CN112804125A - Named data network congestion control method based on fuzzy comprehensive evaluation algorithm - Google Patents
Named data network congestion control method based on fuzzy comprehensive evaluation algorithm Download PDFInfo
- Publication number
- CN112804125A CN112804125A CN202110181006.3A CN202110181006A CN112804125A CN 112804125 A CN112804125 A CN 112804125A CN 202110181006 A CN202110181006 A CN 202110181006A CN 112804125 A CN112804125 A CN 112804125A
- Authority
- CN
- China
- Prior art keywords
- network
- interest
- packet
- fuzzy
- time
- 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
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0882—Utilisation of link capacity
-
- 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
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0823—Errors, e.g. transmission errors
- H04L43/0829—Packet loss
-
- 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
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0852—Delays
- H04L43/0864—Round trip delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/11—Identifying congestion
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A named data network congestion control method based on a fuzzy comprehensive evaluation algorithm is characterized in that four congestion states of network idle, network normal, network busy and network congestion are designed, an accurate network state is obtained by obtaining two network parameters of an average cache queue length and a round-trip time ratio of an interest packet and adopting the fuzzy comprehensive evaluation algorithm, a network state signal is sent to a downstream node by utilizing the existing NACK feedback mechanism, and the downstream node adjusts the sending rate of the interest packet according to the network state signal to ensure the stability of the network.
Description
Technical Field
The invention relates to the technical field of network congestion control, in particular to a named data network congestion control method based on a fuzzy comprehensive evaluation algorithm.
Background
In order to guarantee the stability and the service quality of the NDN network, it is necessary to research the congestion control of the NDN network. The system structure of the NDN network is based on receiving end driving, has the characteristics of multiple sources and multiple paths, and cannot simply adopt a method that a TCP/IP network judges whether the network is congested or not according to time delay. However, the NDN network supports flow balance, that is, one interest packet returns one data packet at most, and according to this characteristic, a scholars proposes to adjust the sending rate of the interest packet by judging whether a network link is congested based on the length of an interest packet buffer queue, increase the sending rate of the interest packet when the length of the buffer queue is small, and decrease the sending rate of the interest packet when the length of the buffer queue is large, but the network environment is uncontrollable, and under the condition that the network performance is relatively good, the method can improve the network throughput to a certain extent, but when the network performance is poor and the length of the buffer queue is small, the method rather aggravates the network congestion.
The end node interest control protocol maintains historical round trip delay for each interest flow according to a method in a TCP/IP network to calculate RTO so as to judge whether the network is congested or not, and then an AIMD controller is arranged at an interest packet requesting party and adjusts the size of a sending window according to the congestion condition of the current network so as to be suitable for the network environment. The method is based on the assumption of single-source congestion, the NDN network is multi-source, data content can be in multiple copies in the network, and therefore the round-trip time obtained through measurement is often inaccurate, and therefore the judgment of the network state is not feasible only according to the round-trip time.
In summary, the above methods have obvious disadvantages in judging the network status.
Disclosure of Invention
In order to solve the technical problems, the invention provides a named data network congestion control method based on a fuzzy comprehensive evaluation algorithm, which is characterized by designing four congestion states of network idle, network normal, network busy and network congestion, obtaining two network parameters of average cache queue length and round-trip time ratio of an interest packet, adopting the fuzzy comprehensive evaluation algorithm to obtain an accurate network state, sending a network state signal to a downstream node by utilizing the conventional NACK feedback mechanism, and adjusting the sending rate of the interest packet by the downstream node according to the network state signal to ensure the stability of the network.
In order to realize the technical purpose, the adopted technical scheme is as follows: a named data network congestion control method based on a fuzzy comprehensive evaluation algorithm comprises the following steps:
step 4.1, defining a factor set U and a decision set V in fuzzy evaluation, wherein the average cache queue length and the round-trip time ratio of the interest packet are the parameters of the factor set U; determine 4 fuzzy subsets of the factor set U: F. n, C and M, μ f (U), μ n (U), μ c (U) and μ M (U) correspond to membership functions of F, N, C, M, F, N, C, M as a mapping of factor set U to decision set V, decision set V ═ V (V ═ V), respectively1,v2,v3,v4) Describing a current congestion status of the network;
and 4.2, expressing the Fuzzy relation from the factor set U to the decision set V by using a Fuzzy matrix R, distributing corresponding weights according to different influences of the average buffer queue length factor and the round-trip time ratio factor of the interest packets on the decision set, marking the weights as Fuzzy vectors W, and obtaining a Fuzzy subset of the decision set V according to the Fuzzy matrix R and the Fuzzy vectors Wb1,b2,b3,b4The membership degree of the decision set V;
4.3, selecting the maximum membership degree and judging which congestion state the network is in;
and 5, according to the congestion state of the network, the downstream routing node adjusts the sending rate of the interest packet and the middle routing node selects other proper ports to control the network congestion of the named data.
