link of the original operator all the way, and 20% of the optical power is connected to the TAP of this system all the way. Distribute equipment to perform data aggregation and convergence.
[0048] S102. Analyze the voice and signaling data, and evaluate the average subjective opinion score (MOS) value on the media stream obtained through the analysis, to obtain voice quality evaluation data.
[0049] Wherein, the P.563 algorithm of the ITU specification is a MOS quality assessment algorithm based on a no-reference model introduced by the ITU, which belongs to an internationally published standard algorithm.
[0050] Specifically, the voice and signaling data are analyzed to obtain the media streams of the above five interfaces, and the media streams of the Mb interface or the Nb interface are screened out through screening, and the P.563 algorithm is used for all RTP packets The MOS value scoring can realize efficient and fast query and processing of large amount of data.
[0051] S103. Add user information and location information to the voice quality assessment data according to the control plane data, and generate a voice quality assessment file according to the voice quality assessment data after adding the user information and location information.
[0052] Wherein, the control plane data is data received from three interfaces S11, Gm, and Mc, including user information and location information.
[0053] S104. Store the speech quality evaluation file in a database server.
[0054] Specifically, this embodiment is mainly based on the speech quality assessment problems of the following two scenarios:
[0055] Scenario 1: In the intercommunication between VoLTE users and VoLTE users, it is necessary to evaluate the voice quality of the Mb interface media stream. like figure 2 As shown, the scenario includes the EPC domain and the IMS domain. The EPC domain includes logical network elements such as MME, SGW, and PGW, and the IMS domain includes logical network elements such as P-CSCF, I/S-CSCF, BGCF, MGCF, VoLTE-SBC, IM-MGW, and CS-MGW. The Mb interface is the interface between the PGW and the VoLTE-SBC. This interface carries the media stream data of the EPC domain and the IMS domain. All media streams of the VoLTE network will be transmitted through this interface. By analyzing the RTP packets of this interface, the voice quality index of the VoLTE network can be obtained. The S11 and Gm interfaces carry the data of the control plane, which can fill the user information and location information for the media stream data of the Mb interface.
[0056] Scenario 2: In the intercommunication between VoLTE users and 2/3G users, it is necessary to evaluate the voice quality of media streams on the Nb interface. like image 3 As shown, the Nb interface is the interface between the IM-MGW and the CS-MGW, which mainly completes the interworking between the IMS domain and the CS domain user plane broadband and narrowband bearers and Codec encoding and decoding conversion. In the scenario where VoLTE users communicate with 2/3G users, when the VoLTE user's RTP media stream passes through this interface, its codec mode will change, and the user plane will change from broadband to narrowband. This conversion may cause the voice quality of VoLTE users to deteriorate, and it is necessary to perform voice quality evaluation and analysis on the RTP media stream of this interface. The control plane data of the Mc interface can fill the user information and location information for the media stream data of the Nb interface.
[0057] Specifically, as Figure 4 As shown, the system corresponding to this embodiment is composed of 6 major parts: acquisition server, processing server (MOS evaluation), loader server (data backfill), hadoop cluster (distributed processing), database server and application server. Distributed architecture, using three-layer switches for data intercommunication.
[0058] The system acquires data sources by means of local acquisition and decoding, performs 2/8 splitting on the CE side corresponding to each interface (including 5 interfaces of Mb, Nb, S11, Gm, and Mc), and uses the TAP splitter to split the split data. Aggregation, and then sent to each collection server according to the IP-based filtering method and load balancing principle. The collection server collects the original code streams of the above five interfaces, and then processes the original code streams through the processing server, only retaining the required processes and fields (for example, the S11 interface only retains the session_setup process). In the process of processing, the MOS value of the media streams of Mb and Nb interfaces is evaluated by the P.563 algorithm embedded in the platform, and then the association mechanism of equal ports and IPs is used by the Lorder server (see the detailed association mechanism below. Part 3) to correlate media stream and control plane data, and fill important user information (such as IMSI, MSISDN, IMEI, etc.) and location information (such as SGW, TAC, ENB, CI, etc.) for media stream data. After filling, the system will convert these data into text files in TXT format, and send these files to the hadoop cluster system, which will store and query these files, and import the query results to ORACLE as required, and finally The data application is based on ORACLE data.
