CN105323751A - Mobile phone short message monitoring device - Google Patents
Mobile phone short message monitoring device Download PDFInfo
- Publication number
- CN105323751A CN105323751A CN201410379283.5A CN201410379283A CN105323751A CN 105323751 A CN105323751 A CN 105323751A CN 201410379283 A CN201410379283 A CN 201410379283A CN 105323751 A CN105323751 A CN 105323751A
- Authority
- CN
- China
- Prior art keywords
- short message
- module
- note
- sms
- monitoring device
- 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.)
- Withdrawn
Links
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a mobile phone short message monitoring device. The mobile phone short message monitoring device is characterized in that the short message monitoring device is connected with a short message service center through a tandem gateway, the internal connection relationship of the mobile phone short message monitoring device is that the monitoring device is connected with a call information number, frequency sending success rate counting module, a short message consistency judgment module, a user confidence grade module, a comprehensive assessment short message category module, a black list-white list management module and a home location register module; a telephone bill acquiring module is connected with a Bayes short message learning module, a keyword dictionary module, a spell fussy keyword matching module and the comprehensive assessment short message category module; a mobile phone short message monitoring device hardware platform is an IBMX39508872-7RC multi-core environment. Through the adoption of the device disclosed by the invention, an interference message in the spam short message can be effectively eliminated, the interception rate of the spam short message and the judgment right rate are improved, the influence of the mobile phone short message monitoring device to the processing ability of the short message service center is reduced.
Description
Technical field
The present invention relates to a kind of SMS supervising device.
Background technology
2000 to 2006, China's short message service amount was respectively: 1,000,000,000,18,900,000,000,90,000,000,000,1,371 hundred million, 2,177 hundred million, 4,296 hundred million, within 2006, daily sent 1,200,000,000.In so numerous notes, more than 30% is refuse messages.So-called refuse messages refer to run counter to recipient's wish, recipient cannot reject, hide the illegal or advertisement SMS of addresser's real information.Refuse messages not only occupies limited Internet resources, causes network congestion, has ruined the corporate image of operator in the public, and has had a strong impact on normal work and the life of user.Even more serious, note has become the carrier that some unprincipled fellows implement crime, utilizes note to carry out swindling, the event of pornographic transaction and reaction publicity happens occasionally, cause social public trust to decline, the realization of the target that is unfavorable for building a harmonious society.
Note application is abroad limited.Reason is that external post and telecommunication tariff is cheap, and user's ability to pay is strong, and user does not widely use the custom of note.For Japan, operator never carries out short message service; Short message service cradle Europe 1991 Jian open up this business, up to now, scale is still very limited.Therefore, the research of developed country in short message monitoring both insufficient not deeply yet, wherein affect larger use more be the MlilifeAnti-SPAM product that Alcatel-Lucent releases, be called for short ASA.ASA adopts the rule engine technique of Bell Laboratory patent, has self study and dynamic conditioning rule functional.ASA is necessarily applied at home, but effect is barely satisfactory.Because SAS is based on the complete matching technique of keyword, refuse messages sender at short encloure interfere information, such as, inserts insignificant character or nearly sound word substitutes, and just can evade the interception of ASA; ASA is according to the interception of short message sending quantity (100/h) rule, and do not consider the behavioural characteristics such as short message content consistency, refuse messages sender only needs to do the interception that simple technique process just can evade ASA.
The domestic research to SMS monitoring center is stuck on " SMS service center and the SMS supervising device interface specification " of promulgation in 2003, there is no substantive behave, to trace it to its cause following 3 points: profit-seeking is the instinct of the operator as enterprise, note is the most important point of economic increase of operator at present, filter note and certainly will affect profit, therefore operator holds a negative attitude for a long time; Based on the first point, the passive attitude of operator has a strong impact on input and the enthusiasm of SMS supervising device developer; The property grasped of relevant specification of country, standard is inadequate, and the Jian that also have impact on SMS supervising device sends out.For the patent of invention of applying for State Intellectual Property Office, applications only ten remainders so far, authorize three.Along with the lasting rising of regulation refuse messages society cry, operator recognizes that enterprise should social responsibility, and since in October, 2006, operator accelerates research to mobile phone text message monitoring device aspect and input.Now with regard to mobile phone text message monitoring device field application RELATED APPLICATIONS involved by technology and strategy be summarized as follows:
A kind of SMS supervising device (application number 200510117449.7), proposes user sets up Keyword List disposal of refuse note method in carrier network side.