Interest package pingAverage buffer queue length AQLfinalIs calculated by the formula
Wherein, QueueLengthcurrent(j)Buffer queue length for current port, i +1 represents the number of times the packet of interest buffer queue length is measured, EjRepresenting QueueLengthcurrent(j)The weight of (c).
Weight EjIs calculated by
Ej+1=aEj+b
Wherein a and b represent parameters with linearly increasing weights.
The method for calculating the round-trip time ratio RTTR is
INT=RTO-(Interesttime+Datatime)
RT is the actual time when the intermediate routing node receives the data packet sent back by the content node, INT is the timeout time from the intermediate routing node to the content node which can meet the request, RTO is the timeout retransmission time, InteresttimeFor the transmission time of interest packets from sender to intermediate node, DatatimeThe time for the data packet to reach the interested requesting end from the intermediate routing node.
Time Data of Data packet from intermediate routing node to interest request endtimeIs calculated by
Wherein, the packageDataSizeIndicates packet size and packageInterestIndicates the size of the Interest packet, InteresttimeIndicating the packet of interest transmission time.
The calculation method of the time-out retransmission time RTO comprises the following steps
Wherein:is the average round trip time, RTT, of the most recently received packetnIs the round trip time of the last received packet,is a constant from 0 to 1 and is,is the average round trip time of the received packet before the current packet reception, f is the constant variation adapted to the RTT, and σ is the estimated value of the RTT standard deviation.
The blur matrix R is represented as
μF(u1)、μN(u1)、μC(u1) And μ M (u)1) Membership functions corresponding to four fuzzy subsets respectively representing average buffer queue length of interest packets, mu F (u2)、μN(u2)、μC(u2) And μ M (u)2) And the membership functions correspond to four fuzzy subsets respectively representing the round trip time ratio.
When the network is idle, the downstream routing nodes adopt a multiplicative increasing algorithm to adjust the sending rate of the interest packets, when the network is normal, the downstream routing nodes adopt an additive increasing algorithm to adjust the sending rate of the interest packets, when the network is busy, the downstream intermediate routing nodes adopt a multiplicative decreasing algorithm to adjust the sending rate of the interest packets, and when the network is congested, other proper ports are selected to reduce the load of the current link.
The invention has the beneficial effects that: the network state is judged by utilizing the fuzzy comprehensive judgment algorithm, and the network state judgment method is more accurate and reliable compared with the prior art in which the network state is judged by adopting a single factor. Through experimental verification, the invention reduces the occurrence of network congestion and improves the network transmission efficiency.
Drawings
FIG. 1 is a link state diagram of membership functions for a set of factors U;
FIG. 2 is a diagram of a dumbbell topology;
FIG. 3 is a multipath topology;
FIG. 4 is a graph of the average delay in a dumbbell topology as a function of the number of interest requests;
FIG. 5 is a graph of the average delay in a multi-path topology as a function of the number of interest requests;
fig. 6 is a graph showing the variation of packet loss rate with the number of interest requests in the dumbbell topology;
fig. 7 is a graph showing the variation of packet loss rate with the number of interest requests in a multi-path topology;
FIG. 8 is a graph of throughput versus packet-of-interest transmission time in a dumbbell topology;
fig. 9 is a graph of throughput as a function of packet of interest transmission time in a multi-path topology.