[0059] This system needs to collect two interfaces Mb and Nb to obtain RTP voice packets, and at the same time it needs to collect three interfaces S11, Gm and Mc to obtain corresponding control plane data. see Figure 5 , first, after the system collects the RTP voice streams of the Mb and Nb interfaces through the acquisition server, it will decode these original streams, and then evaluate the MOS value of these voice packets through the P.563 algorithm embedded in the system. The RTP voice code stream itself has no user information and location information, and needs to be backfilled with related control plane data. Correlate the XDR of the Mb and Nb interfaces after MOS evaluation with the control plane data of S11, Gm, and Mc according to a certain association mechanism, and backfill the corresponding user data information such as IMSI and cell. The XDR of the Mb and Nb interfaces after associated backfilling already has user information and cell location information, which can meet the needs of various functional applications and thematic analysis.
[0060] In this embodiment, voice quality evaluation files are generated by adding user information and location information in the voice quality evaluation data, which is convenient for query, and is applicable to various periods of VoLTE, especially for VoLTE dialing VoLTE users, and VoLTE dialing traditional 2/3G users Scenarios, and by evaluating the MOS value, the evaluation results are reliable and accurate.
[0061] Further, on the basis of the above method embodiments, S101 specifically includes:
[0062] collecting the media streams at the Mb interface, the Nb interface, the S11 interface, the Gm interface, and the Mc interface, performing 2/8 splitting on the media streams, and converging 20% of the splitted media streams, The voice and signaling data are obtained.
[0063] Specifically, the system needs to collect two media stream interfaces (Mb, Nb) and three control plane interfaces (S11, Gm, Mc), Image 6 It is a schematic diagram of the interface that needs to be collected.
[0064] The above 5 interfaces are all data converged on the CE side through 2/8 splitting. After splitting, 80% of the optical power is returned to the main link of the original operator all the way, and 20% of the optical power is connected to the TAP of this system all the way. Distribute equipment to perform data aggregation and convergence. Figure 7 It is the schematic diagram of light splitting: after the data source is converged by the TAP device, the data is distributed to each acquisition server according to certain rules. The acquisition server adopts a distributed architecture and uses a three-layer switch for intercommunication. Multiple servers use protocols such as NTP for time precision synchronization to meet the time precision requirements of 1-50 milliseconds. The key function of the acquisition server is to connect the original code stream to the system, after preliminary processing, convert the file format that the cost system can recognize, and then send it to the processing server for more targeted processing.
[0065] In this embodiment, data sources are converged from the CE side of each interface by using 2/8 optical splitting. After convergence, the TAP distribution device will distribute the original code stream to each collection server. The collection server adopts a distributed architecture and uses a three-layer switch. Data exchange.
[0066] Further, on the basis of the above method embodiments, S102 specifically includes:
[0067] The voice and signaling data are analyzed, and the analyzed voice and signaling data are screened according to the preset process and preset fields to obtain the media stream of the Mb interface or the Nb interface, and adopt the P. The 563 algorithm evaluates the MOS value of the media stream to obtain voice quality evaluation data.
[0068] Specifically, the original code stream of each interface has multiple processes, and each process has several fields. In this embodiment, it is impossible to decode all the processes and fields of the interface, which will greatly increase the load of the system and affect the processing performance of the system. The processing server will filter the data sent by the collection server through process customization (for example, the S11 interface only defines the session_setup process, and the rest of the processes are not output) and field selection, removes unnecessary ones, and retains required ones Processes and fields. At the same time, through the P.563 algorithm embedded in the system, instead of requiring a separate algorithm server like other solutions, the MOS value is scored for the RTP packets of the Mb port and the Nb port, and the MOS value is obtained accordingly.
[0069] After processing server processing and MOS scoring, this embodiment will generate various XDR files and corresponding key fields as shown in the following table on each interface:
[0070]
[0071]
[0072] Further, on the basis of the above method embodiments, S103 specifically includes:
[0073] Add user information in the voice quality assessment data according to the session_setup table in the control plane data, add location information in the voice quality assessment data according to the sipcall table in the control plane data, and add user information and location information according to the The voice quality assessment data generates a voice quality assessment file.
[0074] Specifically, for the evaluation of voice quality, the voicecall_infost of the Mb interface and the Nb interface needs to be analyzed, but the XDR only has data related to voice quality, and does not have any user information and location information, which needs to be filled with data from the control plane.
