CN113225227A - Network measurement method and device based on simplified diagram and considering simplicity and accuracy - Google Patents
Network measurement method and device based on simplified diagram and considering simplicity and accuracy Download PDFInfo
- Publication number
- CN113225227A CN113225227A CN202110479753.5A CN202110479753A CN113225227A CN 113225227 A CN113225227 A CN 113225227A CN 202110479753 A CN202110479753 A CN 202110479753A CN 113225227 A CN113225227 A CN 113225227A
- Authority
- CN
- China
- Prior art keywords
- counter
- network measurement
- layer
- counters
- diagram
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/50—Testing arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/04—Processing captured monitoring data, e.g. for logfile generation
Abstract
The invention relates to a network measurement method and device based on a simplified diagram, which have simplicity and accuracy. The invention adopts the idea of a classical sketch, does not additionally record auxiliary information such as stream ID and the like, and carries out shunting operation by an automatic balancing method; the diagram TowerScut of the invention has a plurality of layers, the counter of each layer is different in size, but the memory is the same, the small stream is recorded in the small counter and the large stream is recorded in the large counter by a proper updating strategy. The invention realizes the consideration of simplicity and accuracy of a diagram, provides a compression and serialization method to facilitate actual deployment, and can obtain higher accuracy in tasks such as obvious flow detection, obvious change detection, flow distribution estimation, flow entropy estimation, flow number estimation and the like.
Description
Technical Field
The invention relates to the field of network measurement, in particular to a network measurement method and device based on a simplified diagram and taking simplicity and accuracy into consideration.
Background
Network measurement is the basis of network operation and maintenance, and is applied to traffic engineering, anomaly detection, fault removal, network charging and the like. The simple algorithm is one of the commonly used measurement algorithms. The diagram records per-flow information in the network instead of per-packet information at the expense of a certain accuracy, thereby greatly saving the memory required in network measurement. Existing schematic algorithms can be divided into the following two categories: a classical schematic and a precise schematic. The classical diagrams, including CM, CU, Count, CMM, CSM, etc., consist of simple counters and are therefore easy to implement, deploy, compress, transmit and aggregate. However, they are less accurate because they do not match the actual network traffic. There is often a large tilt in actual network traffic: most flows are very small and a small number of large flows account for the majority of the flow.
The precise diagram is designed aiming at the network flow distribution on the basis of a classical diagram, has the advantages of high precision, high speed and the like, and inevitably brings complexity. Typical precision diagrams include UnivMon, Hashpipe, Beaucoup, FlowRadar, LossRadar, Elastic, Sonata, Marple, and the like. The precision diagram generally adopts a flow distribution idea to improve accuracy, and stores a big flow and a small flow by using different data structures. In order to distinguish the large stream and the small stream, the existing algorithm usually adopts a structure of a bucket to record the large stream, and auxiliary information such as a stream ID is required to be additionally recorded, so that the accuracy is improved, and meanwhile, the complexity is also improved. The data structure and operation of the fine sketch are more complex than the classical sketch, which prevents their application in real network environments, especially on hardware platforms such as FPGAs and P4-enabled programmable switches. For these hardware platforms adopting pipelines, operations have strict limitations, such as inability to perform division, loop, multiple accesses, copy long memory, etc., which also brings many difficulties to implementation of the diagram on these platforms, and thus most of the precise diagrams are difficult to deploy on the hardware platforms. Due to the complexity of sophisticated schematics, the industry still prefers the simplest, CM, schematic despite its poor accuracy.
Disclosure of Invention
In order to overcome the defect that the existing diagram algorithm is difficult to balance simplicity and accuracy, the invention provides a novel diagram named as TowerScut, and the automatic balancing mode is adopted to realize the consideration of simplicity and accuracy of the diagram. Meanwhile, the invention provides a TowerSketch compression and serialization method so as to facilitate actual deployment.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a network measurement method based on a simplified diagram and considering simplicity and accuracy comprises the following steps:
building a schema, the data structure of which comprises a plurality of layers, wherein layer i comprises wiA counter and a hash function hi(.), each counter contains δiA bit; each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiReduction;
in the insertion step of the network measurement process, network traffic packets are captured by the graph, automatically storing larger flows in larger counters and smaller flows in smaller counters.