Detailed Description
A named data network congestion control method based on a fuzzy comprehensive evaluation algorithm comprises the following steps: step 1, designing a named data network to have four congestion states of network idle, normal network, busy network and network congestion;
there are two message types in the Named Data Network (NDN), namely an interest packet and a data packet, and when a consumer requests content, the interest packet containing the name prefix of the content is sent. The NDN forwarding model mainly has three types of data structures, which are a forwarding information table, a content storage table, and a pending request table. The forwarding information table of the NDN router is similar to the IP router in function, except that the address prefix in the IP is replaced by the stored content name prefix, and meanwhile, for one content name prefix, a plurality of forwarding interfaces may exist in the forwarding information table; the content storage table stores the cache content of the routing node; the pending request table records the name prefix of the interest packet without obtaining the matching data, the information of the arrival interface and the forwarding interface thereof, and the like, and when the data packet is returned, the data packet can be returned along the opposite path of the interest packet through the information in the pending request table. In addition, each NDN router has a policy module for making forwarding decisions for each interest packet.
the internal mechanism of the NDN network is that one interest packet corresponds to one data packet, so that the interest packet buffer queue is adopted to judge how many data packets exist in the current network link, and the network state is judged according to the network link capacity. If a large amount of request data exist at a certain moment, no request data exist for a period of time later, the network cannot be determined to be in a congestion state, or the buffer queue is idle at the current moment, and the network cannot be considered to be in an idle state if a large amount of request data exist for a period of time later. Therefore, in order to avoid the influence of burst traffic on the judgment of the network state, a measurement period T is averagely divided into i sections, the interest packet buffer queue length is monitored once every T/i time, i +1 times are monitored in a period in total, and the average interest packet buffer queue length AQL in a period is obtained through the measured values. The calculation formula of AQL is:
wherein QueueLengthcurrentThe queue length is buffered for the current port.
In order to reflect the variation trend of the network congestion degree, the AQL is optimized:
each monitored interest packet buffer queue length corresponds to a weight value, and the weight value is monotonically increased in a period. Suppose that the current port is detected at time mThe length of the cache queue of the instant interest packet queue is QueueLengthcurrent(m)And the length of the buffer queue of the instant interest packet detected at the moment k is QueueLengthcurrent(k)And m is less than k, QueueLengthcurrent(m)Weight E ofmLess than QueueLengthcurrent(k)Weight E ofk. The design of utilizing the weight increment shows that the newly detected interest packet queue length is more important than the historical interest packet queue length and can better reflect the congestion degree of the current network. In order to enable the weight to be smoothly increased, the algorithm adopts a linear increasing mode, and the calculation formula of the weight is as follows
Ej+1=aEj+b
Wherein i +1 represents the number of times of measuring the length of the packet buffer queue of interest, EjRepresenting QueueLengthcurrent(j)A, b represent parameters of linear increase of the weights.
After optimization, the final average buffer queue length AQL of the interest packetfinalThe calculation formula of (a) is as follows:
the round trip delay reflects the network state to a certain extent, and the round trip delay is treated as a factor of a fuzzy comprehensive evaluation algorithm.
The formula for calculating the time-out retransmission time RTO is as follows:
is the most recently received packetRound trip time, RTTnIs the round trip time of the last received packet,is a constant from 0 to 1 and is,is the average round trip time of the received packet before the current packet is received. The timeout retransmission time RTO is as follows:
f is a constant change to accommodate the RTT, preventing frequent premature timeouts, and σ is an estimate of the standard deviation of RTT.
The calculated RTO is stored in the interest packet sending node as the time-out retransmission time of the interest packet request, and what is done in this document is that a time is stored in the intermediate routing node as a condition for judging network congestion. Therefore, the intermediate routing node needs to recalculate the RTO, and the calculation process is as follows:
because one interest packet corresponds to one data packet and the data packet is returned according to the original path, and data is transmitted by bit stream on a network link, the transmission time ratio of the data packet to the interest packet can be calculated by the ratio of the size of the data packet to the size of the interest packet.
packageDataSizeIndicates packet size and packageInterestIndicates the size of the Interest packet, InteresttimeRepresenting the transmission time of interest packets, DatatimeIndicating the packet return time.
