CN109725836A - User's context compression method and device - Google Patents
User's context compression method and device Download PDFInfo
- Publication number
- CN109725836A CN109725836A CN201711030139.0A CN201711030139A CN109725836A CN 109725836 A CN109725836 A CN 109725836A CN 201711030139 A CN201711030139 A CN 201711030139A CN 109725836 A CN109725836 A CN 109725836A
- Authority
- CN
- China
- Prior art keywords
- node
- content
- context data
- user
- multiplicity
- 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.)
- Granted
Links
Abstract
The present invention proposes user context data compression method and device.Method includes: that LTE base station finds that any context data node of any LTE user produces content, the memory block then is written into the content, and in the context data list of the user of self maintained, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded;When the LTE base station free time, each context data node of each user of LTE base station poll counts the number of repetition of every kind of content of each context data node;At the end of each measurement period, calculate the multiplicity of every kind of content of each node, the highest preset number kind content of multiplicity is put into General Configuration Template, and, discharge the memory block where the node content, meanwhile the corresponding pointer of mark of the node in the context data list of user being replaced with to the pointer of the node content in General Configuration Template.The present invention improves the compression ratio of user context data.
Description
Technical field
The present invention relates to technical field of data compression more particularly to user context data compression methods and device.
Background technique
As LTE (Long Term Evolution, long term evolution) technology is widely popularized industrial application, power train
Unite wireless private network also complete upgrading to LTE standard.More traditional 2G system, LTE system transmission bandwidth, propagation delay time,
Line number of users etc. is improved.Simultaneously because (user density is high, data volume is small, adjusts for the business characteristic of LTE private electric power
Degree interval length), it is online simultaneously that single device needs support tens of thousands of users.So many user management data (the i.e. user of maintenance management
Context data), great challenge is proposed to base station stored resource.Need to propose a kind of suitable embedded system for base station
Context compression method reduces equipment cost to improve base station system storage efficiency.
Traditional LTE base station equipment, for manage user's context storage, by each user's regular length it is direct including
Deposit expansion.Memory overhead is linear with number of users, and as number of users is promoted, memory overhead is huge.
Because of the complexity of LTE protocol stack, generally requires for the context data of a user and occupy 5KByte storage and open
Pin.Under public network scene, 1200 online user of Dan little Qu, storage overhead 6MByte.But it under electric power wireless private network environment, presses
230Mhz frequency range cell coverage area calculates, and Dan little Qu online user is up to the total user of 20,000 or more, Dan Zhanliu cell up to 100,000
More than.Only context storage overhead can not be deployed in general only 512MByte memory in 500MByte or more at this time
In baseband board card.
To ensure the protocol stack deployment for supporting large user's characteristic, it is necessary to be compressed using compress technique to context.But
LTE protocol stack, especially TD-LTE (Time-Division LTE, time-division LTE) system, a processing week extremely sensitive to the time
Phase needs are strictly limited in a TTI (Transmission Time Interval, Transmission Time Interval) (usually 1ms)
It is interior, it is a Hard Real Time Systems.Common Real-time compression algorithm can not ensure such hard real time requirement.
Summary of the invention
The present invention provides user context data compression method and device, to improve LTE user context data compression ratio.
The technical scheme of the present invention is realized as follows:
A kind of user context data compression method, this method comprises:
Long term evolution LTE base station finds that any context data node of any LTE user produces content, then according to pre-
The memory block is written in the content by one memory block of length application of the node first defined, and the user's of self maintained
In context data list, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded;
When the LTE base station free time, each context data node of each user of LTE base station poll counts each context number
According to the number of repetition of every kind of content of node;
At the end of each measurement period, according to the repetition for every kind of content for counting obtained each context data node
Number calculates the multiplicity of every kind of content of each node, and the highest preset number kind content of multiplicity is put into generic configuration
Template, and, for being put into the node content of General Configuration Template, if having the node content of user and being put into General Configuration Template
The node content it is identical, then discharge the memory block where the node content of the user, meanwhile, by the context number of the user
The pointer of the node content in General Configuration Template is replaced with according to the corresponding pointer of mark of the node in list.
