CN101072174A - Tencent voice identifying method based on pay load deep detection and session correlating technology - Google Patents
Tencent voice identifying method based on pay load deep detection and session correlating technology Download PDFInfo
- Publication number
- CN101072174A CN101072174A CNA200710021025XA CN200710021025A CN101072174A CN 101072174 A CN101072174 A CN 101072174A CN A200710021025X A CNA200710021025X A CN A200710021025XA CN 200710021025 A CN200710021025 A CN 200710021025A CN 101072174 A CN101072174 A CN 101072174A
- Authority
- CN
- China
- Prior art keywords
- tengxun
- session
- grouping
- voice
- conversation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Telephonic Communication Services (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Based on payload depth detection and conversation association technique, Tengxun voice conversation recognition (VCR) method is composed of Tengxun conversation recognition (CR) method and Tengxun VCR method. First, using Tengxun CR method recognizes Tengxun conversations. Then, using payload characteristic of Tengxun voice conversation recognizes out all possible Tengxun voice conversations. Afterwards, using the conversation association technique carries out associative analysis between possible Tengxun voice conversations and each Tengxun conversation. Thus, using dual recognitions of Tengxun conversation procedure and Tengxun voice conversation procedure guarantee accuracy of system and success ratio of recognition. The method possesses good expandability and accuracy, and is easy to connect to corresponding application interface of operation manager easily.
Description
Technical field
The present invention be directed to the Internet instant communication software-Tencent QQ and carry out the research of voice flux recognition methods, how main research effectively discerns the QQ speech business based on detection of DPI pay load deep and session association technology, and designed the model of cognition and the algorithm of QQ speech business, relate to the protocal analysis technical field of professional perception of IPv4 and Next Generation Internet and services quality monitoring etc.
Background technology
Along with the development of Internet technology with popularize, change has also taken place in people's communication mode, and traditional communication mode is replaced by network service gradually.Utilize bitcom, not only can carry out text chat, also can chat with voice or video, because easy to use and rate are cheap, increasing people carries out the remote live session by network.A large amount of unlawful VoIP operations are flooded with regular telecommunication market under the attraction of huge interests, not only cause legal operator telephone traffic to run off, more broken the competition situation of original telecommunication market, brought huge impact for traditional voice service, telecom operators are just suffering unlawful VoIP to its huge challenge that brings, and therefore are necessary very much VoIP business on the Internet is included in the category of optimum control.
QQ as one of the most popular instant communication software of present China, though the talking mode of PC2PC only is provided at present, but the Traffic Distribution that brings has thus also caused certain impact to telecom operators, therefore provide the Virtual network operator of voice service wish naturally can optimum control QQ the conversation behavior.So identification and optimum control thereof to QQ and QQ speech business not only are of great importance to telecom operators, and help other MSN is supervised.
Identification at QQ and QQ speech business possesses certain degree of difficulty, and following reason is arranged:
The first, the communication protocol of QQ is underground, and wherein the part signaling has been used cryptographic algorithm.
The second, the version of QQ is numerous, and it is more frequent to upgrade, and different with most softwares be that the upgrading of its client often is accompanied by agreement and changes accordingly.
Three, most of now research to QQ concentrates on QQ and lands on the process of withdrawing from and the text chat interactive mode, rarely has the analysis to its voice process, so it is few to use for reference part.
Four, Tencent QQ adopts the port camouflage, uses 80 ports; But port random arrangement; Server has a plurality of unfixed IP address, is difficult to accomplish complete shutoff;
Five, QQ provides business such as text, data, voice, video, and the session characteristics of miscellaneous service is all inequality, and therefore " cruelty " shutoff to server ip address is not the basic way of dealing with problems, and this can cause normal QQ communication to use.
Therefore, adopt business recognition methods such as traditional ports filter, IP address filtering and protocal analysis to be difficult to identify the voice process of QQ.Therefore, must look for another way.
By the identification to QQ voice and session thereof, we can solve following problem:
(1) concerning telecom operators, can carry out statistical analysis to the QQ speech business, be convenient to control of the influence of QQ audio call to traditional voice transmission;
(2) can make operator to QQ speech business implement optimum supervision, as formulate rational charging policy to ensure the interests of traditional voice transmission;
(3) consider from the angle of national information safety, can implement real-time listening the QQ voice, effectively prevent unlawful activities by QQ as communication medium.