The transmission time of the Interest packet from the sending end to the intermediate node is obtained as Interest through the timestamptimeSince the return of the data packet is a special mechanism for returning according to the request route, the time from the intermediate routing node to the interested request end isDatatimeThe time-out time INT from the intermediate routing node to the content node that can satisfy the request is RTO- (Interest)time+Datatime) And the intermediate routing node maintains the time value. When the intermediate node forwards the interest packet, the time is counted again, the actual time RT when the intermediate node receives the data packet sent back by the content node is calculated, the ratio of the RT to the INT is compared to judge which state the network is in, and a corresponding solution is adopted.
the selection of the membership function is the key for fuzzy comprehensive evaluation and analysis, and the membership function is a trapezoidal function, a trigonometric function, a Gaussian function and the like. Each factor set having a different membership function corresponding to a different congestion level. After reasoning analysis, a mode of combining a trapezoidal function and a Gaussian function is adopted as a membership function on the domain U. As shown in fig. 1.
The abscissa here represents two meanings. When the abscissa represents the average buffer queue length, the formula is shownCalculating to obtain AQLfinalDetermining AQLfinalAnd (4) the value is in the area of the abscissa, and the membership degree of the network state is calculated by using the corresponding membership function. When the abscissa represents the ratio of the predicted round trip time to the actual round trip time. According to the formulaAnd calculating to obtain RTTR, determining that the RTTR value is in the area of the abscissa, and calculating the membership degree of the network state by using the corresponding membership function.
The network state is complicated and changeable, and the state of the network cannot be simply judged according to a certain network factor, so that the fuzzy judgment algorithm in fuzzy mathematics is utilized to make an overall evaluation on the network limited by various factors, and appropriate measures are taken so as to improve the network performance.
Step 4.1, defining a factor set U and a decision set V in fuzzy evaluation, wherein the average cache queue length and the round-trip time ratio of the interest packet are the parameters of the factor set U; determine 4 fuzzy subsets of the factor set U: F. n, C and M, μ f (U), μ n (U), μ c (U) and μ M (U) correspond to membership functions of F, N, C, M, F, N, C, M as a mapping of factor set U to decision set V, decision set V ═ V (V ═ V), respectively1,v2,v3,v4) Describing a current congestion status of the network;
and 4.2, expressing the Fuzzy relation from the factor set U to the decision set V by using a Fuzzy matrix R, distributing corresponding weights according to different influences of the average buffer queue length factor and the round-trip time ratio factor of the interest packets on the decision set, marking the weights as Fuzzy vectors W, and obtaining a Fuzzy subset of the decision set V according to the Fuzzy matrix R and the Fuzzy vectors Wb1,b2,b3,b4Is the degree of membership of decision set V.
According to the definition of Zadeh on the fuzzy subset, establishing a factor set U and a decision set V: and selecting the average buffer queue length and round-trip time ratio of the interest packet as the parameter of the factor set for fuzzy comprehensive judgment.
Determine 4 fuzzy subsets of the factor set U: f, N, C, M, mu F (U), mu N (U), mu C (U), mu M (U) are membership functions corresponding to F, N, C and M respectively, and F, N, C and M can be regarded as the mapping from the factor set U to the decision set V. Using the decision set V ═ V (V) simultaneously1,v2,v3,v4) Describing the current congestion status of the network, v1,v2,v3,v4And respectively corresponding to network idle, network normal, network busy and network congestion according to the grades.
In order to obtain the accurate congestion state of the current network, comprehensive judgment is carried out on the state of the NDN network by utilizing fuzzy comprehensive judgment. A fuzzy relation exists from the factor set U to the decision set V, which is represented by a fuzzy matrix R and is noted as:
μF(u1)、μN(u1)、μC(u1) And μ M (u)1) Membership functions corresponding to four fuzzy subsets respectively representing average buffer queue length of interest packets, mu F (u2)、μN(u2)、μC(u2) And μ M (u)2) And the membership functions correspond to four fuzzy subsets respectively representing the round trip time ratio.
Each factor in the factor set U has a different influence on the decision set transaction, and therefore each factor is assigned a corresponding weight, which is 1 Fuzzy vector on the domain U, and is noted as:
W=(ω1,ω2)∈F(U)
wherein ω is1The weight of the average buffer queue length factor of the interest packets in the factor set U, omega2Is weighted by the round trip time ratio factor in U and satisfies
ω1+ω2=1
The fuzzy comprehensive evaluation result is as follows:
wherein: b is a fuzzy subset of decision set V, B1,b2,b3,b4Is of the order v1,v2,v3,v4And comprehensively judging the membership degree of the fuzzy subset B.
And 4.3, selecting the maximum membership degree and judging the congestion state of the network.