The method further includes: the call number of each context data node of the base station LET real-time statistics;
And at the end of each measurement period, the calling rate of each context data node is calculated,
And described the highest preset number kind content of multiplicity is put into General Configuration Template includes:
According to the weight distributed in advance for calling rate and multiplicity, the calling rate and again of each context data node is calculated
The weighted sum of multiplicity;
After the weighted sum of calling rate and multiplicity that all context data nodes are calculated, according to weighted sum value
Be ranked up from high to low, the first forward preset number node of selected and sorted;
For selected each node, the highest second preset number kind content of the multiplicity of the node is selected, by institute
Second preset number kind content of selection is put into General Configuration Template.
It is preset after the highest second preset number kind content of multiplicity for selecting the node, by selected second
Number kind content further comprises before being put into General Configuration Template:
For every kind of content in selected second preset number kind content, if the multiplicity of the content is greater than default threshold
Value, it is determined that the content is put into General Configuration Template, otherwise, it determines the content is not put into General Configuration Template.
It is used according to the planning of the length of each node in user context data structure predetermined and this LTE base station
Amount initializes barrel type DDR memory pool, wherein each node of each user is at least one piece corresponding in barrel type DDR memory pool
Length is not less than the DDR memory block of the node's length;
According to one memory block of length application of the node predetermined are as follows:
According to the length of the node predetermined, to one DDR memory block of barrel type DDR memory pool application;
The General Configuration Template is located in M2 cache;
It is described the highest preset number kind content of multiplicity is put into General Configuration Template to include:
Every kind of content in preset number kind content highest for multiplicity, according to where this kind of content predetermined
The content is put into the cache blocks applied by the length of node to one cache blocks of M2 cache application.
The method further includes:
When the content of any context data node of any user changes, if finding the context number of the user
It is directed toward General Configuration Template according to the corresponding pointer of the mark of the node in list, then is again one memory block of node application, and
New content is written to the memory block applied, meanwhile, the mark of the node in the context data list of the user is corresponding
The pointer for the memory block that this application that pointer replaces with arrives.
A kind of user context data compression set, the device include:
Context writing module: finding that any context data node of any LTE user produces content, then according to pre-
The memory block is written in the content by one memory block of length application of the node first defined, and the user's of self maintained
In context data list, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded;
Statistical module: when the LTE base station free time, each context data node of each user of poll counts each context
The number of repetition of every kind of content of back end, at the end of each measurement period, according to the obtained each context number of statistics
According to the number of repetition of every kind of content of node, the multiplicity of every kind of content of each node is calculated;
Compression module: being put into General Configuration Template for the highest preset number kind content of multiplicity, and, it is general for being put into
The node content of configuration template, if having, the node content of user is identical as the node content of General Configuration Template is put into,
The memory block where the node content of the user is discharged, meanwhile, by the mark of the node in the context data list of the user
Know the pointer that corresponding pointer replaces with the node content in General Configuration Template.
The statistical module is further used for, the call number of each context data node of real-time statistics;And each
At the end of measurement period, the calling rate of each context data node is calculated;
And the highest preset number kind content of multiplicity is put into General Configuration Template and specifically included by the compression module:
According to the weight distributed in advance for calling rate and multiplicity, the calling rate and again of each context data node is calculated
The weighted sum of multiplicity;After the weighted sum of calling rate and multiplicity that all context data nodes are calculated, according to weighting
With being ranked up from high to low for value, the first forward preset number node of selected and sorted;For selected each section
Point selects the highest second preset number kind content of the multiplicity of the node, by receiving in selected second preset number kind
Enter General Configuration Template.
The compression module select the highest second preset number kind content of the multiplicity of the node after, will be selected
Second preset number kind content further comprises before being put into General Configuration Template:
For every kind of content in selected second preset number kind content, if the multiplicity of the content is greater than default threshold
Value, it is determined that the content is put into General Configuration Template, otherwise, it determines the content is not put into General Configuration Template.
The context writing module is according to one memory block of length application of the node predetermined are as follows:
According to the length of the node predetermined, to one DDR memory block of barrel type DDR memory pool application, wherein preparatory root
According to the length of each node in user context data structure predetermined and the planning number of users of this LTE base station, initially
Change barrel type DDR memory pool, wherein each node of each user at least corresponding block length in barrel type DDR memory pool is not small
In the DDR memory block of the node's length;
And the General Configuration Template is located in M2 cache;The compression module is by the highest present count of multiplicity
Mesh kind content is put into General Configuration Template and includes:
Every kind of content in preset number kind content highest for multiplicity, according to where this kind of content predetermined
The content is put into the cache blocks applied by the length of node to one cache blocks of M2 cache application.
