CN109271390A - Index data structure based on neural network and data retrieval method thereof - Google Patents
Index data structure based on neural network and data retrieval method thereof Download PDFInfo
- Publication number
- CN109271390A CN109271390A CN201811152887.0A CN201811152887A CN109271390A CN 109271390 A CN109271390 A CN 109271390A CN 201811152887 A CN201811152887 A CN 201811152887A CN 109271390 A CN109271390 A CN 109271390A
- Authority
- CN
- China
- Prior art keywords
- data
- bitmap
- label
- index
- data structure
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an index data structure based on a neural network, which comprises a data mapping unit and a dynamic index unit; the data mapping unit is a data mapping unit based on a neural network model, collects name data in a named data network as a sample, calculates an accumulated distribution function value of the sample as a label, and trains to obtain the neural network model for mapping the name data to a corresponding position in an improved bitmap data structure; the dynamic index unit is based on the dynamic index unit of the improved bitmap data structure, the traditional bitmap is averagely divided into a plurality of buckets, the size of each slot in each bucket is expanded, and the improved bitmap data structure with the dynamic label is obtained and is used for storing the address offset corresponding to the name data. Aiming at the characteristics of a forwarding plane of a named data network, the invention can improve the storage efficiency and realize rapid data insertion and retrieval operation under the condition of ensuring the retrieval speed and the misjudgment probability.
Description
Technical field
The invention belongs to high-performance router index data structure design fields, particular for name data network (Named
Data Networking) efficient storage and quick-searching problem of index content in Forwarding plane.
Background technique
It the innovation and applications such as ultra high-definition video, artificial intelligence, cloud computing, technology and calculates mode and continues to bring out, accelerate
Diversification in role of the internet from " communication channel " to " data processing platform (DPP) ".The existing IP Generation Internet knot based on device address
Structure information sharing, mobility, safety, in terms of there are the drawbacks of have become hinder internet development it is great
Problem.Therefore, a kind of novel Future Internet framework names data network, was suggested in 2010, has obtained domestic and international academia
Extensive concern.
It names data network to replace IP address with data name, uses the communication pattern for being perfectly facing data content.Its advantage
It is to realize that data content truly is shared by disposing buffer storage in routing node, greatly reduce net
Network load, effectively improves network data transmission rate.It therefore, is considered as that Future Internet framework field is most promising
One of developing direction.
However, name data network is also faced with a series of urgent problems to be solved and challenge, especially rope in Forwarding plane
Draw the efficient storage and quick-searching problem of content.Wherein, index data structure is the key that improve Forwarding plane performance, but mesh
Preceding main results all respectively have advantage and disadvantage.For example, the data structure lookup speed based on dictionary tree is slower；Based on the grand filtering of cloth
The data structure of device can not direct index data；And the data structure based on Hash table needs great memory space.Therefore, when
Preceding research achievement can not all meet the needs of name data network Forwarding plane is to retrieval rate and memory space simultaneously, be badly in need of mentioning
New Research Thinking out, the index data structure and its data retrievad algorithm of design synthesis excellent performance are to solve this problem.
Summary of the invention
In view of the above-mentioned problems, the present invention provides a kind of index data structure neural network based and its data retrieval sides
Method.The present invention can be promoted for name data network Forwarding plane feature under conditions of guaranteeing retrieval rate and probability of miscarriage of justice
Storage efficiency realizes rapid data insertion and search operaqtion.
In order to solve the above technical problems, the present invention proposes a kind of index data structure neural network based, including
Data mapping unit and dynamic index unit；The data mapping unit is the data mapping unit based on neural network model,
Name data in acquisition name data network calculates its cumulative distribution function value as label as sample, according to sample with
Label training reverse transmittance nerve network, obtains neural network model, for the name data to be retrieved to be mapped to modified
Corresponding position in bitmap bitmap data structure；The dynamic index unit is based on modified bitmap bitmap data structure
Dynamic index unit, traditional bitmap bitmap data structure is divided into several buckets, and the size of slot each in bucket is expanded
Exhibition, obtain can dynamic label modified bitmap bitmap data structure, D-bitmap is denoted as, for storing the title to be retrieved
Address offset amount corresponding to data.