Comparison b1,b2,b3,b4There is a p such that
Bp=max{b1,b2,b3,b4}
According to the principle of maximum membership degree, the network BpThe determined stateAnd thus a solution is selected. And 5, according to the congestion state of the network, the downstream routing node adjusts the sending rate of the interest packet and the middle routing node selects other proper ports to control the network congestion of the named data.
The rate regulation algorithm is based on an AIMD algorithm and adopts a display feedback mechanism, and a downstream routing node learns the specific network congestion state so as to regulate the sending rate of the interest packet. When receiving the NACK _ FREE packet, the network is in an idle state, and in order to occupy idle resources of the network as soon as possible, a Multiplicative Increase (MI) algorithm is used for the rate; when a NACK _ NORMAL packet is received, the rate adopts an Additive Increase (AI) algorithm, so that the impact of burst flow on the network can be prevented; when receiving the NACK _ BUSY packet, it indicates that the network is in a saturated state, and the transmission rate should be rapidly reduced, so a Multiplicative reduction (MD) algorithm is adopted; when receiving the NACK _ CONGESTION packet, the network is already in a high-load state, and at this time, the intermediate routing node selects another suitable port for the request packet to reduce the load of the current link, so that the network can stably operate. The specific algorithm is as follows:
MI:SendRate(t+ep)=SendRate(t)*(1+ζ)
MD:SendRate(t+ep)=SendRate(t)*ψ
In order to verify the effectiveness of the method provided by the invention, different topologies are used for evaluating the algorithm, in order to clearly see the influence of the control method provided by the invention on the network performance, HR-ICP and ARMN which are widely accepted are selected as reference control methods, and the three evaluation indexes of packet loss rate, average time delay and throughput are compared and analyzed. The experiment was basically set up as follows:
a) experimental topologies employing dumbbell and multipath topologies as shown in fig. 3 and 4, the consumer requests data at a rate of 500 packets of interest per second, stopping the request after 10 seconds.
b) The bandwidth of the link between the content requester and the intermediate node is set to 10Mbps, and the bandwidth of the link between the intermediate nodes is set to 2Mbps
c) The size of the interest packet is fixed to 32 bytes, and the size of the data packet is fixed to 1024 bytes
d) Fetching under the condition of average buffer queue length of factor 1 interest packet
e) Taken under the condition of a factor of 2 round trip time ratio
Fig. 4 and 5 show the change of average delay of two topologies with the number of requests, and when the number of interest requests is greater than 1250, the three control methods gradually exhibit different performances. The control method provided by the invention predicts the network state at the intermediate node, and rapidly adjusts the sending rate of the interest packet by setting a feedback mechanism of various congestion signals through fuzzy evaluation. When the network is in a congested situation, the intermediate node can also autonomously select other available ports, so that the lowest average delay is always kept. In the ARMN, the end node uses the NACK packet as congestion information to perform window adjustment, but under the condition that the network is congested, the return of the NACK packet is delayed and the network congestion is further increased, so that the end node cannot acquire the network state in time, and the delay is increased. The HR-ICP end node is provided with a round-trip delay-based window adjustment strategy, and the round-trip delay can be controlled at the end node to a certain extent.
Fig. 6 and 7 show the case that the packet loss rate changes with the number of requests, the change trends of the packet loss rate with the number of interest requests in the two topologies are substantially similar, and the change of the packet loss rate with the increase of the number of interest requests is relatively stable when the number of interest requests is small; when the number of interest requests is gradually increased, the three congestion control methods gradually exhibit different performances. The control method comprises the steps that a rate adjustment mechanism based on various congestion signals is arranged at a node, the design of congestion feedback signals can better meet the actual state of a network, and an overtime timing mechanism based on an NDN multi-source network scene is arranged at the end node, so that the packet loss rate is lowest. The HR-ICP takes RTO as a congestion signal, when a data packet is overtime, the congestion is considered to occur, and a congestion window is multiplicatively reduced, so that the packet loss rate is controlled to a certain extent; ARMN takes NACK packet as congestion signal to carry out window adjustment at end node, NACK packet needs to reach the end node by tracing back hop by hop, the end node can not obtain network state in time to carry out rate adjustment, and the packet loss rate is maximum.