Further, the number of layers d of the diagram is set to be 3 or more, and the size δ of the ith layer counteriIs set as 25-d+iBit, i-th layer counter number wiIs 2d-i×wd。
Further, the inserting step of the network measurement procedure comprises:
given a data packet with flow ID as e, d hash functions h are calculated first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]Then, the d mapped counters are increased by 1;
if a counter overflows, it is set to its maximum value, indicating that it has overflowed, and is not modified in subsequent operations.
Further, the inserting step of the network measurement process adopts a conservative updating mode to improve the accuracy, and the conservative updating mode comprises the following steps:
given a data packet with flow ID as e, d hash functions h are calculated first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]Then, the smallest counter that does not overflow among the d mapped counters is incremented by 1.
Further, the network measurement process further comprises a compression step, the compression step comprising:
given a compression ratio k, the compressed diagram still contains d layers, the i-th layer L 'after compression'iThe counter still contains deltaiA bit;
L′icomprises wiK counters of which the jth counter L'i[j]＝max{Li[j]，Li[j+wi/k]，...，Li[j+(k-1)wi/k]In which L isi[j]Represents the jth counter in the original diagram;
compressed hash function h'i(.)＝hi(.)mod(wi/k)。
Further, the network measurement process further comprises a serialization step, the serialization step comprising:
firstly, outputting parameters of the diagram, including the number d of layers and the size delta of the ith layer counteriNumber wiAnd a hash function hi(.). And then outputting counter arrays of each layer of the diagram layer by layer.
Further, the network measurement process further comprises a query step, the query step comprising:
given a query stream with an ID of e, d hash functions h are first computed1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]And then returns to the minimum value where the counter does not overflow.
A schematic-based network measurement apparatus employing the above method for simplicity and accuracy, comprising:
a schema building module for building a schema, the data structure of the schema comprising a plurality of layers, wherein layer i comprises wiA counter and a hash function hi(.), each counter contains δiA bit; each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiReduction;
and the network measurement module is used for carrying out network measurement, and in the insertion step of the network measurement process, the network flow data packets are captured through the diagram, larger flows are automatically stored in a larger counter, and smaller flows are stored in a smaller counter.
The invention has the advantages that the invention can give consideration to the simplicity and the accuracy of a diagram algorithm and can simultaneously support various tasks. The greatest innovation of the invention is the automatic balancing method. TowerSketch follows the idea of the classic diagram, and does not additionally record auxiliary information such as stream ID, and the shunting operation is carried out by an automatic balancing method. The specific principle is as follows: to increase memory usage efficiency, it is desirable to automatically store larger streams in larger counters and smaller streams in smaller counters. Since the flow distribution is highly skewed, with very few large flows and very many small flows, the ideal diagram requires a large number of small counters, and a small number of large counters. The automatic balancing method of the invention comprises the following steps: TowerSketch comprises multiple layers, and the counter of each layer is different in size but the memory is the same. The recording of small flows in small counters and large flows in large counters is achieved by appropriate update strategies.
In the flow estimation, the average relative error of the invention can reach 1/6.8 of that of ElasticSketch, which is 1/27 of CMSketch. Meanwhile, higher accuracy can be obtained in tasks such as significant flow detection, significant change detection, flow distribution estimation, flow entropy estimation, flow number estimation and the like. At the same time, the invention retains the simplicity of the classical schematic and can be implemented on FPGAs and programmable switches.