(application number 200510117448.2 > proposes user sets up list of numbers disposal of refuse note method in carrier network side to a kind of SMS supervising device.
(application number 200610020383.4 >, propose user and arrange personalized filtering short message rule, operator arranges general filtering short message rule to the method for short message firewall system and setting-up short message fire compartment wall.
The linear dimensions fuzzy matching algorithm (application number 200410061270.X) of short message content, proposes in keyword the technology introducing asterisk wildcard, solves the problem that keyword match technique lost efficacy in the refuse messages of additional interference information.
A kind of method (ZL02137335.3) of mobile phone SMS filtering, proposes user side and arranges keyword filtrating rubbish short message.
A kind of method (ZL200410042668.9) of handset to filter short message proposes user side and arranges telephone number filtrating rubbish short message.
Short message method of realizing fireproof wall (ZL20041008826.9) proposes user side and arranges rubbish number filtering refuse messages.
Above-mentioned achievement in research and benefit our pursuits certain reference, reference value, its weak point can be summarized as follows: the validity introducing asterisk wildcard fuzzy matching technology in keyword has much room for improvement; Do not notice that refuse messages has obvious behavioural characteristic, especially short message content this key element consistent; SMS supervising device real time filtering affects the problem of SMS service center processing ability, lacks corresponding technical measures; Compared with legal note erroneous judgement is judged by accident with refuse messages, the former brings larger loss to user, also causes the trouble more crossed to operator, lacks to this treatment technology and the strategy of taking into account benefits of different parties accordingly; The Auto of keyword dictionary.
Summary of the invention
For solving above-mentioned existing shortcoming, main purpose of the present invention is the SMS supervising device providing a kind of practicality, the interfere information that can effectively remove rubbish in note, improve the interception rate of refuse messages and differentiate accuracy, reducing SMS supervising device to the impact of SMS service center processing ability.
For reaching above-described object, following technical scheme taked by a kind of SMS supervising device of the present invention:
A kind of SMS supervising device, it is characterized in that, SMS supervising device is connected with SMS service center by GNS, the inside annexation of SMS supervising device is: bill record collection module with rise exhale information number, frequency sends successful rate statistics module, short message consistency discrimination module, users to trust grade module, Comprehensive Assessment short message class Modules, black and white lists administration module, attaching position register module connects, ticket Bian collection module and Bayes's short message study module, keyword lexicon module, phonetic obfuscation Keywords matching module, Comprehensive Assessment short message class Modules connects, SMS supervising device hardware platform is IBMX39508872-7RC multi-core environment.
The method of real time filtering mechanism fusion ticket analysis mechanisms is adopted jointly to complete the filtration of refuse messages, ticket Bian collection module reads in data from the accounting module at SMS service center, the behavioural characteristic such as information number, frequency, transmission success rate of statistics short message, short message content consistency then adopts MD5HASH algorithm, dynamic conditioning users to trust grade; Bill record collection module, simultaneously for Bayesian learning module, upgrades the keyword dictionary of short message real time filtering; Phonetic obfuscation Keywords matching formwork module, to the short message real time filtering of authentication request, draws short message rubbish angle value; Consider short message spam degree and users to trust grade factor, provide the discriminating response of short message classification, the black and white lists simultaneously generated feeds back to SMS service center; Communication protocol between SMS supervising device and SMS service center is SMPPV3.X, short message monitoring device hardware platform is IBMX39508872-7RC multi-core environment, runs parallel multithreading real time filtering program under Linux symmetric multiprocessor framework.
Described phonetic obfuscation Keywords matching template method: when after system acceptance to note, system is mated adopting phonetic obfuscation Keywords matching mould plate technique after the obfuscation of note phonetic, obtains the spam degree that note is corresponding.Phonetic obfuscation Keywords matching mould plate technique effectively can eliminate various artificial interfere information in refuse messages; The method of described ticket analysis mechanisms: the behavioural characteristic being conceived to refuse messages, has comprised and has exhaled information number, frequency, transmission success rate, short message content consistency etc., the reliability rating of dynamic conditioning user; In conjunction with the note spam degree that real time filtering is determined, the final classification of this note of Comprehensive Assessment.Short message content consistency discrimination adopts the MD5HASH algorithm in cryptography: carry out to the every bar note entering Surveillance center the integer that MD5 algorithm produces 128, and compare with all HASH numbers that the note that Surveillance center in a period of time receives produces, if there is identical integer, then think that two short message contents are identical, 1 is added to the quantity forwarded of this note, computing function F(X, L, N), X is the reliability rating of user, L is the spam degree of note, N is the quantity of short message sending, calculate and whether can continue to send this note, and process accordingly according to result of calculation.If there is no identical integer, then put 1 by short message sending quantity, and this integer is inserted in HASH queue; Note is divided three classes by described SMS supervising device: legal note, suspicious short message and refuse messages; Legal note is adopted to let pass, intercepting rubbish short message, the strategy that suspicious short message is let pass after adding operator's signal language.The keyword dictionary of SMS supervising device is generated note sample off-line learning automatically by Bayesian Method.
Adopt the present invention of as above technical scheme, there is following beneficial effect:
1, Chinese keyword adopts phonetic obfuscation Keywords matching mould plate technique, the interfere information that can effectively remove rubbish in note.
2, according to the behavioural characteristic of refuse messages, the reliability rating of dynamic conditioning user, in conjunction with based on phonetic obfuscation Keywords matching template real time filtering technology determined note rubbish angle value, the classification of Comprehensive Assessment note, improves the interception rate of refuse messages and differentiates accuracy;
3, on multinuclear hardware platform, run parallel real time filtering program, reduce SMS supervising device to the impact of SMS service center processing ability.
4, SMS classified and process New Policy, note is divided into legal note, suspicious short message and refuse messages, to let pass legal note, catching rubbish note, and the strategy that suspicious short message is let pass after adopting additional carrier signal language, in China's existing laws and regulations framework, take into account the common interest of user and operator.
5, Bian Bayesian Method carries out off-line learning to note sample, automatically generates refuse messages keyword dictionary.
Accompanying drawing explanation
Fig. 1 is the structured flowchart of short message service center (SMS service center) and mobile phone short message supervising device (SMS supervising device).
Fig. 2 is structured flowchart of the present invention (eliminating GNS).
Fig. 3 is phonetic obfuscation Keywords matching template algorithm flow chart.
Fig. 4 (a) is content of short message consistency discrimination MD5 algorithm major cycle schematic diagram.
Fig. 4 (b) is the once-through operation schematic diagram during md5 encryption algorithm is often taken turns.
Fig. 5 is that Bayes generates short message keyword dictionary flow chart automatically.
Embodiment
In order to further illustrate the present invention, be described further below in conjunction with accompanying drawing:
Note real time filtering has multiple method in theory, but the most widely used method remained based on Keywords matching.Adopt for refuse messages and artificially increase interfere information, the monitoring of interference SMS monitoring center, propose phonetic obfuscation Keywords matching template method, reach and effectively eliminate interfere information, the object of filtrating rubbish short message.
Be different from the fact of the remarkable behavioural characteristic of legal note in view of refuse messages, as high transmission frequency and traffic volume, the features such as lower transmission success rate, short message content are consistent, are differentiated in ticket analytical method, and the reliability rating of dynamic conditioning user.In conjunction with the note spam degree that real time filtering is determined, the final classification of Comprehensive Assessment note.Short message content consistency discrimination adopts the MD5HASH algorithm in cryptography.
Usually the fact of SMS service center processing ability decline left and right can be caused in view of SMS monitoring center puts into operation; SMS monitoring center hardware platform selects multi-core environment; under symmetric multiprocessor framework, run parallel multithreading real time filtering program, reduce SMS monitoring center to SMS service center processing capacity.
In view of the fact that legal note erroneous judgement is larger far beyond the negative effect of refuse messages erroneous judgement, consider the constraint of China's relevant laws and regulations to operator, note is divided into legal note, suspicious short message and refuse messages.Legal note is let pass, intercepting rubbish short message, and suspicious short message is the Bian strategy of letting pass after additional operator's signal language then.With regard to user, important information can be avoided to be blocked the loss caused; With regard to operator, interception note is free by regulation, too much interception certainly will affect income, even if suspicious short message refuse messages simultaneously, because addition of prompting, from law angle, operator does duty, moreover whether rubbish is different with people for a same note, and therefore above-mentioned strategy has taken into account the interests of user and operator.
The achievement of filtering spam mail domain used for reference by keyword dictionary, adopts Bayesian Method automatically to generate note sample off-line learning.
As shown in Figure 1, after the short message that each SMS service receive centre domestic consumer sends or entity is submitted to, all send an authentication request to GNS, GNS forwards authentication request to SMS monitoring center.SMS monitoring center judges short message content and behavioural characteristic, returns to the response of SMS service Center Authentication through GNS.If short message is legal, SMS monitoring center returns authentication successful message, and this note is handed down to recipient by SMS service center; If short message is suspicious, SMS monitoring center returns authentication and accepts message, and recipient is handed down to after this note is added operator's signal language in SMS service center; If short message is illegal, return failed authentication message, this note abandons by SMS service center.According to the Ministry of Information Industry " SMS service center and SMS monitoring center interface specification ", information interaction between SMS service center and SMS monitoring center adopts the SMPP agreement of standard, SMS service center will be risen by DELIVER_SM and be exhaled MO (MobileOriginal) information to submit to SMS monitoring center, SMS monitoring center extracts short_message_test field contents in DELIVER_SM, according to returning one, SMS service center SELIVER_SM_RESP authentication response message after established rule process.If the time of SMS service center waits response is greater than 5s, then SMS service central record daily record, direct transmitting short message.
As shown in Figure 2, the inside annexation of SMS monitoring center is: ticket Bian collection module exhales that information number, frequency send successful rate statistics module, short message consistency discrimination module, users to trust grade module, Comprehensive Assessment short message class Modules, black and white lists administration module, HLR attaching position register module connect with rising, and ticket Bian collection module connects with Bayes's short message study module, keyword lexicon module, phonetic obfuscation Keywords matching module, Comprehensive Assessment short message class Modules.
Short message monitoring method adopts the method for real time filtering mechanism fusion ticket analysis mechanisms jointly to complete the filtration of refuse messages, bill record collection module reads in data from the accounting module at SMS service center, the behavioural characteristic such as information number, frequency, transmission success rate of statistics short message, short message content consistency then adopts MD5HASH algorithm, dynamic conditioning users to trust grade; Ticket Bian collection module, simultaneously for Bayesian learning module, upgrades the keyword dictionary of short message real time filtering; Phonetic obfuscation Keywords matching formwork module, to the short message real time filtering of authentication request, draws short message rubbish angle value; Consider short message spam degree and users to trust grade factor, provide the discriminating response of short message classification, the black and white lists simultaneously generated feeds back to SMS service center; Communication protocol between SMS monitoring center and SMS service center is SMPPV3.X, and SMS monitoring center hardware platform is IBMX39508872-RC multi-core environment, runs parallel multithreading real time filtering program under symmetric multiprocessor framework.
As shown in Figure 3, phonetic obfuscation Keywords matching template real time filtering technology is discussed in conjunction with example.Carried out the Bayesian learning of rubbish keyword by 30,000 notes, obtain refuse messages keyword " Falun Gong ", be " FaLunGong " after carrying out alphabetizing, obtain the spam degree that " FaLunGong " is corresponding, be set to M.After the illegal note such as receive " Falun Gong personnel burn oneself to death in the Tian'anmen Square ... " when system, system, to mating after note alphabetizing, matches " FaLunGong " and obtains corresponding spam degree.Lawless person gets following several distortion to the main Bian of refuse messages:
Increase punctuation mark.As: " method, wheel merit personnel burn oneself to death in the Tian'anmen Square ... " etc., first system removes punctuation mark before coupling, and therefore above note can match " FaLunGong " equally.