A kind of data retrieval method of above-mentioned index data structure neural network based is proposed in the present invention simultaneously, it is main
It is included in index data structure and is inserted into data, and data retrieval is carried out to the index data structure after insertion data；
Specific step is as follows:
Step 1: name data is inserted into the index data structure, one name data of every insertion, including following step
It is rapid:
Step 1-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 1-2: the neural computing of data mapping unit: above-mentioned mind will be inputted after name data fixed length processing
Through network model operation, real number value of the range between 0,1 is obtained；
Step 1-3: the position mapping calculation of data mapping unit: the slot that neural computing result is multiplied D-bitmap is total
Number, show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 1-4: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and taking downwards
It is whole, obtain the bucket serial number where the position；
Step 1-5: the maximum label of dynamic index unit is searched: the bucket serial number obtained by step 1-4 traverses institute in the bucket
There is slot, searches existing maximum label；
Step 1-6: the slot internal label of dynamic index unit records: the existing maximum label that judgment step 1-5 is found is
It is no to reach maximum value, if also unmarked arrive maximum value, the label that existing maximum label adds one is recorded in the location groove, it is no
Then, take the label deleted earliest as the label in the location groove from deletion queue；
Step 1-7: for the name data application storage unit: according to the corresponding base of bucket serial number in dynamic index unit
The address offset amount that location and slot internal label represent is the name data application storage unit in data space；
Step 1-8: name data insertion operation terminates；
Step 2: being retrieved in index data structure to a name data, comprising the following steps:
Step 2-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 2-2: the neural computing of data mapping unit: above-mentioned mind will be inputted after name data fixed length processing
Through network model operation, real number value of the range between 0,1 is obtained；
Step 2-3: the position mapping calculation of data mapping unit: the slot that neural computing result is multiplied D-bitmap is total
Number, show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 2-4: the data existence judgement of dynamic index unit: judging whether the label at the position is 0, if mark
It number is not 0, then the name data is present in the index data structure, and continues to execute step 2-5, completes retrieval；Otherwise, should
Name data is not present in the index data structure, that is, shows to execute step 2-7 there is no search result；
Step 2-5: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and taking downwards
It is whole, obtain the bucket serial number where the position；
Step 2-6: the bucket serial number and slot internal label where the position, i.e. data storage cell output search result: are exported
Base address and address offset amount of the name data in data space relative to base address；
Step 2-7: name data search operaqtion terminates.
Compared with prior art, the beneficial effects of the present invention are:
Intel is configured at one by index data structure neural network based and its data retrieval method of the invention
Xeon E5-1650v2 3.50GHz, DDR3 24GB SDRAM computer on carry out deployment realization.Wherein, the name of training
Claim data that can acquire in the actual environment by the Forwarding plane of name data network, and under laboratory environment, we are using often
See the English word name data similar with top level domain 100,000,000 wiht strip-lattice types of construction, obtains that uniform mapping can be achieved by training
Neural network model.In performance test, it is contemplated that the data volume of name data network Forwarding plane processing in million ranks, we
1,000,000 name datas of same distribution rule are taken to test the performance of the index data structure and its data retrieval method.Experiment
The result shows that storage consumption of the invention is far smaller than the storage consumption of hash index, while retrieval rate and hash index phase
When, also can satisfy current internet network False Rate be lower than 1% communication requirement.It is indicated above that designed in the present invention based on
The index data structure and its data retrieval method of neural network, can guaranteeing retrieval rate and under conditions of probability of miscarriage of justice,
Storage efficiency is promoted, it is with good performance.
Detailed description of the invention
Fig. 1 is that the present invention is based on the schematic diagrams of the index data structure of neural network；
Fig. 2 is that the present invention is based on the signals that data mapping unit in the index data structure of neural network runs specific steps
Figure；
Fig. 3 is the flow diagram in data retrieval method of the present invention about data insertion operation；
Fig. 4 is the flow diagram in data retrieval method of the present invention about data retrieval operation.
Specific embodiment
Technical solution of the present invention is described in further detail in the following with reference to the drawings and specific embodiments, it is described specific
Embodiment is only explained the present invention, is not intended to limit the invention.