The context writing module is further used for,
When the content of any context data node of any user changes, if finding the context number of the user
It is directed toward General Configuration Template according to the corresponding pointer of the mark of the node in list, then is again one memory block of node application, and
New content is written to the memory block applied, meanwhile, the mark of the node in the context data list of the user is corresponding
The pointer for the memory block that this application that pointer replaces with arrives.
The present invention passes through the multiplicity for counting every kind of content of each context data node, by the high interior receiving of multiplicity
Enter General Configuration Template preservation, to improve the compression ratio of user context data;
Further, the embodiment of the present invention by the calling rate of context data node in conjunction with the multiplicity of content, to determine
It is fixed which content is put into General Configuration Template, so that user context data compresses more realistic needs.
Detailed description of the invention
Fig. 1 is the user context data compression method flow chart that one embodiment of the invention provides;
Fig. 2 be another embodiment of the present invention provides user context data compression method flow chart;
Fig. 3 is the schematic diagram of user context data structure;
Fig. 4 is the schematic diagram that the present invention carries out annular poll to user context data;
Fig. 5 is the exemplary diagram of user context data provided by the invention compression;
Fig. 6 is user's context number in the case of compression algorithm is not used and uses two kinds of compression algorithm provided by the invention
According to occupancy memory with runing time variation schematic diagram;
Fig. 7 is the structural schematic diagram of user context data compression set provided in an embodiment of the present invention.
Specific embodiment
With reference to the accompanying drawing and specific embodiment the present invention is further described in more detail.
Inventor the analysis found that: although electric power wireless private network networked users' number is huge, terminal type and gathering
It is engaged in more single, such as terminal capability description, business QOS (Quality of Service, service quality) description, power control etc.
It is almost the same.So user's context content similarity is high, this lays the foundation for the feasibility of context compression.
In addition, private electric power business is that length is online, after user's attachment, the probability that type of service is reconfigured is extremely low.Equipment
Interior number of users and its context are also extremely stable.So context, once being formed after pressure texture, maintenance cost is very low.
In view of the high real-time requirements of system, inventor expects: the word of ultrahigh in efficiency (is decompressed) using read operation
Allusion quotation compression scheme.
Fig. 1 is the user context data compression method flow chart that one embodiment of the invention provides, the specific steps of which are as follows:
The base station step 101:LTE finds that any context data node of any LTE user produces content, then according to pre-
The memory block is written in the content by one memory block of length application of the node first defined, and the user's of self maintained
In context data list, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded.
Step 102: when the LTE base station free time, LTE base station according to the context data list of each user of self maintained,
Each context data node of each user of poll, counts the number of repetition of every kind of content of each context data node.
Step 103: at the end of each measurement period, according in every kind that counts obtained each context data node
The number of repetition of appearance calculates the multiplicity of every kind of content of each node, and the highest preset number kind content of multiplicity is put into
General Configuration Template, and, for being put into the node content of General Configuration Template, if having the node content of user be put into it is general
The node content of configuration template is identical, then discharges the memory block where the node content of the user, meanwhile, by the user's
The corresponding pointer of mark of the node replaces with the pointer of the node content in General Configuration Template in context data list.
Fig. 2 be another embodiment of the present invention provides user context data compression method flow chart, specific steps are such as
Under:
Step 201: being each of the structure in LTE base station according to user context data structure predetermined
Context data node distributes unique key assignments respectively.
User context data structure is usually tree structure, as shown in Figure 3, wherein K1, K2, K3 ... be each context
The key assignments of back end, for uniquely identifying corresponding node.
Step 202: according to the length and this LTE base station of each node in user context data structure predetermined
Planning number of users, initialize barrel type DDR (Double Data Rate, Double Data Rate) memory pool, wherein each user
Each node in barrel type DDR memory pool an at least corresponding block length be not less than the DDR memory block of the node's length.
That is, the number of nodes in user context data structure is N, then barrel type if the planning number of users of this LTE base station is M
N*M DDR memory block is included at least in DDR memory pool, each context data node of each user is at least corresponding respectively
A DDR memory block in barrel type DDR memory pool, and the length of DDR memory block is not less than the length of corresponding node.
Any context data node of the base station step 203:LTE discovery any user produces content, then according to the section
To one DDR memory block of barrel type DDR memory pool application the DDR memory block is written in the content by the length of point, and in self maintained
The user context data list in, record the corresponding relationship of the key assignments of the node and the pointer of the DDR memory block.