Drawings
FIG. 1 is an example of a data structure of TowerSketch;
FIG. 2 is an example of an insertion flow for TowerSketch;
FIG. 3 is a query flow example of TowerSketch;
FIG. 4 is an example of a compression flow for TowerSketch;
FIG. 5 is an example of an insertion flow for TowerSketch with conservative updates;
FIG. 6 is a TowerSketch-based flow measurement flow scheme;
fig. 7 is a flow analysis flow based on the TowerSketch.
Fig. 8 is a graph of experimental results of flow estimation, in which (a) the graph is AAE of flow estimation and (b) the graph is ARE of flow estimation.
Fig. 9 is a graph of experimental results of a large flow assay, wherein (a) the graph is the F1 fraction of the large flow assay and (b) the graph is the ARE of the large flow assay.
FIG. 10 is a graph showing the results of an experiment for detecting a flux mutation.
Fig. 11 is a graph of experimental results of entropy estimation.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following examples in the accompanying drawings. It should be understood that the specific examples described herein are intended to be illustrative only and are not intended to be limiting.
The network measurement method based on the simplified diagram and considering simplicity and accuracy comprises the following steps:
(1) the data structure of TowerSketch contains d layers. L for i-th layeriIs represented by comprising wiA counter and a hash function hi(.)(1≤hi≤wi)，h1，h2，...，hdIndependent in pairs. L isiEach counter of (1) contains deltaiBit, for its j-th counter Li[j]And (4) showing.
(2) The TowerSketch realizes the matching of the counter bit number and the flow rate by the following automatic balancing mode: each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiAnd (4) reducing.
(3) Preferably, the number of layers of TowerSketch is set to 5, and the i-th layer counter size δiIs set as 2iBit, i-th layer counter number wiIs 2d-i×wd。wdIs the number of counters at the d-th level. Note that the TowerSketch number of layers is related to the accuracy of the diagram, and the number of layers d should be 3 or more to ensure accuracy.
(4) When inserting, a packet with flow ID e is given. TowerSketch calculates d hash functions h first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]. After that, TowerSketch increments the d mapped counters by 1 each. Note that if a counter overflows, TowerSketch sets the counter to its maximum value, indicating that it has overflowed, and does not modify this counter in subsequent operations.
(5) When inquiring, given inquiry flow, its ID is e. TowerSketch calculates d hash functions h first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]. After that, TowerSketch returns the minimum value of the counter where it did not overflow.
(6) At the time of compression, a compression ratio k is given. The compressed TowerSketch still contains d-layer, i-th layer L'iThe counter still contains deltaiA bit. Except that L'iComprises wiK counters of which the jth counter L'i[j]＝max{Li[j]，Li[j+wi/k]，...，Li[j+(k-1)wi/k]Record the information of k counters in the original TowerSketch. Hash function h'i(.)＝hi(.)mod(wi/k)。
(7) During serialization, firstly outputting parameters of TowerSketch, including the number d of layers and the size delta of the ith layer counteriNumber wiAnd a hash function hi(.). Then, the counter arrays of the TowerSketch layers are output layer by layer.
(8) The TowerSketch can adopt a conservative updating method to perform an insertion process so as to obtain higher accuracy. During insertion, TowerSketch instead maps d to the counters, and the smallest counter that does not overflow is incremented by 1 each.
The "insertion" in step (4) refers to inserting a packet with a given flow ID into the diagram during the network measurement process, where the flow ID usually adopts a source IP address, < source IP address, destination IP address >, or TCP/IP quintuple (< source IP address, source port, destination IP address, destination port, transport layer protocol >), etc. The "query" in step (5) refers to querying the traffic of a given flow in the network measurement process, where the traffic is defined as the number of packets in one flow. The compression in the step (6) refers to compressing the size of the simplified diagram in the network measurement process so as to reduce the occupation of the memory and the transmission bandwidth in the measurement process. The step (7) of "serialization" refers to the step of converting the diagram into a byte sequence in the network measurement process so as to facilitate storage and transmission.