As shown in Figure 1, a kind of index data structure neural network based proposed by the present invention, including data mapping unit
With dynamic index unit；It is described as follows:
The data mapping unit is the data mapping unit based on neural network model, in acquisition name data network
Name data calculates its cumulative distribution function value as label, according to sample and label training Back propagation neural as sample
Network (Back Propagation Neural Network, BPNN), obtains neural network model, for the name to be retrieved
Data are claimed to be mapped to the corresponding position in modified bitmap bitmap data structure.Wherein, the title number in data network is named
According to acquisition name data network actual deployment in Forwarding plane realize, pass through record certain time in Forwarding plane processing
All name datas, obtain sufficient amount of sample data.Its cumulative distribution function value is calculated to be as the reason of label,
Being uniformly distributed between 0,1 is obeyed in the distribution that the data of obedience Arbitrary distribution are worth after cumulative distribution function converts.Therefore, will
The name data of acquisition calculates its cumulative distribution function value as label, the uniform mapping of data can be realized as sample, from
And optimize storage unit utilization efficiency.
It is as shown in Figure 2 that the data mapping unit runs specific steps.In the training stage, name data and work as sample
Training dataset is collectively formed for the cumulative distribution function value of label, utilizes training dataset training reverse transmittance nerve network
BPNN is fitted the cumulative distribution function of training dataset, obtains the relevant parameters such as weight and the biasing of hidden layer and output layer, shape
At the neural network model of achievable uniform mapping.The actual prediction stage after the completion of training inputs a name data, passes through
The value that neural network model operation obtains multiplies mapping size, obtains the mapping position of the name data.
The dynamic index unit is the dynamic index unit based on modified bitmap bitmap data structure, by traditional bit
Figure bitmap data structure is divided into several buckets, and the size of slot each in bucket is extended, obtain can dynamic label improvement
Type bitmap bitmap data structure, is denoted as D-bitmap, for storing address offset amount corresponding to the name data to be retrieved.
It is to traditional improved specific method of bitmap bitmap data structure, the bitmap bitmap having a size of m slot is averagely divided into n
A bucket, each bucket contain m/n slot, the data space of a corresponding Fixed Base location.The size of each slot is by original 1
Bit expanded is used for recording address offset at j bit.D-bitmap improved so is dynamically marked in data insertion process
Number, the sequence namely name data that the label in each slot represents bucket where name data enters are on the ground of place memory space
Location offset.It is name data application storage unit according to the address offset amount that the corresponding base address of bucket and slot internal label represent,
The address for being achieved in data storage cell dynamically distributes.
For the data retrieval method of above-mentioned index data structure neural network based, it is included in index data structure
Data are inserted into, and data retrieval is carried out to the index data structure after insertion data；Specific step is as follows:
Step 1: name data is inserted into the index data structure, one name data of every insertion, as shown in figure 3,
The following steps are included:
Step 1-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 1-2: the neural computing of data mapping unit: above-mentioned mind will be inputted after name data fixed length processing
Through network model operation, real number value of the range between 0,1 is obtained；
Step 1-3: the position mapping calculation of data mapping unit: the slot that neural computing result is multiplied D-bitmap is total
Number, show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 1-4: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and taking downwards
It is whole, obtain the bucket serial number where the position；
Step 1-5: the maximum label of dynamic index unit is searched: the bucket serial number obtained by step 1-4 traverses institute in the bucket
There is slot, searches existing maximum label；
Step 1-6: the slot internal label of dynamic index unit records: the existing maximum label that judgment step 1-5 is found is
It is no to reach maximum value, if also unmarked arrive maximum value, the label that existing maximum label adds one is recorded in the location groove, it is no
Then, take the label deleted earliest as the label in the location groove from deletion queue；
Step 1-7: for the name data application storage unit: according to the corresponding base of bucket serial number in dynamic index unit
The address offset amount that location and slot internal label represent is the name data application storage unit in data space；
Step 1-8: name data insertion operation terminates；
In the present invention, the example that name data is inserted into index data structure is as shown in Figure 1.In the index data knot
In structure, data mapping unit is the neural network model that training is completed, dynamic index unit, that is, modified bitmap data structure D-
bitmap.Slot number is set as 32, and every 16 slots are divided into a bucket, therefore the D-bitmap is divided into 2 buckets altogether, right
Answer the data space of 2 Fixed Base locations.The current index data structure has been inserted into two name datas.When third name
When claiming data/ndn/name/3rd insertion, neural computing and position mapping calculation are carried out in data mapping unit, is somebody's turn to do
Mapping position of the name data in D-bitmap is the 6th slot.Into dynamic index unit, the bucket serial number of the position is calculated
It is 1.Search existing maximum label in the 1st bucket, lookup result 1, also not up to maximum value, therefore in the 6th slot internal standard
Remember the label that existing maximum label adds one, is 2 namely the name data in the data store relative to the 1st base address
Address offset amount.
Step 2: being retrieved in index data structure to a name data, as shown in Figure 4, comprising the following steps:
Step 2-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 2-2: the neural computing of data mapping unit: above-mentioned mind will be inputted after name data fixed length processing
Through network model operation, real number value of the range between 0,1 is obtained；
Step 2-3: the position mapping calculation of data mapping unit: the slot that neural computing result is multiplied D-bitmap is total
Number, show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 2-4: the data existence judgement of dynamic index unit: judging whether the label at the position is 0, if mark
It number is not 0, then the name data is present in the index data structure, and continues to execute step 2-5, completes retrieval；Otherwise, should
Name data is not present in the index data structure, that is, shows to execute step 2-7 there is no search result；
Step 2-5: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and taking downwards
It is whole, obtain the bucket serial number where the position；
Step 2-6: the bucket serial number and slot internal label where the position, i.e. data storage cell output search result: are exported
Base address and address offset amount of the name data in data space relative to base address；
Step 2-7: name data search operaqtion terminates.
Embodiment: index data structure neural network based and its data retrieval method of the invention are configured at one
For Intel Xeon E5-1650v2 3.50GHz, DDR3 24GB SDRAM computer on carry out deployment realization.Wherein, it instructs
The name data practiced can be acquired in the actual environment by the Forwarding plane of name data network, and under laboratory environment, I
Construct the similar name data of 100,000,000 wiht strip-lattice types using common English word and top level domain, obtain can be achieved by training equal
The neural network model of even mapping.In performance test, it is contemplated that the data volume of name data network Forwarding plane processing is million
Rank, we take 1,000,000 name datas of same distribution rule to test the index data structure and its data retrieval method
Performance.The experimental results showed that storage consumption of the invention is far smaller than the storage consumption of hash index, while retrieval rate and Kazakhstan
Uncommon index quite, also can satisfy the communication requirement that current internet network False Rate is lower than 1%.It is indicated above that being set in the present invention
The index data structure neural network based and its data retrieval method of meter can guarantee retrieval rate and probability of miscarriage of justice
Under the conditions of, storage efficiency is promoted, it is with good performance.
Although above in conjunction with attached drawing, invention has been described, and the invention is not limited to above-mentioned specific implementations
Mode, the above mentioned embodiment is only schematical, rather than restrictive, and those skilled in the art are at this
Under the enlightenment of invention, without deviating from the spirit of the invention, many variations can also be made, these belong to of the invention
Within protection.
Claims (2)
1. a kind of index data structure neural network based, which is characterized in that including data mapping unit and dynamic index list
Member；
The data mapping unit is the data mapping unit based on neural network model, the title in acquisition name data network
Data calculate its cumulative distribution function value as label as sample, according to sample and label training reverse transmittance nerve network,
Neural network model is obtained, it is corresponding in modified bitmap bitmap data structure for the name data to be retrieved to be mapped to
Position；
The dynamic index unit is the dynamic index unit based on modified bitmap bitmap data structure, by traditional bitmap
Bitmap data structure is divided into several buckets, and the size of slot each in bucket is extended, obtain can dynamic label modified
Bitmap bitmap data structure, is denoted as D-bitmap, for storing address offset amount corresponding to the name data to be retrieved.