The call number of each context data node of the base station step 204:LET real-time statistics.
The call number of each context data node is equal to the call number of the context data node of all users
The sum of.Such as: for node Kn, as long as there is the node Kn of any user called, the call number of node Kn is added 1.
Step 205: when the LTE base station free time, LTE base station according to the context data list of each user of self maintained,
Each context data node of each user of poll, counts the number of repetition of every kind of content of each context data node.
Such as: obtained by each context data node of each user of poll: for node Kn, there are two types of contents: Vn_
1, Vn_2, wherein the content for having the node Kn of p user is Vn_1, and the content for having the node Kn of q user is Vn_2, then: section
The number of repetition of the content Vn_1 of point Kn is p, and the number of repetition of the content Vn_2 of node Kn is q.
As shown in figure 4, all users constitute a ring, each context of each user on LTE base station cyclic polling ring
Back end counts the number of repetition of the various contents of each context data node.
It should be noted that the section of each free time of LTE base station be it is very short, start in next free time section
When, LTE base station will the last one user that then last free time section is polled to continue poll.
Step 206: at the end of each measurement period, the calling rate and multiplicity of each context data node are calculated,
And according to the weight distributed in advance for calling rate and multiplicity, the calling rate and multiplicity of each context data node are calculated
Weighted sum.
Wherein,
Total call number/the statistical time of the calling rate of each context data node=context data node, if
For formula one;
When the multiplicity of each context data node=context data node content highest number of repetition/statistics
Between, it is set as formula two, wherein the highest number of repetition of the context data node content weighs that is, in all the elements of the node
Again the number of repetition of that highest content is counted.
In general, statistical time=measurement period, accordingly, " total calling time of the context data node in formula one
Total call number-upper one of the context data node obtained at the end of number "=this measurement period obtains when counting end cycle
Total call number of the context data node arrived;
It is obtained at the end of " the highest number of repetition of the context data node content "=this measurement period in formula two
The context data node content a highest number of repetition-upper statistics end cycle when the obtained context data node
The highest number of repetition of content.
Step 207: after the weighted sum of calling rate and multiplicity that all context data nodes are calculated, according to adding
Power is ranked up from high to low with value, the first forward preset number node of selected and sorted.
Step 208: for selected each node, selecting in the highest second preset number kind of the multiplicity of the node
Hold, for selected every kind of content, if the multiplicity of the content is greater than preset threshold, according to the length of the node to M2 high
Speed caching one cache blocks of application, which is put into the cache blocks.
Further, in this step, for selected every kind of content, if the multiplicity of the content is greater than preset threshold,
First according to the corresponding pointer of node where the content, judge whether the pointer has pointed to M2 cache, if so, determining that this is interior
Appearance has deposited in M2 cache, is not for further processing;Otherwise, it is slow to M2 high speed to execute the subsequent length according to the node
One cache blocks of application are deposited, which is put into the movement in the cache blocks.
In addition, if the multiplicity of the content is greater than preset threshold, confirming in the node for selected every kind of content
Appearance should be put into M2 cache, after having determined all node contents that should be put into M2 cache in this measurement period,
It will be stored in M2 cache but be not belonging to determine in this measurement period should be put into the node of M2 cache
M2 cacheline release where holding, meanwhile, according to the pointer and corresponding node identification of the M2 cacheline of release
The mark of each node in the context data list of each user and the corresponding relationship of pointer are searched, if matching, basis should
The length of node applies for a DDR memory block for the node content of the user again, should by the node content write-in of the user
DDR memory block, and the corresponding pointer of mark of the node in the context data list of the user is replaced with into the DDR memory block
Pointer.
Step 209: the node content for being put into M2 cacheline inquires the content of the node of each user, if having
The content of the node of user is identical as the node content of M2 cacheline is put into, then where the content for discharging the node
DDR memory block, meanwhile, the key assignments of the node is searched in the context data list of the user, by the key-value pair of the node
The pointer for the DDR memory block answered replaces with the pointer of the M2 cacheline.