Summary of the invention
Technical problem: the objective of the invention is to set up a kind of Tengxun's audio recognition method based on pay load deep detection and session association technology, and design its model of cognition and algorithm, by identification to the QQ speech business, QQ voice signaling traffic and Media Stream are sorted out from QQ session stream, be convenient to analyze the details such as calling IP address, encoding and decoding speech type, QQ voice server address of both call sides, thus the dialog context of analysis QQ voice that can be more deep.
Technical scheme: the present invention proposes the technological frame of a kind of effective identification QQ speech business, and detailed design recognizer, as shown in Figure 1.As can be seen from the figure, system is divided into four levels, is successively from the bottom up: data collection layer, protocal analysis layer, flow identification (professional perception) layer and QQ speech service application layer and presentation layer.
Here need to distinguish the term of two this paper definition.All QQ interbehaviors after QQ session general reference user lands comprise that the user lands, authentication, text chat, voice call, video session, 3D recreation, withdraw from or the like QQ reciprocal process.And the voice and video communication process in the QQ session is refered in particular in the QQ voice call.Therefore, the corresponding QQ session of QQ number, the QQ voice call is refered in particular to a QQ user with another QQ user's voice communication process.
The key method of this paper is at the flow identification layer, and this layer mainly comprises two methods: QQ session recognition methods and QQ voice conversation correlating method.By at first identifying the QQ conversation, carry out the association of QQ voice conversation again and determine real QQ voice conversation.By test and data analysis, find that the QQ session possesses certain payload characteristic, data packet format is divided into head in QQ login process or the connection request process, and three parts of content and afterbody are fixed as: 0 * 02 client release command sequence QQ number content 0 * 03.Therefore can identify the grouping of QQ session by DPI pay load deep testing mechanism according to initial sum end payload characteristic 0 * 02 and 0 * 03.According to request login token, identifying packet the 8th to the 11 byte is caller QQ number, to identify a QQ session again.As shown in Figure 2.And the reciprocal process of test QQ voice communication finds that voice connect when setting up, and also has the feature of initial sum end payload 0 * 02/0 * 03, afterwards, then adopts the communication interaction mechanism of similar Session Initiation Protocol to set up voice conversation.Therefore also can adopt pay load deep testing mechanism and simple protocal analysis technology to discern the voice conversation of QQ.Payload characteristic coupling string is " SIP/user-agent:Tencent-VQQ " or " SIP/reason=100 " etc.As shown in Figure 3.
Yet the change of QQ version or the change of agreement all can bring the variation of QQ payload characteristic, therefore also will inevitably make above-mentioned recognition methods that certain variation takes place.How can not change system and be a major challenge of algorithm with regard to the adaptation of finishing QQ new business feature by simple configuration.Regular expression is an extraordinary solution just, and this method adopts regular expression to show the feature of session characteristics and the voice conversation of QQ.Therefore change or feature changes when the QQ version, the feature configuration file that this algorithm only needs simply to revise regular expression gets final product, and need not to remodify code and method and promptly accomplishes rapidly and efficiently renewal.
Below introduce each aspect and the voice conversation recognition methods thereof of this design in detail.
1. data collection layer
Function: this aspect provides for the data acquisition of different links or reproduction technology, as the collection or the reproduction technology of 100/1000M FE, ATM, SDH different rates, to ensure data integrity, to be sent to last layer face one protocal analysis layer reliably.
Interface: the interface of this aspect and last layer face is a bitstream data, provides various grouping informations to the upper strata.
2. protocal analysis layer
Function: this aspect provides the protocol analysis for the TCP/IP data, purpose is for enough IP packet header and the header information of TCP/UDP and the packet payload information of necessity thereof is provided to the upper strata, to satisfy identification and the perception of last layer surface current amount identification layer to business.
Interface: the protocal analysis degree of depth of this aspect should be analyzed to the 4th layer of the ICP/IP protocol stack, i.e. transport layer.It is stream (flow) to the interface that the upper strata provides.Stream should be determined by a five-tuple, i.e. flow=(source IP, purpose IP, source port, destination interface, protocol type).Protocol type herein refers to TCP or UDP.If necessary, but storage part payload also in this stream, and the payload size of catching is configurable.