2. a kind of data retrieval method of index data structure neural network based according to claim 1, feature exist
In, be included in index data structure and be inserted into data, and to insertion data after index data structure carry out data retrieval；
Specific step is as follows:
Step 1: name data is inserted into the index data structure, one name data of every insertion, comprising the following steps:
Step 1-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 1-2: the neural computing of data mapping unit: above-mentioned nerve net will be inputted after name data fixed length processing
Network model calculation obtains real number value of the range between 0,1；
Step 1-3: neural computing result: being multiplied the slot sum of D-bitmap by the position mapping calculation of data mapping unit,
Show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 1-4: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and being rounded downwards, obtains
Bucket serial number where the position out；
Step 1-5: the maximum label of dynamic index unit is searched: the bucket serial number obtained by step 1-4 traverses in the bucket and owns
Slot searches existing maximum label；
Step 1-6: the slot internal label record of dynamic index unit: whether the existing maximum label that judgment step 1-5 is found reaches
To maximum value, if also unmarked arrive maximum value, the label that existing maximum label adds one is recorded in the location groove, otherwise, from
Deleting in queue takes the label deleted earliest as the label in the location groove；
Step 1-7: for the name data application storage unit: according in dynamic index unit the corresponding base address of bucket serial number with
The address offset amount that slot internal label represents is the name data application storage unit in data space；
Step 1-8: name data insertion operation terminates；
Step 2: being retrieved in index data structure to a name data, comprising the following steps:
Step 2-1: input name data: the name data being inserted into is input in above-mentioned index data structure；
Step 2-2: the neural computing of data mapping unit: above-mentioned nerve net will be inputted after name data fixed length processing
Network model calculation obtains real number value of the range between 0,1；
Step 2-3: neural computing result: being multiplied the slot sum of D-bitmap by the position mapping calculation of data mapping unit,
Show that the name data is mapped to the position on D-bitmap, i.e. the slot serial number of D-bitmap；
Step 2-4: the data existence judgement of dynamic index unit: judge whether the label at the position is 0, if label is not
It is 0, then the name data is present in the index data structure, and continues to execute step 2-5, completes retrieval；Otherwise, the title
Data are not present in the index data structure, that is, show to execute step 2-7 there is no search result；
Step 2-5: the bucket serial number of dynamic index unit calculates: with slot serial number divided by each barrel of slot number, and being rounded downwards, obtains
Bucket serial number where the position out；
Step 2-6: the bucket serial number and slot internal label where the position, i.e. data storage cell base output search result: are exported
Location and address offset amount of the name data in data space relative to base address；
Step 2-7: name data search operaqtion terminates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811152887.0A CN109271390B (en) | 2018-09-30 | 2018-09-30 | Index data structure based on neural network and data retrieval method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811152887.0A CN109271390B (en) | 2018-09-30 | 2018-09-30 | Index data structure based on neural network and data retrieval method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109271390A true CN109271390A (en) | 2019-01-25 |
| CN109271390B CN109271390B (en) | 2022-03-01 |
Family
ID=65196197
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811152887.0A Active CN109271390B (en) | 2018-09-30 | 2018-09-30 | Index data structure based on neural network and data retrieval method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109271390B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110096458A (en) * | 2019-04-02 | 2019-08-06 | 天津大学 | Name data network content storage pool data retrieval method neural network based |
| CN110109616A (en) * | 2019-04-02 | 2019-08-09 | 天津大学 | Name data network content storage pool data-erasure method neural network based |
| CN110138661A (en) * | 2019-04-02 | 2019-08-16 | 天津大学 | Name data network content storage pool neural network based |
| CN110196938A (en) * | 2019-04-02 | 2019-09-03 | 天津大学 | Named data network content storage pool data insertion method based on neural network |
| CN110474844A (en) * | 2019-06-28 | 2019-11-19 | 天津大学 | The learning-oriented index data structure of high-performance intelligent router and its training method |
| CN110851658A (en) * | 2019-10-12 | 2020-02-28 | 天津大学 | Tree index data structure, content storage pool, router and tree index method |
| CN113220683A (en) * | 2021-05-08 | 2021-08-06 | 天津大学 | Content router PIT structure supporting flooding attack detection and data retrieval method thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090271412A1 (en) * | 2008-04-29 | 2009-10-29 | Maxiscale, Inc. | Peer-to-Peer Redundant File Server System and Methods |
| CN104536958A (en) * | 2014-09-26 | 2015-04-22 | 杭州华为数字技术有限公司 | Composite index method and device |