The throughput is shown by fig. 8 and 9 as a function of the packet transmission time. When the number of interest requests is small, the throughput increases obviously along with the increase of the number of interest requests; when the number of interest requests is larger than 2000, the throughput is not obviously increased under the influence of the number of interest requests, and different mechanisms gradually show different performances. The adjusting mechanism of the ICP adopted by the HR-ICP at the end node is mainly based on RTT, when the RTT is too large, the window is sharply reduced, so that a port queue is always in a smaller state, and the final throughput of the end node is lower; ARMN utilizes the idea of multipath forwarding, and when the network is congested, the interface is searched for the NACK packet again instead of delayed forwarding, so that the throughput is ensured; according to the method, on one hand, the network state is monitored at the intermediate node and the sending rate is adjusted in a self-adaptive mode, the MIAIMD algorithm is adopted for adjusting the sending rate, the link bandwidth is occupied quickly when the network is idle, and on the other hand, other available interfaces can be searched again when the node rate adjustment algorithm cannot adapt to the network well, so that the algorithm provided by the method can fully utilize network resources and always keep higher throughput.
Claims (8)
1. A named data network congestion control method based on a fuzzy comprehensive evaluation algorithm is characterized by comprising the following steps: the method comprises the following steps:
step 1, designing a named data network to have four congestion states of network idle, normal network, busy network and network congestion;
step 2, obtaining the average buffer queue length of the interest packet;
step 3, obtaining a round-trip time ratio;
step 4, carrying out fuzzy evaluation according to the average buffer queue length of the interest packet and the round-trip time ratio;
step 4.1, defining a factor set U and a decision set V in fuzzy evaluation, wherein the average cache queue length and the round-trip time ratio of the interest packet are the parameters of the factor set U; determine 4 fuzzy subsets of the factor set U: F. n, C and M, μ f (U), μ n (U), μ c (U) and μ M (U) correspond to membership functions of F, N, C, M, F, N, C, M as a mapping of factor set U to decision set V, decision set V ═ V (V ═ V), respectively1,v2,v3,v4) Describing a current congestion status of the network;
and 4.2, expressing the Fuzzy relation from the factor set U to the decision set V by using a Fuzzy matrix R, distributing corresponding weights according to different influences of the average buffer queue length factor and the round-trip time ratio factor of the interest packets on the decision set, marking the weights as Fuzzy vectors W, and obtaining a Fuzzy subset of the decision set V according to the Fuzzy matrix R and the Fuzzy vectors Wb1,b2,b3,b4The membership degree of the decision set V;
4.3, selecting the maximum membership degree and judging which congestion state the network is in;
and 5, according to the congestion state of the network, the downstream routing node adjusts the sending rate of the interest packet and the middle routing node selects other proper ports to control the network congestion of the named data.
2. The named data network congestion control method based on the fuzzy comprehensive judgment algorithm as claimed in claim 1, wherein: average buffer queue length AQL of interest packetfinalIs calculated byIs of the formula
Wherein, QueueLengthcurrent(j)Buffer queue length for current port, i +1 represents the number of times the packet of interest buffer queue length is measured, EjRepresenting QueueLengthcurrent(j)The weight of (c).
4. The named data network congestion control method based on the fuzzy comprehensive judgment algorithm as claimed in claim 1, wherein: the method for calculating the round-trip time ratio RTTR is
INT=RTO-(Interesttime+Datatime)
RT is the actual time when the intermediate routing node receives the data packet sent back by the content node, INT is the timeout time from the intermediate routing node to the content node which can meet the request, RTO is the timeout retransmission time, InteresttimeFor the transmission time of interest packets from sender to intermediate node, DatatimeThe time for the data packet to reach the interested requesting end from the intermediate routing node.
5. The named data network congestion control method based on fuzzy comprehensive judgment algorithm as claimed in claim 4, characterized in that: time Data of Data packet from intermediate routing node to interest request endtimeIs calculated by
Wherein, the packageDataSizeIndicates packet size and packageInterestIndicates the size of the Interest packet, InteresttimeIndicating the packet of interest transmission time.