[0075] (1) User data backfill mechanism of Mb port
[0076] The S11 port directly decodes to generate a session_setup table, which contains information such as IMSI and MSISDN; the Gm port generates a sipcall table, which contains LAC, CI and other information; the Mb port generates a voicecall_infost table, which has MOS values (evaluated by the previous P.563 algorithm ) and other voice quality index data, but the data about user information and cell location information cannot be directly obtained. These data need to be filled through the backfill mechanism. Among them, the cell location information needs to be obtained by associating the sipcall table, and user information such as IMSI needs to be obtained. Obtained indirectly through the session_setup table.
[0077] The process of backfilling user data is as follows: Figure 8 shown, including the following specific steps:
[0078] A1. Construct a relationship table between MSISDN and IMSI. The session_setup process of the S11 interface comes with MSISDN and IMSI, and the relationship table between MSISDN and IMSI can be obtained and automatically updated in real time.
[0079] A2. Fill the IMSI for the sipcall process. The sipcall process of the Gm interface can directly decode the location information of TAC and CI, with the MSISDN of the calling party and the called party, but lacks the IMSI, and needs to be backfilled through the S11 interface. The specific process is as follows:
[0080] A21. Use the MSISDN-IMSI relationship table constructed in the first step to fill the IMSI. When the call type is outgoing (that is, Call_Type=1), use sipcall.sip_from=MSISDN to fill in the association of IMSI; when the call type is inbound (that is, Call_Type=0), use sipcall.sip_to=MSISDN to fill in the association of IMSI . Note: The number formats of MSISDN, sip_from, and sip_to need to be normalized, otherwise the number format will affect the matching.
[0081] A22. After passing a, there are still a small number of users who fail to fill in the IMSI. For this part of the users, use the principle of equal user IP addresses to backfill. When the field User_IP_Value in the sipcall process of the Gm interface is equal to the field User_IP_Value_v6 in the Session_Setup process of the S11 interface, find the relevant Session_Setup record and fill in the IMSI.
[0082] A3. Backfill the user data of the Mb port. The XDR of voice_info can be obtained by directly decoding the RTP media stream of the Mb interface. According to the principle that the quaternion of the XDR is equal to the quaternion in the sipcall process of the Gm interface, the user data information (including IMSI, MSISDN, TAC, CI). Quadruple equality means: srcipv6, destipv6, srcport, and destport of voice_info are strictly equal to CallingVoice_Addr, CalledVoice_Addr, CallingVoice_Port, and CalledVoice_Port in the sipcall process of the Gm interface.
[0083] (2) User data backfill mechanism of Nb port
[0084] The direct decoding of the Mc port will eventually generate 3 XDRs, which are 3 original tables of H248, voicecall_moc and voicecall_mtc. H248 is for the control plane protocol of the Nb port; voicecall_moc and voicecall_mtc are the control plane data of the calling party and the called party of the Mc port respectively, which contain detailed information such as IMSI, MSISDN, LAC, and CI.
[0085] The Nb port directly decodes and generates the VOICE_INFOST table, which contains voice quality index data such as MOS value, but does not have any user information and location information. The user data of the VoLTE network needs to be backfilled through the control plane information of the Gm interface and the S11 interface, while the user data of the 2/3G network needs to be backfilled through the control plane information of the Mc interface.
[0086] The process of backfilling user data is as follows: Figure 9 As shown, it specifically includes the following steps:
[0087] B1. User data backfilling of the VoLTE network. The backfill mechanism here is the same as the above-mentioned user backfill mechanism of the Mb port, and will not be described again.
[0088] User data backfilling of B2 and 2/3G networks. There are two cases: the first one is for non-IP users of 2G network, and the second one is for users of 2/3G network IP.
[0089] B21. For the first type: when the triplets of H248 are strictly equal to the triplets of voicecall_moc and voicecall_mtc, fill in information such as IMSI, MSISDN, LAC, and CI for H248. The triplet here refers to: pcm, timeslot and opc.
[0090] B22. For the second type: when the 2-tuples of H248 are strictly equal to the 2-tuples of voicecall_moc and voicecall_mtc, fill in information such as IMSI, MSISDN, LAC, and CI for H248. The two-tuple here refers to: locconnaddr (local connection address) and loctransport (local connection port).