Increase decollator.As: " Falun Gong personnel burn oneself to death in the Tian'anmen Square ... ", system is after removal punctuation mark, and also can remove decollator, therefore above note can match " FaLunGong " equally
Increase irrelevant word.As: " method wheel merit personnel burn oneself to death in the Tian'anmen Square ... " short message keyword template matching technique, its core concept is that common keyword is decomposed into single character.In view of the length of a note, "=75 Chinese characters, the irrelevant word interference that refuse messages inserts in keyword is limited, and too much irrelevant word can make the readability of refuse messages degenerate simultaneously, losing deceitful effect.Keyword " Falun Gong " is considered as three words by keyword template matching technique: " method " " wheel " " merit "; Suppose that the length of template is length keywords X2, for keyword " Falun Gong ", template length is 6.During real time filtering, scanning note in full, " method " word mates with the lead-in of refuse messages keyword " Falun Gong ", from note, then intercept template length < 6 character > contents mate with keyword " Falun Gong ", therefore above note can match " FaLunGong " equally.The deformable bodys such as such refuse messages " method wheel merit " also cannot evade the real time filtering interception of SMS monitoring center.Template length adjusts at any time depending on operation result.
Replace by homophone, as: " sending out wheel merit personnel to burn oneself to death in the Tian'anmen Square ... ", because system phonetic mates, therefore " send out opinion merit " and can match " FaLunGong " too when being converted into phonetic.
Replace with nearly sound word, as " floral whorl merit personnel burn oneself to death in the Tian'anmen Square ... ", system is when carrying out Chinese character to the conversion of phonetic, similar pinyin can be carried out Fuzzy Processing, as " hua & " arrives " fa ", " run " arrives " lun ", and " neng " arrives " nen " etc.Therefore " floral whorl merit " also can mate " FaLunGong " after carrying out the fuzzy conversion of phonetic.
In a word, phonetic obfuscation Keywords matching mould plate technique effectively can eliminate various artificial interfere information in refuse messages.
As shown in Fig. 4 (a) (b), in view of refuse messages has obvious malicious act feature, the feature such as consistent in high transmission frequency and traffic volume, lower transmission success rate, short message content, the therefore behavioural characteristic of analytic induction refuse messages, the reliability rating of dynamic conditioning user is feasible.The effectiveness of SMS monitoring center is can further improve with real time filtering combine with technique.The statistics of transmission frequency, traffic volume, transmission success rate belongs to known technology, but short message content consistency discrimination is infeasible by traditional full text scanning comparison method one by one, scanning comparison method one by one will expend the ample resources of SMS monitoring center in full, cannot meet requirement requirement of real-time being bordering on to harsh SMS monitoring center.
The full name of MD5 is Message-DigestAlgorithm5 (md5-challenge), and its typical apply produces informative abstract to a segment information.Whole file is used as a Long Binary information, by its irreversible character string mapping algorithm, creates the MD5 informative abstract that this is unique.If propagated afterwards in the process of this file, no matter the content of file there occurs any type of change (comprising the error of transmission etc. that in artificial amendment or process of transmitting, circuit instability causes) as long as will find that informative abstract is not identical when recalculating MD5 to this file, can determine the just incorrect file obtained thus.In simple terms, the effect of MD5 allows note be ' compressed ' into a kind of secret form before filtration, namely the byte serial of a random length is transformed into the big integer of a fixed length.
Consistency discrimination message M is divided into from M
1, M
2... M
kk subgroup altogether.The input of one-way hash function is message grouping M
iwith last message grouping M
i-1the hashed value h of hashed value
i-1, export hashed value h
i, export hashed value hi=f (M
i, h
i-1).This hashed value h
iwith the message grouping M of next round
i+1together, the next round as one-way hash function inputs, and so repeatedly, the output of last group is exactly the hashed value of whole message to be encrypted.
Md5 encryption algorithm is a kind of one-way hash function that chi RonRivest designs, it is grouped into processed in units input expressly with 512, each 512 groupings are divided into again 16 32 seat groupings, and the output of algorithm is 32 groupings of 4 cascades, forms the hashed value of 128.The initialization of MD5 algorithm, major cycle and afterbody process three steps are as follows:
1. initialization.Algorithm initialization is divided into original plaintext initialization and link variable initialization two parts.Expressly initialization first by length be L plaintext P to be encrypted after cascade byte 0X8O, then several 0X00 of cascade are as filling, make its length for (512 ﹡ K-64) position (K is positive integer).Then cascade 64 is in the length of the plaintext of byte again, and now length of the plaintext is filled to be the integral multiple of 512, guarantees that isolog can not be not identical after filling simultaneously.Four 32 link variables are initialized as: A=0x01234567, B=0x89ABCDEF, C=0xFEDCBA98, D=0x76543210.These four link variables are most important for whole algorithm, and its value is directly connected to writing of corollary equipment software, these four link variables must be write correctly.