6. The named data network congestion control method based on fuzzy comprehensive judgment algorithm as claimed in claim 4, characterized in that: the calculation method of the time-out retransmission time RTO comprises the following steps
Wherein:is the average round trip time, RTT, of the most recently received packetnIs the round trip time of the last received packet,is a constant from 0 to 1 and is,is the average round trip time of the received data packet before the current data packet reception, f is the constant change adapted to the RTT, and σ is the estimation value of the standard deviation of the RTT。
7. The named data network congestion control method based on the fuzzy comprehensive judgment algorithm as claimed in claim 1, wherein: the blur matrix R is represented as
μF(u1)、μN(u1)、μC(u1) And μ M (u)1) Membership functions corresponding to four fuzzy subsets respectively representing average buffer queue length of interest packets, mu F (u2)、μN(u2)、μC(u2) And μ M (u)2) And the membership functions correspond to four fuzzy subsets respectively representing the round trip time ratio.
8. The named data network congestion control method based on the fuzzy comprehensive judgment algorithm as claimed in claim 1, wherein: when the network is idle, the downstream routing nodes adopt a multiplicative increasing algorithm to adjust the sending rate of the interest packets, when the network is normal, the downstream routing nodes adopt an additive increasing algorithm to adjust the sending rate of the interest packets, when the network is busy, the downstream intermediate routing nodes adopt a multiplicative decreasing algorithm to adjust the sending rate of the interest packets, and when the network is congested, the intermediate routing nodes select other proper ports to reduce the load of the current link.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110181006.3A CN112804125B (en) | 2021-02-09 | 2021-02-09 | Named data network congestion control method based on fuzzy comprehensive evaluation algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110181006.3A CN112804125B (en) | 2021-02-09 | 2021-02-09 | Named data network congestion control method based on fuzzy comprehensive evaluation algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112804125A true CN112804125A (en) | 2021-05-14 |
CN112804125B CN112804125B (en) | 2022-03-18 |
Family
ID=75814987
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110181006.3A Active CN112804125B (en) | 2021-02-09 | 2021-02-09 | Named data network congestion control method based on fuzzy comprehensive evaluation algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112804125B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113746748A (en) * | 2021-09-10 | 2021-12-03 | 中南民族大学 | Explicit congestion control method in named data network |
CN114615199A (en) * | 2022-04-25 | 2022-06-10 | 曲阜师范大学 | TCP network congestion control method, device, terminal and readable storage medium |
CN114827036A (en) * | 2022-04-18 | 2022-07-29 | 天津大学 | NDN hop-by-hop congestion control method with cache perception based on SDN |
CN115002036A (en) * | 2022-05-26 | 2022-09-02 | 国网河北省电力有限公司电力科学研究院 | NDN network congestion control method, electronic device and storage medium |
CN116506354A (en) * | 2023-06-30 | 2023-07-28 | 天地信息网络研究院(安徽)有限公司 | NDN router and active congestion packet loss recovery method thereof |
CN117955979A (en) * | 2024-03-27 | 2024-04-30 | 中国电子科技集团公司第五十四研究所 | Cloud network fusion edge information service method based on mobile communication node |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102006230A (en) * | 2010-11-26 | 2011-04-06 | 中南大学 | Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network |
US20160043963A1 (en) * | 2014-08-11 | 2016-02-11 | Cisco Technology, Inc. | Maintaining Named Data Networking (NDN) Flow Balance with Highly Variable Data Object Sizes |
CN107634911A (en) * | 2017-10-31 | 2018-01-26 | 河南科技大学 | Adaptive congestion control method based on deep learning in a kind of information centre's network |
WO2018225039A1 (en) * | 2017-06-09 | 2018-12-13 | Telefonaktiebolaget Lm Ericsson (Publ) | Method for congestion control in a network |
CN111526096A (en) * | 2020-03-13 | 2020-08-11 | 北京交通大学 | Intelligent identification network state prediction and congestion control system |
-
2021
- 2021-02-09 CN CN202110181006.3A patent/CN112804125B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102006230A (en) * | 2010-11-26 | 2011-04-06 | 中南大学 | Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network |