[0091] After B21 and B22, relevant backfilling of user information and location information has been completed for H248 form. Afterwards, when the VOICE_INFOST table is strictly equal to the H248 quadruple, fill the VOICE_INFOST table with information such as IMSI, MSISDN, LAC, and CI. The quadruple here refers to: srcip, srcport, destip and destport of the VOICE_INFOST table, locconnaddr, loctransport, remconnaddr and remtransport of the voicecall_moc and voicecall_mtc.
[0092] Through the principle of quaternion equality, the data association of multiple interfaces of the VoLTE network is carried out, and finally the key user information (IMSI/MSISDN) and location information (SGW/TAC/ENB/CI) are backfilled for the media stream data of the Mb port.
[0093] Further, on the basis of the above method embodiment, after S104, it also includes:
[0094] S105. Perform a real-time query on each speech quality evaluation file stored in the database server according to preset conditions, generate an aggregation table for a preset time period, and store the aggregation table in the database server.
[0095] Wherein, the preset conditions are preset according to specific requirements.
[0096] Specifically, after the original code stream is collected and decoded, MOS evaluated, and user data backfilled, it will be pushed to the hadoop cluster in TXT text format through the socket. Hadoop uses storm to receive and convert the TXT text files pushed by the system, and then compress them and send them to HDFS for storage according to a certain mechanism. Hive and Spark mainly query and process files stored in HDFS. The former focuses on non-real-time data query processing, while the latter focuses on real-time data query processing. The results after query processing can be imported into the ORACLE database for storage, while the front-end applications are more developed based on ORACLE. The hadoop ecosystem architecture diagram of the system is as follows Figure 10 Shown: The TXT file received by Storm and stored in HDFS is the original XDR (session_setup, sipcall, voicecall_infost, voicecall_mtc, voicecall_moc, and H248) decoded by each interface and the XDR (mb_voicecall and nb_voicecall) after data backfilling. The Hadoop cluster can perform corresponding data integration on the above XDRs, generate various custom XDRs, and import them into the ORACLE database, which is convenient for the application server to call. In this system, the hadoop cluster mainly completes the following data query functions:
[0097] (1) Non-real-time data query.
[0098] Use the Hive function to purposely query the XDR text files stored in HDFS, and directly display the query results, export them in TXT/CSV format or import them into ORACLE.
[0099] Application Scenario 1: It is necessary to temporarily and randomly query the voice quality index data of some users or cells. The amount of data to be queried is large, the query time is long, and the real-time requirements are not high. This type of scenario belongs to instant query, which is obviously temporary and random.
[0100] Application Scenario 2: Based on the two XDR text files voicecall_mb and voicecall_nb, various aggregation table XDRs can be customized and automatically generated, and the results can be imported into ORACLE. For example, a day granularity aggregation table (mb_imsi_day or mb_ci_day) of user or cell dimension may be generated based on voicecall_mb.
[0101] (2) Real-time data query.
[0102] Use the Spark function to query XDR text files efficiently and quickly, generate aggregation table XDR with various small time granularities (such as 5-minute or 15-minute granularity), and import them into ORACLE. For example, based on voicecall_nb, a 15-minute aggregation table (nb_imsi_15min or nb_ci_15min) at the user or cell level can be generated and imported into ORACLE. The application program can display the voice quality indicators of TOP N users or TOP N cells in real time by calling these aggregation tables in ORACLE.
[0103] This embodiment can solve voice quality evaluation analysis in two scenarios of intercommunication between VoLTE users and VoLTE users, and intercommunication between VoLTE users and 2/3G users. The following will focus on these two scenarios to understand various application analyzes using this method.
[0104] Scenario 1: VoLTE users communicate with VoLTE users.
[0105] This scenario is based on the media stream of the Mb port, and the application analysis can be performed based on the XDR data of voicecall_mb that has completed data backfilling. The voice quality evaluation and analysis of this scenario is mainly performed through the following applications.
[0106] Application 1: Real-time monitoring of TOP N cells
[0107] Using the Hive function, the 15-minute granular aggregation table mb_ci_15min of the community dimension is generated in real time based on voicecall_mb, and imported into the ORACLE database. The application server calls mb_ci_15min through a custom task to execute the corresponding program, and generates TOP N cells with the worst MOS value every 15 minutes. These TOP N cells can be rendered and presented in the form of maps, and reports can also be generated or imported into the ORACLE database .