| CN105975587A (en) * | 2016-05-05 | 2016-09-28 | 诸葛晴凤 | Method for organizing and accessing memory database index with high performance |
| US20170228407A1 (en) * | 2016-02-05 | 2017-08-10 | Amadeus S.A.S. | Database table index |
| CN107832343A (en) * | 2017-10-13 | 2018-03-23 | 天津大学 | A kind of method of MBF data directories structure based on bitmap to data quick-searching |
-
2018
- 2018-09-30 CN CN201811152887.0A patent/CN109271390B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090271412A1 (en) * | 2008-04-29 | 2009-10-29 | Maxiscale, Inc. | Peer-to-Peer Redundant File Server System and Methods |
| CN104536958A (en) * | 2014-09-26 | 2015-04-22 | 杭州华为数字技术有限公司 | Composite index method and device |
| US20170228407A1 (en) * | 2016-02-05 | 2017-08-10 | Amadeus S.A.S. | Database table index |
| CN105975587A (en) * | 2016-05-05 | 2016-09-28 | 诸葛晴凤 | Method for organizing and accessing memory database index with high performance |
| CN107832343A (en) * | 2017-10-13 | 2018-03-23 | 天津大学 | A kind of method of MBF data directories structure based on bitmap to data quick-searching |
Non-Patent Citations (1)
| Title |
|---|
| 刘哲益: "内存数据库存储结构及索引的研究与设计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110096458A (en) * | 2019-04-02 | 2019-08-06 | 天津大学 | Name data network content storage pool data retrieval method neural network based |
| CN110109616A (en) * | 2019-04-02 | 2019-08-09 | 天津大学 | Name data network content storage pool data-erasure method neural network based |
| CN110138661A (en) * | 2019-04-02 | 2019-08-16 | 天津大学 | Name data network content storage pool neural network based |
| CN110196938A (en) * | 2019-04-02 | 2019-09-03 | 天津大学 | Named data network content storage pool data insertion method based on neural network |
| CN110096458B (en) * | 2019-04-02 | 2022-03-01 | 天津大学 | Named data network content storage pool data retrieval method based on neural network |
| CN110474844A (en) * | 2019-06-28 | 2019-11-19 | 天津大学 | The learning-oriented index data structure of high-performance intelligent router and its training method |
| CN110474844B (en) * | 2019-06-28 | 2021-06-08 | 天津大学 | Training method and chip for learning type index data structure of high-performance intelligent router |
| CN110851658A (en) * | 2019-10-12 | 2020-02-28 | 天津大学 | Tree index data structure, content storage pool, router and tree index method |
| CN113220683A (en) * | 2021-05-08 | 2021-08-06 | 天津大学 | Content router PIT structure supporting flooding attack detection and data retrieval method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109271390B (en) | 2022-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109271390A (en) | Index data structure based on neural network and data retrieval method thereof | |
| CN105320775B (en) | The access method and device of data | |
| CN102521334B (en) | Data storage and query method based on classification characteristics and balanced binary tree | |
| Cambazoglu et al. | Scalability challenges in web search engines | |
| CN110472004B (en) | Method and system for multi-level cache management of scientific and technological information data | |
| CN102163226A (en) | Adjacent sorting repetition-reducing method based on Map-Reduce and segmentation | |
| CN110321325A (en) | File inode lookup method, terminal, server, system and storage medium | |
| CN109245879A (en) | A kind of double hash algorithms of storage and lookup IP address mapping relations | |
| CN103218443A (en) | Blogging webpage retrieval system and retrieval method | |
| CN109166615A (en) | A kind of medicine CT image storage and retrieval method of random forest Hash | |
| CN110460529A (en) | Content router FIB storage organization and its data processing method | |
| CN103714149A (en) | Self-adaptive incremental deep web data source discovery method | |
| CN105404677B (en) | A kind of search method based on tree structure | |
| CN110096458B (en) | Named data network content storage pool data retrieval method based on neural network | |
| Wu et al. | Cloud-based connected component algorithm | |
| CN104765767A (en) | Knowledge storage algorithm for intelligent learning | |
| CN107908773A (en) | The search method for focused web crawler that link based on precious deposits figure is combined with content | |
| CN110109616B (en) | Named data network content storage pool data deletion method based on neural network | |
| CN105426490B (en) | A kind of indexing means based on tree structure | |
| CN110138661A (en) | Name data network content storage pool neural network based | |
| CN107169020A (en) | A kind of orientation web retrieval method based on keyword | |
| CN110413724A (en) | A kind of data retrieval method and device | |
| CN107341227A (en) | Document handling method, server and computer-readable recording medium | |
| CN110196938B (en) | Named data network content storage pool data insertion method based on neural network | |
| CN103646056B (en) | Method for storing and extracting historical data based on characteristic value storage |
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 | ||
| CP02 | Change in the address of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: 300452 Binhai Industrial Research Institute Campus of Tianjin University, No. 48 Jialingjiang Road, Binhai New Area, Tianjin Patentee after: Tianjin University Address before: 300072 Tianjin City, Nankai District Wei Jin Road No. 92 Patentee before: Tianjin University |