US20160043963A1 (en) * | 2014-08-11 | 2016-02-11 | Cisco Technology, Inc. | Maintaining Named Data Networking (NDN) Flow Balance with Highly Variable Data Object Sizes |
WO2018225039A1 (en) * | 2017-06-09 | 2018-12-13 | Telefonaktiebolaget Lm Ericsson (Publ) | Method for congestion control in a network |
CN107634911A (en) * | 2017-10-31 | 2018-01-26 | 河南科技大学 | Adaptive congestion control method based on deep learning in a kind of information centre's network |
CN111526096A (en) * | 2020-03-13 | 2020-08-11 | 北京交通大学 | Intelligent identification network state prediction and congestion control system |
Non-Patent Citations (1)
Title |
---|
曾振平等: "因特网拥塞控制研究综述", 《小型微型计算机系统》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113746748A (en) * | 2021-09-10 | 2021-12-03 | 中南民族大学 | Explicit congestion control method in named data network |
CN114827036A (en) * | 2022-04-18 | 2022-07-29 | 天津大学 | NDN hop-by-hop congestion control method with cache perception based on SDN |
CN114827036B (en) * | 2022-04-18 | 2023-09-29 | 天津大学 | SDN-based NDN hop-by-hop congestion control method with cache perception |
CN114615199A (en) * | 2022-04-25 | 2022-06-10 | 曲阜师范大学 | TCP network congestion control method, device, terminal and readable storage medium |
CN114615199B (en) * | 2022-04-25 | 2023-10-13 | 曲阜师范大学 | TCP network congestion control method, device, terminal and readable storage medium |
CN115002036A (en) * | 2022-05-26 | 2022-09-02 | 国网河北省电力有限公司电力科学研究院 | NDN network congestion control method, electronic device and storage medium |
CN115002036B (en) * | 2022-05-26 | 2023-07-25 | 国网河北省电力有限公司电力科学研究院 | NDN network congestion control method, electronic equipment and storage medium |
CN116506354A (en) * | 2023-06-30 | 2023-07-28 | 天地信息网络研究院(安徽)有限公司 | NDN router and active congestion packet loss recovery method thereof |
CN116506354B (en) * | 2023-06-30 | 2023-09-12 | 天地信息网络研究院(安徽)有限公司 | NDN router and active congestion packet loss recovery method thereof |
CN117955979A (en) * | 2024-03-27 | 2024-04-30 | 中国电子科技集团公司第五十四研究所 | Cloud network fusion edge information service method based on mobile communication node |
Also Published As
Publication number | Publication date |
---|---|
CN112804125B (en) | 2022-03-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112804125B (en) | Named data network congestion control method based on fuzzy comprehensive evaluation algorithm | |
US9088510B2 (en) | Universal rate control mechanism with parameter adaptation for real-time communication applications | |
CN106160953B (en) | A kind of transmission method based on learning-oriented energy efficiency model | |
CN113746748B (en) | Explicit congestion control method in named data network | |
JP2004538719A (en) | Method for providing a non-linear, highly scalable increase-decrease congestion control mechanism | |
Wu et al. | GTP: Group transport protocol for lambda-grids | |
CN106878192B (en) | Data scheduling method of self-adaptive MPTCP | |
CN113613286A (en) | Space-ground integrated intelligent network multi-path congestion control method | |
Mamatas et al. | Approaches to congestion control in packet networks | |
Aboobaker et al. | Streaming media congestion control using bandwidth estimation | |
Qazi et al. | Congestion control with multipacket feedback | |
Zhong et al. | A rate-based multipath-aware congestion control mechanism in named data networking | |
CN116389375A (en) | Network queue management method, device and router for live video stream | |
Asmaa et al. | EC-elastic an explicit congestion control mechanism for named data networking | |
Shorten et al. | On queue provisioning, network efficiency and the transmission control protocol | |
JP3853784B2 (en) | Data communication management method | |
Cui et al. | NDN congestion control based on fuzzy comprehensive evaluation algorithm | |
Bouras et al. | Adaptive smooth multicast protocol for multimedia data transmission | |
CN114553783B (en) | Load balancing method for self-adaptive regulation of cell granularity of data center network | |
Chakareski et al. | Rate-distortion optimized packet scheduling over bottleneck links | |
Li | PFECC: a precise feedback-based explicit congestion control algorithm in nameddata networking | |
CN116760777B (en) | Multipath congestion control method based on ABEA3C | |
US10833999B2 (en) | Active request management apparatus in stateful forwarding networks and method thereof | |
Suchara et al. | Implementation of provably stable maxnet | |
Kim et al. | Loss proportional decrease based congestion control in the future Internet |
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 |