[0108]Based on mb_ci_15min, the aggregation tables mb_ci_hour and mb_ci_day with hour granularity and day granularity can be easily obtained.
[0109] Application 2: NE-level real-time monitoring
[0110] Based on the voicecall_mb table, the aggregation table mb_ne_15min of the network element dimension and 15-minute granularity is generated in real time. The network elements in the aggregation table may include TACs, SGWs, and townships. Based on the aggregation table, the voice quality indicators of TAC, SGW, and towns can be monitored in real time, such as MOS value, jitter, delay, packet loss rate, etc.
[0111] Application 3: TOP N user analysis
[0112] Based on voicecall_mb, the user dimension hour-grained or day-grained aggregation table mb_imsi_hour or mb_imsi_day is generated in real time. Using this aggregation table, the TOP N user list with the worst MOS value can be automatically generated, and the cell information of the TOP N users can be obtained to analyze whether it is caused by a problem on the wireless side of the cell.
[0113] Application 4: User Complaint Analysis
[0114] For temporary and random batches of user complaint numbers, use Hive to query voicecall_mb to obtain the information of the community where the complaining user is located 2 hours before the complaint time (this time can be customized) and the corresponding MOS value index, and analyze whether the reason for the complaint is related to the wireless environment of the community related.
[0115] Scenario 2: VoLTE users communicate with 2/3G users.
[0116] This scenario is based on the media flow of the Nb interface (the interface between the IM-MGW and the CS-MGW), and the application analysis can be performed based on the XDR data of the voicecall_nb that has completed data backfilling. The voice quality evaluation and analysis of this scenario is mainly performed through the following applications.
[0117] Application 1: Voice quality analysis of VoLTE network users based on Nb port
[0118] In the XDR of voicecall_nb, the rat_type field represents the network type, and the values of this field are 2, 3, and 4, representing 2G, 3G, and 4G networks respectively. When rat_type=4, the overall voice quality of the VoLTE user at the Nb interface can be analyzed in the scenario where the VoLTE user communicates with the 2/3G user.
[0119] Application 2: Difference analysis of voice quality based on Mb port and Nb port
[0120] Both the voicecall_mb of the Mb interface and the voicecall_nb of the Nb interface have backfilled IMSI and MSISDN fields. Through this field, the same VoLTE users can be found on the two interfaces, and the voice quality of the Mb interface and the Nb interface are evaluated for these users respectively. Analyze the differences.
[0121] In this embodiment, based on the data of the Mb interface, a comprehensive voice quality assessment is performed for a scenario in which a VoLTE user communicates with a VoLTE user. The existing methods have not realized the collection of the interface, reasonable server combination and algorithm, and analysis and processing technology of massive data. This embodiment can evaluate the voice quality from the dimensions of users, terminals, cells, and TAC, and analyze the impact of various dimensions. Factors of voice quality, and find optimization methods; data based on Nb interface is aimed at the scenario where VoLTE users communicate with 2/3G users. In the current few years, this scenario accounts for more than 90%. It innovatively proposes the collection, storage and analysis based on IMGW to evaluate the overall voice quality of VoLTE users on Nb ports; evaluate the voice quality of Mb ports and Nb ports for the same VoLTE users at the same time. Analyze the differences.
[0122] This embodiment carries out association analysis for the data of multiple interfaces (comprising Mb, Nb, S11, Gm and Mc interface); Adopt distributed framework, utilize the combination of acquisition server, processing server, loader server, hadoop cluster, database server and application server , can process the data of the above five interfaces in real time (15-minute granularity presentation), massive (2W users in the whole network), and efficient (minute-level query is completed); for the voice quality evaluation of RTP media stream, P.563 of ITU standard is adopted Algorithms are used to evaluate, and the evaluation results are reliable and accurate. Compared with the traditional drive test results, the fitting degree of the MOS results based on the P.563 algorithm reaches 0.82, and the average error is less than 0.3, while the average error between the MOS results based on the E-Model model algorithm and the traditional drive test results exceeds 0.45; using hadoop The cluster performs data cleaning, query and processing. It can customize and generate various reports and aggregation tables, and the aggregation tables can be imported into the ORACLE database. The custom aggregation table can reduce the load of ORACLE. Compared with directly storing the original table, this method can significantly improve the I/O performance of ORACLE; the function is widely used, and it can be used for intercommunication between VoLTE users and VoLTE users (Mb interface media real-time (15-minute granularity), network-wide voice quality analysis in two scenarios, VoLTE user and 2/3G user interworking (Nb interface media flow).