2. algorithm major cycle.Algorithm major cycle is as shown in Fig. 4 (a).Major cycle number of times is the business of the clear data figure place through filling divided by 512 gained.Link variable A, B, C, D are assigned to respectively a, b, c, d tetra-variablees.Each major cycle has four-wheel, and often wheel carries out 16 operations, and each operation first carries out a nonlinear function computing to three in a, b, c, d, then acquired results is added successively the 4th variable, M
j(subgroup in text) and a C
i(constant) by constant S indefinite for acquired results ring shift left one, and then one of to add in a, b, c or d, finally one of to replace in a, b, c or d by this result.Each operating process is as shown in Fig. 4 (b).
3. afterbody process.A, B, C, D are added a, b, c, d after all completing by major cycle respectively each time.Then continue executing arithmetic with next grouping, last output hashed value is the cascade of A, B, C and D.Definition nonlinear function FF(a, b, c, d, M
j, S, C
i):
a=b+[F(b,c,d)+a+M
j+C
i]＜＜＜S
Wherein, S is the figure place of ring shift left, is a constant; Constant C
ibe as 2 in i-th step
32﹡ | sin(i) | integer part (i unit is radian).Four-wheel is altogether 16 × 4=64 time, and nonlinear function is each time not identical.
Determine whether that the note method repeating to send is as follows based on MD5HASH algorithm: the integer that MD5 algorithm produces 128 is carried out to the every bar note entering Surveillance center, and compare with all HASH numbers that the note that Surveillance center in a period of time receives produces, if there is identical integer, then think that two short message contents are identical, 1 is added to the quantity forwarded of this note, computing function F(X, L, N), X is the reliability rating of user, L is the spam degree of note, N is the quantity of short message sending, calculate and whether can continue to send this note, and process accordingly according to result of calculation.If there is no identical integer, then send note, and short message sending quantity is put 1, and this integer is inserted in HASH queue.
As shown in Figure 5, at SMS monitoring center idle period, Bayes's short message study module automatic/manual screening short message sample carries out Bayesian learning to the sample filtered out; The rubbish short message keyword generated carries out short message real time filtering for phonetic obfuscation Keywords matching module.
Existing research and test show, SMS monitoring center puts into operation and usually reduces the disposal ability at SMS service center 30%, because short message real time filtering---the software and hardware resources expense of phonetic obfuscation Keywords matching template algorithm is comparatively large, and this is the thorny problem that operator faces.The technical measures solving this difficult problem adopt multinuclear hardware platform, runs parallel real time filtering program.The difficult point of multinuclear application is at software, and short message real time filtering has natural concurrency fortunately, and therefore parallel filter difficulty reduces greatly, is exemplified below.
The IBMX39508872-7RC server that hardware platform adopts contains 2 IntelXeon7020 dual core processors, meets the symmetric multiprocessor framework of system.The difference of polycaryon processor and multiprocessor is the channel whether adopting Front Side Bus as data sharing.The parallel processing of short message should adopt the pattern of one process multiple threads, because contrast thread and process, establishment and the handover overhead of thread mode that is less than process, inter-thread communication is many and simple also more efficient, multithreading has the base library support that so many as to make the ox carrying them perspire and to fill a house to the rafters, the program of multithreading is easier to understand than the program of multi-process and revises.For the operating system supporting SMP framework, the scheme that each thread is assigned to different IPs carrys out Automatic dispatching by system kernel, and scheduling scheme is that platform is correlated with for kernel, and application programs is irrelevant.
Background application needs to comprise (pthread.h) header file to create multithreading.The function prototype creating new thread is externintpthread_create_P ((pthread_t ﹡ _ thread, _ constpthread_attr_t ﹡ _ attr, void ﹡ (﹡ _ start_routine) (void ﹡), void ﹡ _ arg)); First parameter is point to the pointer of thread identifier, and second parameter is used for arranging thread attribute, and the 3rd parameter is the initial address that thread runs function, and last parameter runs the parameter of function.After sub-thread completes, main
Thread externintpthread_join_P ((pthread_t_th, vido ﹡ ﹡ _ thread, _ return)) or externintpthread_exit__P ((vido ﹡ _ _ retval)) _ _ attribute__ ((_ _ noreyurn__)) wait for its end.