As shown in figure 5, at a time the storage condition of the context data of a certain user is as follows:
The content V1 of K1 node is written with DDR memory block DB3-1;
The content V2 of K2 node is written with DDR memory block DB1-1;
The content V4 of K4 node is written with DDR memory block DB2-1;
The content V5 of K5 node is written with DDR memory block DB3-3;
The content V6 of K6 node is written with M2 cacheline T2-4;
In current statistic end cycle, the content V5 of discovery K5 node should be transferred to M2 cacheline T3-2,
Then discharge the DDR memory block DB3-3 at the original place V5.
Further, when the content of any context data node of any user changes, if finding the user
Context data list in the corresponding pointer of key assignments of the node be directed toward M2 cacheline, then again to barrel type DDR memory
Pond is the one DDR memory block of node application, and new content is written to the DDR memory block applied, meanwhile, above and below the user
The corresponding pointer of the key assignments of the node replaces with the pointer for the DDR memory block that this application arrives in literary data list.
Present invention is particularly suitable for the high scenes of user context data multiplicity, such as electric power wireless private network.
Application example of the invention is given below:
If being narrowband/broadband 8:1 configuration in network termination in private electric power, user context data multiplicity is 94%,
100,000 LTE user terminals access LTE network is subjected to shock-testing in 2 minutes.
In the case of Fig. 6 gives unused compression algorithm and uses two kinds of compression algorithm provided by the invention, above and below user
The memory that literary data occupy with runing time variation schematic diagram, wherein heavy black, which represents, is not used compression algorithm, grey white line generation
Table has used compression algorithm provided by the invention.
As shown in the heavy black of Fig. 6, the context data of each user is directly stored in memory if compression algorithm is not used,
Then the memory overhead of context data is substantially linear with number of users.Memory overhead is 510MB under 10 general-purpose family scenes.
As shown in the grey white line of Fig. 6, after compression algorithm provided by the invention, measurement period is set as 20S, and memory is opened
Before pin peak value appears in 20S, reach 75MB;Perfect with context data node, memory overhead is remarkably decreased.In 10 general-purposes
Memory overhead is 55MB under the scene of family.The Rate of average load of the digital processing unit (DSP) of LTE base station is promoted by 58% to 71%,
Analysis reason is consumed by running compression algorithm provided by the invention under Idle state, and compression algorithm provided by the invention will increase
13% or so DSP treating capacity is a kind of scheme that space is changed with the time.
Advantageous effects of the invention are as follows:
By counting the multiplicity of every kind of content of each context data node, the high content of multiplicity is put into general
Configuration template saves, to improve the compression ratio of user context data;
Further, by the calling rate of context data node in conjunction with the multiplicity of content, to determine which content
It is put into General Configuration Template, so that user context data compresses more realistic needs;
In addition, carrying out multiplicity calculating and statistical operation in the LTE base station free time, the shadow to regular traffic is reduced
It rings.
Fig. 7 is the structural schematic diagram of user context data compression set provided in an embodiment of the present invention, which is located at
In LTE base station, the device mainly includes: context writing module 71, statistical module 72 and compression module 73, in which:
Context writing module 71: find that any context data node of any LTE user produces content, then basis
The memory block is written in the content by one memory block of length application of the node predetermined, and in the user of self maintained
Context data list in, record the mark and the corresponding relationship of the pointer of the memory block of the node.
Statistical module 72: when the LTE base station free time, the context number for each user that based on context writing module 71 is safeguarded
According to list, each context data node of each user of poll counts the repetition time of every kind of content of each context data node
Number, at the end of each measurement period, according to the number of repetition for every kind of content of each context data node that statistics obtains,
Calculate the multiplicity of every kind of content of each node.
Compression module 73: according to the multiplicity of every kind of content of each node of the calculating of statistical module 72, most by multiplicity
High preset number kind content is put into General Configuration Template, and, for being put into the node content of General Configuration Template, if there is user
The node content it is identical as the node content of General Configuration Template is put into, then where discharging the node content of the user
Memory block, meanwhile, the mark of the node is corresponding in the context data list for the user that context writing module 71 is safeguarded
Pointer replace with the pointer of the node content in General Configuration Template.
Statistical module 72 is further used for, the call number of each context data node of real-time statistics;And in each system
When counting end cycle, the calling rate of each context data node is calculated;
And the highest preset number kind content of multiplicity is put into General Configuration Template and specifically included by compression module 73: according to
The weight distributed in advance for calling rate and multiplicity, according to the calling rate for each context data node that statistical module 72 calculates
With the content multiplicity of the node, the calling rate of each context data node and the weighted sum of content multiplicity are calculated;Work as meter
After calculation obtains the calling rate of all context data nodes and the weighted sum of content multiplicity, according to weighted sum value slave height to
It is low to be ranked up, the first forward preset number node of selected and sorted;For selected each node, the node is selected
Selected second preset number kind content is put into General Configuration Template by the highest second preset number kind content of multiplicity.