Fig. 1 is a data structure example of the towersketh. As shown in FIG. 1, TowerSketch contains 5 levels, counter size δi2bit, 4bit, 8bit, 16bit and 32bit in sequence, and the number w of the counter i32, 16, 8, 4 and 2 in sequence, and the hash function is h in sequence1，h2，h3，h4，h5。
Fig. 2 is an example of an insertion flow of towersketh. As shown in fig. 2, towersketh first computes d hash functions: h is1(e)＝25，h2(e)＝9，h3(e)＝3，h4(e)＝4，h5(e) 1 is ═ 1; locate d mapped counters: l is1[h1(e)]＝3，L2[h2(e)]＝14，L3[h3(e)]＝17，L4[h4(e)]＝14，L5[h5(e)]＝130。L1[h1(e)]Has overflowed. After that, TowerSketch increments the counter that did not overflow by 1: l is2[h2(e)]＝14+1＝15，L3[h3(e)]＝17+1＝18，L4[h4(e)]＝14+1＝15，L5[h5(e)]130+1 131. After insertion, L2[h2(e)]Marked as overflowed.
FIG. 3 is an example of a query flow for TowerSketch. As shown in fig. 3, towersketh first computes d hash functions: h is1(e)＝25，h2(e)＝9，h3(e)＝3，h4(e)＝4，h5(e) 1 is ═ 1; locating d mapped counters L1[h1(e)]＝3，L2[h2(e)]＝15，L3[h3(e)]＝18，L4[h4(e)]＝15，L5[h5(e)]＝131。L1[h1(e)]、L2[h2(e)]Has overflowed. After that, TowerSketch returns the minimum value of the overflow not counter: min {18, 15, 131} -, 15.
Fig. 4 is a compression flow example of towersketh. As shown in FIG. 4, before compression, TowerSketch contains 5 levels, and the counter size δi2bit, 4bit, 8bit, 16bit and 32bit in sequence, and the number w of the counter i32, 16, 8, 4 and 2 in sequence, and the hash function is h in sequence1，h2，h3，h4，h5. After compression, TowerSketch still contains 5 layers, counter size δiSequentially comprises 2 bits, 4 bits, 8 bits, 16 bits and 32 bits. Number of counters wiAre sequentially reduced to 16, 8, 4, 2 and 1, and the hash function is sequentially h'1(.)＝h1(.)mod 16，h′2(.)＝h2(.)mod 8，h′3(.)＝h3(.)mod 4，h′4(.)＝h4(.)mod 2，h′5(.)＝h5(.) mod 1. The partial counter compressed values are as follows:
L′5[0]＝max{L5[0]，L5[1]}＝max{131，127}＝131，L′4[0]＝max{L4[0]，L4[2]}＝max{8，0}＝8，L′4[1]＝max{L4[1]，L4[3]}＝max{40，18}＝40。
fig. 5 is an example of an insertion flow for topersketh with conservative updates. As shown in fig. 5, towersketh first computes d hash functions: h is1(e)＝25，h2(e)＝9，h3(e)＝3，h4(e)＝4，h5(e) 1 is ═ 1; locate d mapped counters: l is1[h1(e)]＝3，L2[h2(e)]＝14，L3[h3(e)]＝17，L4[h4(e)]＝14，L5[h5(e)]＝130。L1[h1(e)]Has overflowed, L2[h2(e)]、L4[h4(e)]The smallest counter that does not overflow. Thereafter, the minimum counters that TowerSketch does not overflow are each incremented by 1: l is2[h2(e)]＝14+1＝15，L4[h4(e)]14+1 15. After insertion, L2[h2(e)]Marked as overflowed.
Fig. 6 is a flow measurement process based on the topersketh, which includes the processes of data insertion, compression, and serialization.
Fig. 7 is a flow of traffic analysis based on the TowerSketch, in which a query process of data is included.