After creating sub-thread, thread can be set with pthread_attr_setscope () and whether be tied to the real-time process light process done and does not need to dispatch; Pthread_attr_setdetachstate (pthread_attr_t ﹡ attr, intdetachstate) can be used to set the mode of Thread Termination self.
Claims (3)
1. a SMS supervising device, it is characterized in that, short message monitoring device is connected with SMS service center by GNS, the inside annexation of SMS supervising device is: with rise exhale information number, frequency sends successful rate statistics module, short message consistency discrimination module, users to trust grade module, Comprehensive Assessment short message class Modules, black and white lists administration module, attaching position register module connects, bill record collection module and Bayes's short message study module, keyword lexicon module, phonetic obfuscation Keywords matching module, Comprehensive Assessment short message class Modules connects, described SMS supervising device hardware platform is IBMX39508872-7RC multi-core environment.
2. a kind of SMS supervising device according to claim 1, it is characterized in that, the method of real time filtering mechanism fusion ticket analysis mechanisms is adopted jointly to complete the filtration of refuse messages, ticket Bian collection module reads in data from the accounting module at SMS service center, the behavioural characteristic such as information number, frequency, transmission success rate of statistics short message, short message content consistency then adopts MD5HASH algorithm, dynamic conditioning users to trust grade; Bill record collection module, simultaneously for Bayesian learning module, upgrades the keyword dictionary of short message real time filtering; Phonetic obfuscation Keywords matching formwork module, to the short message real time filtering of authentication request, draws short message rubbish angle value; Consider short message spam degree and users to trust grade factor, provide the discriminating response of short message classification, the black and white lists simultaneously generated feeds back to SMS service center; Communication protocol between SMS supervising device and SMS service center is SMPPV3.X, short message monitoring device hardware platform is IBMX39508872-7RC multi-core environment, runs parallel multithreading real time filtering program under Linux symmetric multiprocessor framework.
3. a kind of SMS supervising device according to claim 2, it is characterized in that, described phonetic obfuscation Keywords matching template method: when after system acceptance to note, system is mated adopting phonetic obfuscation Keywords matching mould plate technique after the obfuscation of note phonetic, obtain the spam degree that note is corresponding, phonetic obfuscation Keywords matching mould plate technique effectively can eliminate various artificial interfere information in refuse messages, the method of described ticket analysis mechanisms: the behavioural characteristic being conceived to refuse messages, has comprised and has exhaled information number, frequency, transmission success rate, short message content consistency etc., the reliability rating of dynamic conditioning user, in conjunction with the note spam degree that real time filtering is determined, the final classification of this note of Comprehensive Assessment, short message content consistency discrimination adopts the MD5HASH algorithm in cryptography: carry out to the every bar note entering Surveillance center the integer that MD5 algorithm produces 128, and compare with all HASH numbers that the note that Surveillance center in a period of time receives produces, if there is identical integer, then think that two short message contents are identical, 1 is added to the quantity forwarded of this note, computing function F(X, L, N), X is the reliability rating of user, L is the spam degree of note, N is the quantity of short message sending, calculate and whether can continue to send this note, and process accordingly according to result of calculation, if there is no identical integer, then short message sending quantity is put 1, and this integer is inserted in HASH queue, note is divided three classes by described SMS supervising device: legal note, suspicious short message and refuse messages, adopt legal note to let pass, intercepting rubbish short message, the strategy that suspicious short message is let pass after adding operator's signal language, the keyword dictionary of SMS supervising device is generated note sample off-line learning automatically by Bayesian Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410379283.5A CN105323751A (en) | 2014-08-04 | 2014-08-04 | Mobile phone short message monitoring device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410379283.5A CN105323751A (en) | 2014-08-04 | 2014-08-04 | Mobile phone short message monitoring device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105323751A true CN105323751A (en) | 2016-02-10 |