Compression module 73 select the highest second preset number kind content of multiplicity of the node after, by selected the
Two preset number kind contents further comprise before being put into General Configuration Template: for selected second preset number kind content
In every kind of content, if the multiplicity of the content be greater than preset threshold, it is determined that the content is put into General Configuration Template, it is no
Then, which is not put into General Configuration Template by determination.
Context writing module 71 is according to one memory block of length application of the node predetermined are as follows: according to pre-defined
The node length, to one DDR memory block of barrel type DDR memory pool application, wherein above and below user predetermined
The length of each node in literary data structure and the planning number of users of this LTE base station initialize barrel type DDR memory pool, wherein
Each node of each user at least corresponding block length in barrel type DDR memory pool is not less than the DDR memory of the node's length
Block;
And General Configuration Template is located in M2 cache;Compression module 73 will be in the highest preset number kind of multiplicity
Receiving General Configuration Template includes: every kind of content in preset number kind content highest for multiplicity, according to fixed in advance
The content is put into the caching applied to one cache blocks of M2 cache application by the length of node where this kind of content of justice
In block.
Context writing module 71 is further used for, when the content of any context data node of any user changes
When change, if finding, the corresponding pointer of the mark of the node in the context data list of the user is directed toward General Configuration Template,
Again it is one memory block of node application, and new content is written to the memory block applied, meanwhile, by the context number of the user
The pointer for the memory block that this application arrives is replaced with according to the corresponding pointer of mark of the node in list.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (10)
1. a kind of user context data compression method, which is characterized in that this method comprises:
Long term evolution LTE base station finds that any context data node of any LTE user produces content, then according to fixed in advance
The memory block is written in the content by one memory block of length application of the node of justice, and above and below the user of self maintained
In literary data list, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded;
When the LTE base station free time, each context data node of each user of LTE base station poll counts each context data section
The number of repetition of every kind of content of point;
At the end of each measurement period, according to the repetition for every kind of content for counting obtained each context data node time
Number, calculates the multiplicity of every kind of content of each node, the highest preset number kind content of multiplicity is put into generic configuration mould
Plate, and, for being put into the node content of General Configuration Template, if having the node content of user and being put into General Configuration Template
The node content is identical, then discharges the memory block where the node content of the user, meanwhile, by the context data of the user
The corresponding pointer of mark of the node replaces with the pointer of the node content in General Configuration Template in list.
2. the method according to claim 1, wherein the method further includes: the base station LET real-time statistics are every
The call number of a context data node;
And at the end of each measurement period, the calling rate of each context data node is calculated,
And described the highest preset number kind content of multiplicity is put into General Configuration Template includes:
According to the weight distributed in advance for calling rate and multiplicity, the calling rate and multiplicity of each context data node are calculated
Weighted sum;
When the weighted sum of calling rate and multiplicity that all context data nodes are calculated after, according to weighted sum value from
It is high to Low to be ranked up, the first forward preset number node of selected and sorted;
For selected each node, the highest second preset number kind content of the multiplicity of the node is selected, it will be selected
The second preset number kind content be put into General Configuration Template.
3. according to the method described in claim 2, it is characterized in that, the multiplicity highest second for selecting the node is default
Further comprise before being put into General Configuration Template after number kind content, by selected second preset number kind content:
For every kind of content in selected second preset number kind content, if the multiplicity of the content is greater than preset threshold,
It then determines and the content is put into General Configuration Template, otherwise, it determines the content is not put into General Configuration Template.
4. the method according to claim 1, wherein according in user context data structure predetermined
The length of each node and the planning number of users of this LTE base station initialize barrel type DDR memory pool, wherein each of each user
Node at least corresponding block length in barrel type DDR memory pool is not less than the DDR memory block of the node's length;
According to one memory block of length application of the node predetermined are as follows:
According to the length of the node predetermined, to one DDR memory block of barrel type DDR memory pool application;
The General Configuration Template is located in M2 cache;
It is described the highest preset number kind content of multiplicity is put into General Configuration Template to include:
Every kind of content in preset number kind content highest for multiplicity, according to node where this kind of content predetermined
Length the content is put into the cache blocks applied to one cache blocks of M2 cache application.