[0123] Figure 11 A schematic structural diagram of a VoLTE voice quality assessment system provided in this embodiment is shown, and the system includes: an acquisition server 1101, a processing server 1102, a data backfill loader server 1103, a distributed processing hadoop cluster system 1104, and a database server 1105;
[0124] The collection server 1101 collects voice and signaling data after optical splitting and aggregation at each interface, and sends the voice and signaling data to the processing server 1102;
[0125] The processing server 1102 analyzes the voice and signaling data, and evaluates the average subjective opinion score MOS value of the media stream obtained by the analysis, obtains voice quality assessment data, and sends the voice quality assessment data to the loader server 1103;
[0126] The loader server 1103 adds user information and location information to the voice quality evaluation data according to the control plane data, and generates a voice quality evaluation file according to the voice quality evaluation data after adding user information and location information, and sends the Voice quality evaluation file is sent to hadoop cluster system 1104;
[0127] The hadoop cluster system 1104 stores the speech quality evaluation file to a database server;
[0128] The database server 1105 is used for storing the voice quality assessment file.
[0129] In this embodiment, voice quality evaluation files are generated by adding user information and location information in the voice quality evaluation data, which is convenient for query, and is applicable to various periods of VoLTE, especially for VoLTE dialing VoLTE users, and VoLTE dialing traditional 2/3G users Scenarios, and by evaluating the MOS value, the evaluation results are reliable and accurate.
[0130] Further, on the basis of the above-mentioned device embodiment, the collection server 1101 is specifically configured to collect the media streams at the Mb interface, the Nb interface, the S11 interface, the Gm interface, and the Mc interface, and process the media streams 2/8 of the light is split, and the 20% of the split media streams are aggregated to obtain the voice and signaling data.
[0131] Further, on the basis of the above device embodiments, the processing server 1102 is specifically configured to analyze the voice and signaling data, and analyze the voice and signaling data after analysis according to preset procedures and preset fields. Screening is carried out to obtain the media stream of the Mb interface or the Nb interface, and the P.563 algorithm of the ITU standard is used to evaluate the MOS value of the media stream to obtain voice quality evaluation data.
[0132] Further, on the basis of the above device embodiment, the loader server 1103 is specifically configured to add user information to the voice quality evaluation data according to the session_setup table in the control plane data, and add user information to the voice quality evaluation data according to the sipcall table in the control plane data. Add location information to the voice quality evaluation data, and generate a voice quality evaluation file according to the voice quality evaluation data after adding user information and location information.
[0133] Further, on the basis of the above-mentioned device embodiment, the hadoop cluster system 1104 is also used to perform real-time query on each speech quality evaluation file stored in the database server according to preset conditions, and generate a collection table for a preset time period , and store the aggregation table in the database server.
[0134] The speech quality evaluation and storage system described in this embodiment can be used to implement the above method embodiments, and its principles and technical effects are similar, and will not be repeated here.
[0135] refer to Figure 12 , the electronic device includes: a processor (processor) 1201, a memory (memory) 1202 and a bus 1203;
[0136] in,
[0137] The processor 1201 and the memory 1202 communicate with each other through the bus 1203;
[0138] The processor 1201 is configured to invoke program instructions in the memory 1202 to execute the methods provided by the above method embodiments, for example, including:
[0139] Collect voice and signaling data after optical splitting and aggregation at each interface;
[0140] Analyzing the voice and signaling data, and evaluating the average subjective opinion score MOS value on the media stream obtained by the analysis, to obtain voice quality evaluation data;
[0141] Adding user information and location information to the voice quality assessment data according to the control plane data, and generating a voice quality assessment file according to the voice quality assessment data after adding the user information and location information;
[0142] The voice quality evaluation file is stored in a database server.
[0143] This embodiment discloses a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed by the computer, the computer The methods provided by the above-mentioned method embodiments can be implemented, for example, including:
[0144] Collect voice and signaling data after optical splitting and aggregation at each interface;
[0145] Analyzing the voice and signaling data, and evaluating the average subjective opinion score MOS value on the media stream obtained by the analysis, to obtain voice quality evaluation data;
[0146] Adding user information and location information to the voice quality assessment data according to the control plane data, and generating a voice quality assessment file according to the voice quality assessment data after adding the user information and location information;
[0147] The voice quality evaluation file is stored in a database server.
[0148] This embodiment provides a non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the methods provided in the foregoing method embodiments, for example including :
[0149] Collect voice and signaling data after optical splitting and aggregation at each interface;
[0150]Analyzing the voice and signaling data, and evaluating the average subjective opinion score MOS value on the media stream obtained by the analysis, to obtain voice quality evaluation data;
[0151] Adding user information and location information to the voice quality assessment data according to the control plane data, and generating a voice quality assessment file according to the voice quality assessment data after adding the user information and location information;
[0152] The voice quality evaluation file is stored in a database server.
[0153] Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.
[0154] The device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in One place, or it can be distributed to multiple network elements. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. It can be understood and implemented by those skilled in the art without any creative effort.
[0155] Through the above description of the implementations, those skilled in the art can clearly understand that each implementation can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic discs, optical discs, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.
[0156] It should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it still can The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more PUM


Description & Claims & Application Information
We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more Similar technology patents
Fusion query method based on heterogeneous data source and distributed file system
InactiveCN106970943AConvenient queryImplement Fusion QuerySpecial data processing applicationsDistributed File SystemData ingestion
Owner:南京中新赛克科技有限责任公司
Mobile information intelligent nursing station
InactiveCN103110461AConvenient queryReduce doctor-patient disputesEnergy efficient ICTSurgeryFingerprintLocal area network
Owner:SUZHOU DERPIN MEDICAL SCI & TECH CO LTD
Medical information system
InactiveCN102419797AConvenient queryConvenient consultationSpecial data processing applicationsPatient informationInformation system
Owner:苏州伟士立德管理咨询有限公司
Industrial waste material storage and management system
ActiveCN110803434AConvenient queryConvenient inventoryStorage devicesIndustrial wasteProcess engineering
Owner:全威 +1
Non-contact measuring device for automatically detecting regular surface deformation
Owner:上海米度测量技术有限公司 +1
Classification and recommendation of technical efficacy words
- Evaluation results are reliable
- Convenient query
Safety warning method of electric vehicle power battery
ActiveCN110133508AEvaluation results are reliableElectrical testingVehicular energy storageElectric vehicleRemaining life
Owner:SHANGHAI RICHPOWER MICROELECTRONICS
Detailed terrain data-based transmission line failure-shielding and lightning-protection performance evaluation method
ActiveCN102072992AEvaluation results are reliableImprove lightning protection performanceElectrical testingAssessment methodsTerrain
Owner:WUHAN NARI LIABILITY OF STATE GRID ELECTRIC POWER RES INST
Assessment method for rainstorm disaster risks of sectional power transmission line pole-tower foundation slopes
ActiveCN102930348AEvaluation results are reliableImprove disaster resilienceForecastingAssessment methodsEngineering
Owner:ELECTRIC POWER RES INST OF GUANGDONG POWER GRID
Equipment evaluation method and device based on IOT (Internet of Things), and computer readable storage medium
ActiveCN109120451AEvaluation results are reliableData switching networksInternet of ThingsDevice Evaluation Method
Owner:深圳市麦斯杰网络有限公司
Method for evaluating service condition of high-temperature alloy turbine blade
ActiveCN110411850AExclude complex effectsEvaluation results are reliableMaterial strength using tensile/compressive forcesNeural architecturesAssessment methodsTurbine blade
Owner:UNIV OF SCI & TECH BEIJING
Two-dimension code-based shopping method
Owner:广州闪购软件服务有限公司
Food safety tracing system and method
Owner:ANHUI JINGHUI CAILANZI E COMMERCE CO LTD
Instrument for monitoring gas emission quantity in roadway
Owner:SHANDONG UNIV OF SCI & TECH
Flood forecasting method for area without runoff data
ActiveCN103488871AConvenient queryEasy to analyzeSpecial data processing applicationsModel parametersCorrelation analysis
Owner:STATE GRID CORP OF CHINA +2
Intelligent district log system and log recording method thereof
InactiveCN102195795AImprove processing efficiencyConvenient queryData switching networksUser Datagram ProtocolProcess module
Owner:TCL CORPORATION