Family
ID=55250166
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410379283.5A Withdrawn CN105323751A (en) | 2014-08-04 | 2014-08-04 | Mobile phone short message monitoring device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105323751A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106445815A (en) * | 2016-09-06 | 2017-02-22 | 合网络技术（北京）有限公司 | Automated testing method and device |
| CN109885645A (en) * | 2018-11-06 | 2019-06-14 | 北京奇虎科技有限公司 | A kind of method, apparatus, storage medium and equipment that refuse messages identify |
| CN109933775A (en) * | 2017-12-15 | 2019-06-25 | 腾讯科技（深圳）有限公司 | UGC content processing method and device |
| CN110996272A (en) * | 2019-11-05 | 2020-04-10 | 武汉普斯顿网络科技有限公司 | Game service verification method, device, equipment and medium |
| CN113316153A (en) * | 2020-04-02 | 2021-08-27 | 阿里巴巴集团控股有限公司 | Short message inspection method, device and system |
-
2014
- 2014-08-04 CN CN201410379283.5A patent/CN105323751A/en not_active Withdrawn
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106445815A (en) * | 2016-09-06 | 2017-02-22 | 合网络技术（北京）有限公司 | Automated testing method and device |
| CN106445815B (en) * | 2016-09-06 | 2019-04-23 | 优酷网络技术（北京）有限公司 | A kind of automated testing method and device |
| CN109933775A (en) * | 2017-12-15 | 2019-06-25 | 腾讯科技（深圳）有限公司 | UGC content processing method and device |
| CN109885645A (en) * | 2018-11-06 | 2019-06-14 | 北京奇虎科技有限公司 | A kind of method, apparatus, storage medium and equipment that refuse messages identify |
| CN110996272A (en) * | 2019-11-05 | 2020-04-10 | 武汉普斯顿网络科技有限公司 | Game service verification method, device, equipment and medium |
| CN113316153A (en) * | 2020-04-02 | 2021-08-27 | 阿里巴巴集团控股有限公司 | Short message inspection method, device and system |
| CN113316153B (en) * | 2020-04-02 | 2024-03-26 | 阿里巴巴集团控股有限公司 | Short message inspection method, device and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100576940C (en) | Short message monitoring center and method for supervising | |
| US11050690B2 (en) | Method for providing recording and verification service for data received and transmitted by messenger service, and server using method | |
| CN105323751A (en) | Mobile phone short message monitoring device | |
| CN103853841A (en) | Method for analyzing abnormal behavior of user in social networking site | |
| CN102231873A (en) | Method and system for monitoring garbage message and monitor processing apparatus | |
| US11722503B2 (en) | Responsive privacy-preserving system for detecting email threats | |
| KR20170024777A (en) | Apparatus and method for detecting smishing message | |
| TW202121329A (en) | Method and apparatus for evaluating risk of leakage of private data | |
| CN101909261A (en) | Method and system for monitoring spam | |
| CN108388802A (en) | A kind of alarm method and warning system of script injection attacks | |
| Liu et al. | Cyberthreat-intelligence information sharing: Enhancing collaborative security | |
| CN201066901Y (en) | SMS monitoring center | |
| CN102271331B (en) | Method and system for detecting reliability of service provider (SP) site | |
| Sun et al. | Design and Application of an AI‐Based Text Content Moderation System | |
| CN112699088A (en) | Method, system and medium for sharing fraud-related data | |
| CN104301804B (en) | One kind realizes internet electronic government affairs information exchange method via DTV | |
| Ryan | Weibo diplomacy and censorship in China | |
| CN107341375A (en) | A kind of method and system for the attacker that traced to the source based on Web page picture secret mark | |
| CN114363839B (en) | Fraud data early warning method, device, equipment and storage medium | |
| KR102531661B1 (en) | System for providing phishing warning service using contents based blocking | |
| CN112347502A (en) | Information processing method and system | |
| Nguyen et al. | Context-driven policies enforcement for edge-based iot data sharing-as-a-service | |
| CN116862679B (en) | Block chain-based data processing method, device, equipment and readable storage medium | |
| KR20200040038A (en) | Block chain based civil affair document nature authenticity system | |
| CN112910763B (en) | Method, device, equipment and medium for providing real-time data interface service |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C04 | Withdrawal of patent application after publication (patent law 2001) | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20160210 |