3. flow is discerned (professional perception) layer
Function: this aspect is the core aspect of whole framework, and main is that features such as the header information of the IP packet header that provides of protocal analysis layer and TCP/UDP and payload information thereof effectively identify the QQ business according to lower floor is provided, and the grouping that it fails to match then abandons.
Interface: the interface that provides to application should be a five-tuple, and promptly (source IP, purpose IP, source port, destination interface is used details).
This layer mainly comprises two methods: QQ session recognition methods and QQ speech business identification and session association method.By at first identifying the QQ conversation, carry out the association of QQ voice conversation again and determine real QQ voice conversation.
◆ QQ session recognition methods.The method processing procedure is as follows
1.) initialization Hash table: this Hash table is to be used to store Tengxun's session identification, it is Tengxun's session id, this sign is represented with Tengxun's number and its IP address two tuples, Tengxun's number can only be corresponding to an IP address, elements all in the Hash table are initialized as 0, and promptly the IP address of all Tengxun's number correspondences is initialized as 0;
2.) receive the grouping of the IP network that will monitor;
3. whether, to judge this grouping be Tengxun conversation, judge that more whether this grouping is that token packet is landed in the request of Tengxun's session according to Tengxun's session payload characteristic if) carrying out DPI and detecting; In this way, then obtain Tengxun's number, change step 4); As it fails to match, abandon grouping, change step 2);
4.) judge this session whether Already in the Hash table, if, then abandon grouping, change step 2); If not, change step 5);
5.) preserve Tengxun's session identification, session identification is made up of login IP address two tuples of Tengxun's number and this Tengxun's number;
6) Tengxun's session is discerned successfully, finishes.
Detect and the session association method identifies the voice conversation of Tengxun by pay load deep, its method step is:
1.) receive grouping: this receiving course is same process with Tengxun's conversation procedure, is that same packet copies is used for the different grouping characteristic matching afterwards;
2.) at the grouping that receives: the class SIP characteristic according to Tengxun's speech payload is carried out characteristic matching, as the match is successful, then changes step 3); Otherwise, abandon grouping, change step 1);
3.) Tengxun's voice conversation is carried out related identification with Tengxun's session: owing to only can't judge fully also that by the signature analysis of Tengxun's speech payload this grouping is exactly Tengxun's voice conversation grouping, therefore this Tengxun's voice conversation grouping must be carried out association with the session of existing Tengxun detects, exist as this Tengxun's session, then the judgement of this Tengxun's packets of voice will may be accurately greatly; Concrete association process promptly with Tengxun's calling number of obtaining in this packets of voice as key, in Tengxun's session Hash table, inquire about, as the element that checks out is an IP address, prove that so this Tengxun's session is to exist, continue relatively IP address, if identical, then this voice conversation of decidable belongs to this Tengxun's session, changes step 4); If inequality, illustrate that then this Tengxun's voice conversation is not real voice conversation, abandon grouping, change step 1);
4.) preservation and renewal Tengxun voice conversation information: calling and called Tengxun address and port, calling and called Tengxun number, encoding and decoding speech type, the information of initiation time, end of calling time of calling out of Tengxun's voice conversation are preserved; When the grouping of other these voice conversations arrives, upgrade relevant information accordingly, form the call detail record CDR of Tengxun's voice conversation;
5) Tengxun's voice conversation is discerned successfully, finishes.
4.QQ speech service application layer and presentation layer
Identification for the QQ speech business has very wide significance and using value.Mainly can be applied in:
◆ QQ speech business traffic statistics are analyzed;
◆ QQ speech business performance evaluation;
◆ control of QQ voice flux and call follow;
◆ the estimation of QQ rate factor of influence:
◆ QQ voice flux abnormality detection;
◆ the security monitoring of QQ voice messaging.
Description of drawings
Fig. 1 is QQ session identification process figure.Provided each processing procedure of identification QQ session among the figure.
Fig. 2 is QQ speech business identification and session association method flow diagram.Provided each processing procedure of QQ speech business identification and session association method among the figure.
Fig. 3 is Tengxun's voice conversation recognition methods technological frame structural representation.