5. the method according to claim 1, wherein the method further includes:
When the content of any context data node of any user changes, if finding the context data column of the user
The corresponding pointer of the mark of the node is directed toward General Configuration Template in table, then is again one memory block of node application, and will be new
The memory block applied is written in content, meanwhile, by the corresponding pointer of mark of the node in the context data list of the user
Replace with the pointer for the memory block that this application arrives.
6. a kind of user context data compression set, which is characterized in that the device includes:
Context writing module: finding that any context data node of any LTE user produces content, then according to fixed in advance
The memory block is written in the content by one memory block of length application of the node of justice, and above and below the user of self maintained
In literary data list, the mark and the corresponding relationship of the pointer of the memory block of the node are recorded;
Statistical module: when the LTE base station free time, each context data node of each user of poll counts each context data
The number of repetition of every kind of content of node, at the end of each measurement period, according to the obtained each context data section of statistics
The number of repetition of every kind of content of point, calculates the multiplicity of every kind of content of each node;
Compression module: being put into General Configuration Template for the highest preset number kind content of multiplicity, and, for being put into generic configuration
The node content of template discharges if there is the node content of user identical as the node content of General Configuration Template is put into
Memory block where the node content of the user, meanwhile, by the mark pair of the node in the context data list of the user
The pointer answered replaces with the pointer of the node content in General Configuration Template.
7. device according to claim 6, which is characterized in that the statistical module is further used for, and real-time statistics are each
The call number of context data node;And at the end of each measurement period, the calling of each context data node is calculated
Rate;
And the highest preset number kind content of multiplicity is put into General Configuration Template and specifically included by the compression module:
According to the weight distributed in advance for calling rate and multiplicity, the calling rate and multiplicity of each context data node are calculated
Weighted sum;After the weighted sum of calling rate and multiplicity that all context data nodes are calculated, taken according to weighted sum
Value is ranked up from high to low, the first forward preset number node of selected and sorted;For selected each node, choosing
The highest second preset number kind content of multiplicity for selecting the node, selected second preset number kind content is put into general
Configuration template.
8. device according to claim 7, which is characterized in that the compression module selects the multiplicity of the node highest
Taking a step forward for General Configuration Template is put into after second preset number kind content, by selected second preset number kind content
Include:
For every kind of content in selected second preset number kind content, if the multiplicity of the content is greater than preset threshold,
It then determines and the content is put into General Configuration Template, otherwise, it determines the content is not put into General Configuration Template.
9. device according to claim 6, which is characterized in that the context writing module is according to the section predetermined
One memory block of length application of point are as follows:
According to the length of the node predetermined, to one DDR memory block of barrel type DDR memory pool application, wherein previously according to pre-
The length of each node in user context data structure first defined and the planning number of users of this LTE base station initialize bucket
Formula DDR memory pool, wherein each node of each user at least corresponding block length in barrel type DDR memory pool is not less than should
The DDR memory block of node's length;
And the General Configuration Template is located in M2 cache;The compression module is by the highest preset number kind of multiplicity
Content is put into General Configuration Template
Every kind of content in preset number kind content highest for multiplicity, according to node where this kind of content predetermined
Length the content is put into the cache blocks applied to one cache blocks of M2 cache application.