In the experiment 5 schematic diagrams were compared: TowerSketch (abbreviated as Tower), TowerSketch with conservative updates (abbreviated as Tower + CU), CM, CU and Elastic. The following 4 tasks were tested in the experiment: flow estimation, large flow detection, flow sudden change detection, and entropy estimation.
1. Experimental device
The experimental environment is as follows: 12-core CPU Server (
Xeon E5-2620, 64G, GTX 1080). Each core contains a 32KB L1 cache, a 256KB L2 cache, and a 15MB L3 cache.
Data set: the data set is the internet data set from CAIDA and is cropped. Each data set contains a 5s monitoring time window, 170K stream, 2.3M packets.
Indexes are as follows:
AAE (mean absolute error):
wherein n is the number of streams, fiAnd
respectively true and estimated values of the flow.
ARE (average relative error):
f1 score:
where PR (precision rate) indicates how many of the samples predicted to be positive are true positive samples and RR (recall rate) indicates how many of the positive samples are predicted to be correct.
RE (relative error):
wherein val and
true and estimated values, respectively.
2. Results of the experiment
And (3) flow estimation: as shown in FIG. 8, the accuracy of Tower and Tower + CU is at least 6.8 times, 29 times higher than CM, CU, Elastic. When the memory is 900KB, the AAE of the Tower and the Tower + CU respectively reaches 0.30 and 0.021, and the AAE of the CM, the CU and the Elastic respectively reaches 2.6, 1.3 and 0.58; towerAnd TowerThe AREs for + CU ARE 0.060 and 0.014 respectively, while those for CM, CU, Elastic ARE 1.6, 1.0, 0.40.
And (3) large flow detection: as shown in FIG. 9, the F1 score was higher for Tower + CU and ARE was at least 13 times lower compared to CM, CU, Elastic. When the memory is 150KB, the F1 score of the Tower + CU is 0.998, and the F1 scores of the CM, the CU and the Elastic are 0.909, 0.996 and 0.990; the ARE for Tower + CU is 0.00083, while the ARE for CM, CU, Elastic is 0.042, 0.0051, 0.0063.
And (3) detecting flow mutation: as shown in FIG. 10, the F1 score is higher for Tower + CU compared to CM, CU, Elastic. When the memory is 150KB, the F1 score of the Tower + CU is 0.99, and the F1 scores of the CM, CU and Elastic are 0.91, 0.94 and 0.97.
Entropy estimation: as shown in FIG. 11, the RE of Tower + CU is at least 12.8 times lower than that of CM, CU, Elastic. When the memory is 900KB, the RE of Tower + CU is 0.00024, and the RE of CM, CU, Elastic is 0.0092, 0.0072, 0.0031.
Based on the same inventive concept, another embodiment of the present invention provides a network measurement apparatus based on a simplified diagram and using the method of the present invention, which has simplicity and accuracy, and comprises:
a schema building module for building a schema, the data structure of the schema comprising a plurality of layers, wherein layer i comprises wiA counter and a hash function hi(.), each counter contains δiA bit; each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiReduction;
and the network measurement module is used for carrying out network measurement, and in the insertion step of the network measurement process, the network flow data packets are captured through the diagram, larger flows are automatically stored in a larger counter, and smaller flows are stored in a smaller counter.
Based on the same inventive concept, another embodiment of the present invention provides an electronic device (computer, server, smartphone, etc.) comprising a memory storing a computer program configured to be executed by the processor, and a processor, the computer program comprising instructions for performing the steps of the inventive method.
Based on the same inventive concept, another embodiment of the present invention provides a computer-readable storage medium (e.g., ROM/RAM, magnetic disk, optical disk) storing a computer program, which when executed by a computer, performs the steps of the inventive method.
The particular embodiments of the present invention disclosed above are illustrative only and are not intended to be limiting, since various alternatives, modifications, and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The invention should not be limited to the disclosure of the embodiments in the present specification, but the scope of the invention is defined by the appended claims.
Claims (10)
1. A network measurement method based on a simplified diagram and considering simplicity and accuracy is characterized by comprising the following steps:
building a schema, the data structure of which comprises a plurality of layers, wherein layer i comprises wiA counter and a hash function hi(.), each counter contains δiA bit; each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiReduction;
in the insertion step of the network measurement process, network traffic packets are captured by the graph, automatically storing larger flows in larger counters and smaller flows in smaller counters.
2. The method according to claim 1, wherein the number of layers d of the diagram is set to 3 or more, and the ith layer counter size δ is set to be equal to or larger thaniIs set as 25-d+iBit, i-th layer counter number wiIs 2d-i×wd。
3. The method of claim 1, wherein the step of inserting the network measurement procedure comprises:
given a data packet with flow ID as e, d hash functions h are calculated first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]Then, the d mapped counters are increased by 1;
if a counter overflows, it is set to its maximum value, indicating that it has overflowed, and is not modified in subsequent operations.
4. The method of claim 1, wherein the step of inserting the network measurement procedure employs a conservative update to improve accuracy, the conservative update comprising:
given a data packet with flow ID as e, d hash functions h are calculated first1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]Then, the smallest counter that does not overflow among the d mapped counters is incremented by 1.
5. The method of claim 1, wherein the network measurement process further comprises a compression step, the compression step comprising:
given a compression ratio k, the compressed diagram still contains d layers, the i-th layer L 'after compression'iThe counter still contains deltaiA bit;
L′icomprises wiK counters of which the jth counter L'i[j]＝max{Li[j]，Li[j+wi/k]，...，Li[j+(k-1)wi/k]In which L isi[j]Represents the jth counter in the original diagram;
compressed hash function h'i(.)＝hi(.)mod(wi/k)。
6. The method of claim 1, wherein the network measurement procedure further comprises a serialization step, the serialization step comprising:
firstly, outputting parameters of the diagram, including the number d of layers and the size delta of the ith layer counteriNumber wiAnd a hash function hi(.). And then outputting counter arrays of each layer of the diagram layer by layer.
7. The method of claim 1, wherein the network measurement procedure further comprises a query step, the query step comprising:
given a query stream with an ID of e, d hash functions h are first computed1(e)，h2(e)，...，hd(e) To locate d mapped counters L1[h1(e)]，L2[h2(e)]，...，Ld[hd(e)]And then returns to the minimum value where the counter does not overflow.
8. A simplified schematic-based network measurement device using the method of any one of claims 1-7 for simplicity and accuracy, comprising:
a schema building module for building a schema, the data structure of the schema comprising a plurality of layers, wherein layer i comprises wiA counter and a hash function hi(.), each counter contains δiA bit; each layer containing spaces of the same size, delta, as the number of layers increasesiIncrease, wiReduction;
and the network measurement module is used for carrying out network measurement, and in the insertion step of the network measurement process, the network flow data packets are captured through the diagram, larger flows are automatically stored in a larger counter, and smaller flows are stored in a smaller counter.
9. An electronic apparatus, comprising a memory and a processor, the memory storing a computer program configured to be executed by the processor, the computer program comprising instructions for performing the method of any of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a computer, implements the method of any one of claims 1 to 7.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2021103194916 | 2021-03-25 | ||
| CN202110319491 | 2021-03-25 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113225227A true CN113225227A (en) | 2021-08-06 |
| CN113225227B CN113225227B (en) | 2023-01-17 |
Family
ID=77090419
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110479753.5A Active CN113225227B (en) | 2021-03-25 | 2021-04-30 | Network measurement method and device based on simplified diagram and considering simplicity and accuracy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113225227B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113992541A (en) * | 2021-09-11 | 2022-01-28 | 西安电子科技大学 | Network flow measuring method, system, computer equipment, storage medium and application |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180302152A1 (en) * | 2014-10-07 | 2018-10-18 | Sedonasys Systems Ltd | Systems and methods for managing multilayer communication networks |
| CN110830322A (en) * | 2019-09-16 | 2020-02-21 | 北京大学 | Network flow measuring method and system based on probability measurement data structure Sketch with approximate zero error |
| CN111262756A (en) * | 2020-01-20 | 2020-06-09 | 长沙理工大学 | High-speed network elephant flow accurate measurement method and structure |
| CN111782700A (en) * | 2020-08-05 | 2020-10-16 | 中国人民解放军国防科技大学 | Data stream frequency estimation method, system and medium based on double-layer structure |
| CN112416950A (en) * | 2021-01-25 | 2021-02-26 | 中国人民解放军国防科技大学 | Design method and device of three-dimensional sketch structure |
-
2021
- 2021-04-30 CN CN202110479753.5A patent/CN113225227B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180302152A1 (en) * | 2014-10-07 | 2018-10-18 | Sedonasys Systems Ltd | Systems and methods for managing multilayer communication networks |
| CN110830322A (en) * | 2019-09-16 | 2020-02-21 | 北京大学 | Network flow measuring method and system based on probability measurement data structure Sketch with approximate zero error |
| CN111262756A (en) * | 2020-01-20 | 2020-06-09 | 长沙理工大学 | High-speed network elephant flow accurate measurement method and structure |
| CN111782700A (en) * | 2020-08-05 | 2020-10-16 | 中国人民解放军国防科技大学 | Data stream frequency estimation method, system and medium based on double-layer structure |
| CN112416950A (en) * | 2021-01-25 | 2021-02-26 | 中国人民解放军国防科技大学 | Design method and device of three-dimensional sketch structure |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113992541A (en) * | 2021-09-11 | 2022-01-28 | 西安电子科技大学 | Network flow measuring method, system, computer equipment, storage medium and application |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113225227B (en) | 2023-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9819637B2 (en) | Efficient longest prefix matching techniques for network devices | |
| US11558302B2 (en) | Data transmission method and apparatus | |
| US10230626B2 (en) | Service specific route selection in communication networks | |
| US9507756B2 (en) | Space efficient counters in network devices | |
| Lei et al. | An entropy-based probabilistic forwarding strategy in named data networking | |
| WO2019184640A1 (en) | Indicator determination method and related device thereto | |
| CN107317879A (en) | The distribution method and system of a kind of user's request | |
| US7224670B2 (en) | Flow control in computer networks | |
| CN107689919B (en) | Dynamic adjustment weight fuzzy routing method for SDN network | |
| CN110730469B (en) | Method for predicting bandwidth based on extended Kalman wireless network and congestion control thereof | |
| CN113225227B (en) | Network measurement method and device based on simplified diagram and considering simplicity and accuracy | |
| CN113452676B (en) | Detector distribution method and Internet of things detection system | |
| JP2002223240A (en) | Method for approximating weighted random early detection buffer admittance algorithm | |
| US9992081B2 (en) | Scalable generation of inter-autonomous system traffic relations | |
| WO2019052307A1 (en) | Adaptive code rate algorithm optimization system and method, and terminal | |
| CN116545936A (en) | Congestion control method, system, device, communication equipment and storage medium | |
| CN108768738B (en) | Rapid self-adaptive network bandwidth detection method based on CDN network technology | |
| KR100567320B1 (en) | Flow generation method for Internet traffic measurement | |
| CN116055416B (en) | Method and device for adjusting transmission rate in long-distance communication network scene | |
| US10320874B2 (en) | Site-based server selection | |
| CN115277877B (en) | Protocol session distribution method compatible with multiple network communications | |
| CN114884893B (en) | Forwarding and control definable cooperative traffic scheduling method and system | |
| CN108566335B (en) | Network topology generation method based on NetFlow | |
| CN114448872B (en) | Path selection method and system in CDN | |
| CN110868332B (en) | SDN-based network-level flow measurement method |
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 |