Embodiment
The VoIP detection system of developing according to this method has obtained concrete checking on the 10G backbone network of Guangxi branch company of China Telecom.System adopts the beam split mode that the 10G flow load balance is branched on some the traffic identification processors, and the traffic identification processor is finished the realization of core algorithm, extracts from the grouping of numerous and complicated, analyzes, the voice conversation of identification and the related QQ of going out.
By actual motion and test calls at Guangxi telecommunications 10G backbone network, be 100% at the recognition accuracy of QQ speech business, well embodied the implementation result of algorithm, verified algorithm accuracy.
QQ voice monitoring system is divided into entities such as light-dividing device, QQ voice monitoring equipment, core database server and application server.The 10G flow is divided toward some QQ voice monitoring server apparatus by light-dividing device, the flow of every QQ voice monitoring server apparatus carrying gigabit identifies after the service traffics, and business information is sent to core database in real time, and by application server issue, access topology.
System's access way is divided into two kinds: a kind of for series model, and be about to QQ voice monitoring system and connect and implement to detect and control in the backbone network; Another kind is a paralleling model, promptly adopts the mode of monitoring to finish and detects and control.Series model can influence whole network topology, and more or less can be for legacy network brings hidden danger, and the paralleling model of therefore more recommending legacy network is had no effect inserts.
The light-dividing device of system is after real-time beam split is got off from the 10G link, it is divided into the several flow points to some QQ watch-dogs, watch-dog adopts high performance flow collection technology to receive all flows, and call QQ speech business identification engine automatically flow is carried out real-time identification, and control according to user-defined strategy, as shutoff, interference or clearance etc.
Detect and the step of Tengxun's audio recognition method of session association technology is based on pay load deep:
1.) initialization Hash table: this Hash table is to be used to store Tengxun's session identification, it is Tengxun's session id, this sign is represented with Tengxun's number and its IP address two tuples, Tengxun's number can only be corresponding to an IP address, elements all in the Hash table are initialized as 0, and promptly the IP address of all Tengxun's number correspondences is initialized as 0;
2.) receive the grouping of the IP network that will monitor;
3. whether, to judge this grouping be Tengxun conversation, judge that more whether this grouping is that token packet is landed in the request of Tengxun's session according to Tengxun's session payload characteristic if) carrying out DPI and detecting; In this way, then obtain Tengxun's number, change step 4); As it fails to match, abandon grouping, change step 2);
4.) judge this session whether Already in the Hash table, if, then abandon grouping, change step 2); If not, change step 5);
5.) preserve Tengxun's session identification, session identification is made up of login IP address two tuples of Tengxun's number and this Tengxun's number;
6) Tengxun's session is discerned successfully, finishes.
Detect and the session association method identifies the voice conversation of Tengxun by pay load deep, its method step is:
1.) receive grouping: this receiving course is same process with Tengxun's conversation procedure, is that same packet copies is used for the different grouping characteristic matching afterwards;
2.) at the grouping that receives: the class SIP characteristic according to Tengxun's speech payload is carried out characteristic matching, as the match is successful, then changes step 3); Otherwise, abandon grouping, change step 1);
3.) Tengxun's voice conversation is carried out related identification with Tengxun's session: owing to only can't judge fully also that by the signature analysis of Tengxun's speech payload this grouping is exactly Tengxun's voice conversation grouping, therefore this Tengxun's voice conversation grouping must be carried out association with the session of existing Tengxun detects, exist as this Tengxun's session, then the judgement of this Tengxun's packets of voice will may be accurately greatly; Concrete association process promptly with Tengxun's calling number of obtaining in this packets of voice as key, in Tengxun's session Hash table, inquire about, as the element that checks out is an IP address, prove that so this Tengxun's session is to exist, continue relatively IP address, if identical, then this voice conversation of decidable belongs to this Tengxun's session, changes step 4); If inequality, illustrate that then this Tengxun's voice conversation is not real voice conversation, abandon grouping, change step 1);
4.) preservation and renewal Tengxun voice conversation information: calling and called Tengxun address and port, calling and called Tengxun number, encoding and decoding speech type, the information of initiation time, end of calling time of calling out of Tengxun's voice conversation are preserved; When the grouping of other these voice conversations arrives, upgrade relevant information accordingly, form the call detail record CDR of Tengxun's voice conversation;
5) Tengxun's voice conversation is discerned successfully, finishes.
Claims (2)
1. one kind is detected based on pay load deep and Tengxun's audio recognition method of session association technology, it is characterized in that steps of the method are:
1.) initialization Hash table: this Hash table is to be used to store Tengxun's session identification, it is Tengxun's session id, this sign is represented with Tengxun's number and its IP address two tuples, Tengxun's number can only be corresponding to an IP address, elements all in the Hash table are initialized as 0, and promptly the IP address of all Tengxun's number correspondences is initialized as 0;
2.) receive the grouping of the IP network that will monitor;
3. whether, to judge this grouping be Tengxun conversation, judge that more whether this grouping is that token packet is landed in the request of Tengxun's session according to Tengxun's session payload characteristic if) carrying out DPI and detecting; In this way, then obtain Tengxun's number, change step 4); As it fails to match, abandon grouping, change step 2);
4.) judge this session whether Already in the Hash table, if, then abandon grouping, change step 2); If not, change step 5);
5.) preserve Tengxun's session identification, session identification is made up of login IP address two tuples of Tengxun's number and this Tengxun's number;
6) Tengxun's session is discerned successfully, finishes.
2. one kind is detected based on pay load deep and Tengxun's voice conversation recognition methods of session association technology, it is characterized in that identifying by pay load deep detection and session association method the voice conversation of Tengxun, and its method step is:
1.) receive grouping: this receiving course is same process with Tengxun's conversation procedure, is that same packet copies is used for the different grouping characteristic matching afterwards;
2.) at the grouping that receives: the class SIP characteristic according to Tengxun's speech payload is carried out characteristic matching, as the match is successful, then changes step 3); Otherwise, abandon grouping, change step 1);
3.) Tengxun's voice conversation is carried out related identification with Tengxun's session: owing to only can't judge fully also that by the signature analysis of Tengxun's speech payload this grouping is exactly Tengxun's voice conversation grouping, therefore this Tengxun's voice conversation grouping must be carried out association with the session of existing Tengxun detects, exist as this Tengxun's session, then the judgement of this Tengxun's packets of voice will may be accurately greatly; Concrete association process promptly with Tengxun's calling number of obtaining in this packets of voice as key, in Tengxun's session Hash table, inquire about, as the element that checks out is an IP address, prove that so this Tengxun's session is to exist, continue relatively IP address, if identical, then this voice conversation of decidable belongs to this Tengxun's session, changes step 4); If inequality, illustrate that then this Tengxun's voice conversation is not real voice conversation, abandon grouping, change step 1);
4.) preservation and renewal Tengxun voice conversation information: calling and called Tengxun address and port, calling and called Tengxun number, encoding and decoding speech type, the information of initiation time, end of calling time of calling out of Tengxun's voice conversation are preserved; When the grouping of other these voice conversations arrives, upgrade relevant information accordingly, form the call detail record CDR of Tengxun's voice conversation;
5) Tengxun's voice conversation is discerned successfully, finishes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA200710021025XA CN101072174A (en) | 2007-03-23 | 2007-03-23 | Tencent voice identifying method based on pay load deep detection and session correlating technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA200710021025XA CN101072174A (en) | 2007-03-23 | 2007-03-23 | Tencent voice identifying method based on pay load deep detection and session correlating technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101072174A true CN101072174A (en) | 2007-11-14 |
Family
ID=38899169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA200710021025XA Pending CN101072174A (en) | 2007-03-23 | 2007-03-23 | Tencent voice identifying method based on pay load deep detection and session correlating technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101072174A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101582897A (en) * | 2009-06-02 | 2009-11-18 | 中兴通讯股份有限公司 | Deep packet inspection method and device |
| CN101621587B (en) * | 2008-06-30 | 2012-08-08 | 成都市华为赛门铁克科技有限公司 | Method, device and system for network monitoring |
| CN102932817A (en) * | 2012-10-10 | 2013-02-13 | 福建星网锐捷网络有限公司 | Method and device for recognizing voice stream and wireless access equipment |
| US8416695B2 (en) | 2008-06-30 | 2013-04-09 | Huawei Technologies Co., Ltd. | Method, device and system for network interception |
| CN103400579A (en) * | 2013-08-04 | 2013-11-20 | 徐华 | Voice recognition system and construction method |
| CN108206788A (en) * | 2016-12-16 | 2018-06-26 | 中国移动通信有限公司研究院 | The business recognition method and relevant device of a kind of flow |
| CN110519169A (en) * | 2019-08-30 | 2019-11-29 | 成都安恒信息技术有限公司 | A kind of method of application layer multiplexed network header |
-
2007
- 2007-03-23 CN CNA200710021025XA patent/CN101072174A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101621587B (en) * | 2008-06-30 | 2012-08-08 | 成都市华为赛门铁克科技有限公司 | Method, device and system for network monitoring |
| US8416695B2 (en) | 2008-06-30 | 2013-04-09 | Huawei Technologies Co., Ltd. | Method, device and system for network interception |
| CN101582897A (en) * | 2009-06-02 | 2009-11-18 | 中兴通讯股份有限公司 | Deep packet inspection method and device |
| CN102932817A (en) * | 2012-10-10 | 2013-02-13 | 福建星网锐捷网络有限公司 | Method and device for recognizing voice stream and wireless access equipment |
| CN102932817B (en) * | 2012-10-10 | 2015-02-25 | 福建星网锐捷网络有限公司 | Method and device for recognizing voice stream and wireless access equipment |
| CN103400579A (en) * | 2013-08-04 | 2013-11-20 | 徐华 | Voice recognition system and construction method |
| CN103400579B (en) * | 2013-08-04 | 2015-11-18 | 徐华 | A kind of speech recognition system and construction method |
| CN108206788A (en) * | 2016-12-16 | 2018-06-26 | 中国移动通信有限公司研究院 | The business recognition method and relevant device of a kind of flow |
| CN108206788B (en) * | 2016-12-16 | 2021-07-06 | 中国移动通信有限公司研究院 | Traffic service identification method and related equipment |
| CN110519169A (en) * | 2019-08-30 | 2019-11-29 | 成都安恒信息技术有限公司 | A kind of method of application layer multiplexed network header |
| CN110519169B (en) * | 2019-08-30 | 2021-11-26 | 成都安恒信息技术有限公司 | Method for multiplexing network message header by application layer |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101072174A (en) | Tencent voice identifying method based on pay load deep detection and session correlating technology | |
| CN106550155B (en) | Swindle sample is carried out to suspicious number and screens the method and system sorted out and intercepted | |
| US8416695B2 (en) | Method, device and system for network interception | |
| CN103023670B (en) | Message traffic kind identification method and device based on DPI | |
| US20130294449A1 (en) | Efficient application recognition in network traffic | |
| CN101321097A (en) | Tencent network living broadcast business recognition method based on payload depth detection | |
| CN112511696A (en) | System and method for identifying bad content of call center AI engine | |
| CN102571946A (en) | Realization method of protocol identification and control system based on P2P (peer-to-peer network) | |
| CN110620849B (en) | Centralized sorting method and system for IMS telephone terminal call records | |
| CN106789728A (en) | A kind of voip traffic real-time identification method based on NetFPGA | |
| CN101631174B (en) | Network telephone real-time identification and filtering method based on session initiation protocol | |
| CN106572103A (en) | Hidden port detection method based on SDN network architecture | |
| US8725901B2 (en) | Analysis tool for intra-node application messaging | |
| CN101420336A (en) | Method for recognizing network telephone flow quantity in network and system thereof | |
| CN109067782A (en) | IMS network session abnormal interrupt attack detecting device and method | |
| CN107070741B (en) | A kind of voip network topology detection method based on the analysis of gateway space time correlation | |
| CN107846520A (en) | single-pass detection method and device | |
| Matousek et al. | Fast RTP detection and codecs classification in internet traffic | |
| CN105357166B (en) | A kind of method of next generation firewall system and its detection messages | |
| CN108234498B (en) | Communication call tracing method and system | |
| CN110404267A (en) | A kind of plug-in detection method of game based on HTTP flow HOST field feature | |
| CN108055287A (en) | Voice gateways system and method based on Session Initiation Protocol | |
| KR101287588B1 (en) | Security System of the SIP base VoIP service | |
| US20080031231A1 (en) | System and method for handling media streams | |
| Anwar et al. | Data analysis and summarization to detect illegal VoIP traffic with call detail records |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20071114 |