10. the apparatus according to claim 1, which is characterized in that the context writing module is further used for,
When the content of any context data node of any user changes, if finding the context data column of the user
The corresponding pointer of the mark of the node is directed toward General Configuration Template in table, then is again one memory block of node application, and will be new
The memory block applied is written in content, meanwhile, by the corresponding pointer of mark of the node in the context data list of the user
Replace with the pointer for the memory block that this application arrives.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711030139.0A CN109725836B (en) | 2017-10-30 | 2017-10-30 | User context compression method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711030139.0A CN109725836B (en) | 2017-10-30 | 2017-10-30 | User context compression method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109725836A true CN109725836A (en) | 2019-05-07 |
CN109725836B CN109725836B (en) | 2021-11-26 |
Family
ID=66292463
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711030139.0A Active CN109725836B (en) | 2017-10-30 | 2017-10-30 | User context compression method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109725836B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110895492A (en) * | 2019-12-11 | 2020-03-20 | Oppo(重庆)智能科技有限公司 | Device control method, device, storage medium and electronic device |
CN110955612A (en) * | 2019-11-07 | 2020-04-03 | 浪潮电子信息产业股份有限公司 | Data caching method and related device |
CN113051175A (en) * | 2021-04-19 | 2021-06-29 | 杭州至千哩科技有限公司 | Extensible general workflow framework system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103177111A (en) * | 2013-03-29 | 2013-06-26 | 西安理工大学 | System and method for deleting repeating data |
US20150254020A1 (en) * | 2014-03-04 | 2015-09-10 | Netapp, Inc. | Background checking for lost writes and data corruption |
CN105824881A (en) * | 2016-03-10 | 2016-08-03 | 中国人民解放军国防科学技术大学 | Repeating data and deleted data placement method and device based on load balancing |
CN106230564A (en) * | 2016-07-27 | 2016-12-14 | 重庆重邮汇测通信技术有限公司 | The weight fragment data storage of wireless chain control layer determination transmission mode and method for sorting |
-
2017
- 2017-10-30 CN CN201711030139.0A patent/CN109725836B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103177111A (en) * | 2013-03-29 | 2013-06-26 | 西安理工大学 | System and method for deleting repeating data |
US20150254020A1 (en) * | 2014-03-04 | 2015-09-10 | Netapp, Inc. | Background checking for lost writes and data corruption |
CN105824881A (en) * | 2016-03-10 | 2016-08-03 | 中国人民解放军国防科学技术大学 | Repeating data and deleted data placement method and device based on load balancing |
CN106230564A (en) * | 2016-07-27 | 2016-12-14 | 重庆重邮汇测通信技术有限公司 | The weight fragment data storage of wireless chain control layer determination transmission mode and method for sorting |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110955612A (en) * | 2019-11-07 | 2020-04-03 | 浪潮电子信息产业股份有限公司 | Data caching method and related device |
CN110955612B (en) * | 2019-11-07 | 2022-03-08 | 浪潮电子信息产业股份有限公司 | Data caching method and related device |
CN110895492A (en) * | 2019-12-11 | 2020-03-20 | Oppo(重庆)智能科技有限公司 | Device control method, device, storage medium and electronic device |
CN110895492B (en) * | 2019-12-11 | 2023-01-10 | Oppo(重庆)智能科技有限公司 | Device control method, device, storage medium and electronic device |
CN113051175A (en) * | 2021-04-19 | 2021-06-29 | 杭州至千哩科技有限公司 | Extensible general workflow framework system |
Also Published As
Publication number | Publication date |
---|---|
CN109725836B (en) | 2021-11-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105183560B (en) | Method for adjusting picture, picture adjusting device and system | |
CN107547154B (en) | Method and device for establishing video traffic prediction model | |
CN109725836A (en) | User's context compression method and device | |
CN108462785B (en) | Method and device for processing malicious call | |
JP2018525859A (en) | Method and apparatus for network capacity expansion | |
CN109327856B (en) | Passenger flow volume statistical method, network equipment and storage medium | |
CN1611087A (en) | Method of logging call processing information in a mobile communication network | |
CN104618162B (en) | A kind of management method of system docking, device and system | |
CN108055701A (en) | A kind of resource regulating method and base station | |
CN109547393B (en) | Malicious number identification method, device, equipment and storage medium | |
CN101568182B (en) | Wireless resource allocation method and device | |
CN108632088B (en) | Method for processing business, device and server | |
CN104486769A (en) | Method and device for selecting valuable cell | |
CN113795032B (en) | Method and device for judging invisible faults of indoor division, storage medium and equipment | |
CN113966602A (en) | Distributed storage of blocks in a blockchain | |
CN106912066B (en) | A kind of calculation method and device of KPI Key Performance Indicator | |
CN104462422A (en) | Object processing method and device | |
CN110134547B (en) | Middleware-based repeated data deleting method and related device | |
CN109428774A (en) | A kind of data processing method and relevant DPI equipment of DPI equipment | |
CN110868732B (en) | VoLTE radio access failure problem positioning method, system and equipment | |
CN102724687B (en) | Method for acquiring wireless network utilization rate and network management equipment | |
CN111144687A (en) | Novel electric power big data analysis system | |
CN105187472A (en) | Data synchronization method and data synchronization device | |
CN110650483A (en) | Equipment model selection method and device | |
CN111581931B (en) | Report generation method and device based on data leakage prevention system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |