CN116455821A - Rate-based multipath perceived congestion control method in named data network - Google Patents
Rate-based multipath perceived congestion control method in named data network Download PDFInfo
- Publication number
- CN116455821A CN116455821A CN202310721754.5A CN202310721754A CN116455821A CN 116455821 A CN116455821 A CN 116455821A CN 202310721754 A CN202310721754 A CN 202310721754A CN 116455821 A CN116455821 A CN 116455821A
- Authority
- CN
- China
- Prior art keywords
- packet
- interface
- data
- rate
- data packet
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000005540 biological transmission Effects 0.000 claims description 67
- 230000004044 response Effects 0.000 claims description 8
- 238000011144 upstream manufacturing Methods 0.000 claims description 7
- 239000004973 liquid crystal related substance Substances 0.000 claims description 6
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 3
- 230000008447 perception Effects 0.000 abstract 1
- 230000007246 mechanism Effects 0.000 description 14
- 238000002474 experimental method Methods 0.000 description 9
- 230000008713 feedback mechanism Effects 0.000 description 8
- 238000004088 simulation Methods 0.000 description 4
- 230000003139 buffering effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- 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
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/22—Alternate routing
-
- 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
- H04L47/26—Flow control; Congestion control using explicit feedback to the source, e.g. choke packets
- H04L47/263—Rate modification at the source after receiving feedback
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Abstract
The invention discloses a multi-path perception congestion control method based on speed in a named Data network, which comprises the steps of calculating the average sending speed of an Intrest packet and the average sending speed of a Data packet; determining a final forwarding interface through a wheel disc selection algorithm; calculating the allowable sending rate of the Data stream to which the Data packet belongs and the allowable sending rate allocation value of the Data packet; the consumer updates the allowed sending rate of the Interest packet by feeding back the Data packet allowed sending rate. The present invention achieves higher overall throughput by fully utilizing network resources on all available paths.
Description
Technical Field
The invention relates to the field of computer network data congestion control, in particular to a multi-path sensing congestion control method based on speed in a named data network.
Background
NDN is a very promising future network architecture, which is quite different from TCP/IP. In NDN, there is no IP address and all data contents are named by unique data content names. The communication is initiated by the consumer requesting the desired data content from the network by sending an Interest packet with a unique desired data content name. The routing node forwards the Interest packet to the corresponding node according to the name of the data content, and the node can be a producer with the data content or a routing node for caching the data content. In TCP/IP, there are two schemes for implementing congestion control, a sliding window based congestion control mechanism and a rate based congestion control mechanism, respectively. Congestion control mechanisms based on sliding windows have been widely studied and deployed that utilize a sliding window to limit the transmission of packets and to adjust the window size based on congestion indications (e.g., receipt of normal ACK messages, packet timeouts, and out-of-order messages). Rate-based congestion control is the latter of TCP/IP congestion control, which calculates the sending rate of each data flow in a routing node and feeds it back to the endpoint. But the rate-based protocol is not friendly to TCP/IP and is not compatible with sliding window based mechanisms and therefore not widely used in practical applications.
Fortunately, NDN is a new network architecture. So far, there is no standard congestion control mechanism in NDN, nor is there a congestion control foundation as strong as the congestion control mechanism based on sliding window in TCP/IP, which brings new opportunities for the congestion control mechanism based on rate. Traditional congestion control mechanisms based on sliding windows increase the congestion window when normal ACK messages are received, and reduce the congestion window by measuring RTT, timeout and out of order data delivery when congestion is detected. In NDN, however, data packets may be retrieved from multiple sources through multiple paths in the network cache, thereby varying RTT, making RTO values difficult to calculate, and Data packets from different caches may arrive out of order. This makes RTT-based timeout and unordered data delivery unreliable congestion indicators. However, the rate-based congestion control mechanism does not depend on these metrics.
The invention adopts a scheme based on the rate, explores the combination of the multipath forwarding characteristics of RCP and NDN, designs, realizes and evaluates a multipath sensing congestion control mechanism based on the rate, allows routing nodes to fairly feed back rate information and avoid congestion, and simultaneously provides a multipath forwarding strategy to realize higher total throughput by fully utilizing network resources on all available paths.
Disclosure of Invention
The invention aims to solve the problems in the prior art, and provides a multi-path sensing congestion control method based on speed in a named data network, which can be used for detecting and responding to congestion states more quickly and realizing low packet loss rate and quick data transmission under high link utilization.
The above object of the present invention is achieved by the following technical solutions:
a rate-based multipath aware congestion control method in a named data network, comprising the steps of:
step 1, a routing node divides an Intrest packet and a Data packet into different Intrest flows and Data flows according to different name prefixes of a named Data network, and calculates an average sending rate of the Intrest packetAnd an average transmission rate D (t) of Data packets;
step 2, determining a final forwarding interface through a wheel disc selection algorithm according to the allowable transmission rate A (t) of an Interest packet recorded by the FIB table entry;
step 3, calculating the allowable sending rate of the Data flow f to which the Data packet belongsAnd the allowed transmission rate allocation value of the Data packet +.>When receiving a Data packet, a routing node transmits an allowable transmission rate R of a Data stream to which the Data packet belongs f (t) assigned to each interface, and when the Data packet is forwarded to interface k, setting the allowable transmission rate B (t) of the Data packet to be the allowable transmission rate assignment value +.>；
And 4, updating the allowable sending rate A (t) of the Intrest packet by the consumer through feeding back the allowable sending rate B (t) of the Data packet.
Step 1 as described above comprises the steps of:
step 1.1, dividing an Interest packet and a Data packet into different Interest flows and Data flows according to different name prefixes of named Data networks;
step 1.2, calculating an average sending rate of an Interest packet at a time t corresponding to each interface:
，
wherein I (t) is the average sending rate of the Intrest packet at the t moment corresponding to each interface, I (t-d) is the average sending rate of the Intrest packet at the t-d moment corresponding to each interface, eta is the target link utilization, alpha and beta are parameters for controlling stability and performance respectively,for the number of packets that pass through the interface at time t, and (2)>For the number of Data packets passing through the interface at time t, C is the available bandwidth of the link, y (t) is the interface's +.>The used bandwidth, q (t) represents the length of the buffer queue at time t, d t Is the average response time between the consumer sending the Interest packet to the matching Data packet reaching the current routing node at time t, s is the average of the size ratio of the Data packet to the Interest packet,
step 1.3, calculating an average sending rate D (t) of the Data packets of the same interface:。
step 1 as described above further comprises the steps of:
step 1.4, when the interface receives a new Interest packet, the interface is processed by the following common formulaUpdating the average sending rate of the Interest packet:，
wherein I is new (t) represents the average transmission rate of the updated Intrest packets, N i (t) represents the number of packets passing through the interface at time t, N d (t) represents the number of Data packets passing through the interface at time t,
step 1.5, when the interface receives a new Data packet, updating the average sending rate of the Data packet by using the following formula:，
wherein D is new (t) represents the average transmission rate of the updated Data packets, N i (t) represents the number of packets passing through the interface at time t, N d (t) indicates the number of Data packets passing through the interface at time t.
Step 2 as described above comprises the steps of:
step 2.1, recording an allowable sending rate A (t) of each Interest packet by an FIB table entry, setting the allowable sending rate A (t) of the Interest packet as a forwarding weight of the Interest packet, recording N available forwarding interfaces in the FIB table entry when a routing node receives a request of the Interest packet, and selecting the probability of an interface hThe method is calculated by the following formula: />，
Wherein, the liquid crystal display device comprises a liquid crystal display device,allowing a transmission rate for an Interest packet of the interface h at the time t; />The transmission rate is allowed for the interval packet of interface i at time t,
step 2.2 cumulative probability per interface hThe method is calculated by the following formula: />，
Generating a random value r between 0 and 1 using a random function ifThe first interface is selected, otherwise, when +.>，/>And selecting a kth interface to forward the Interest packet.
Step 3 as described above comprises the steps of:
step 3.1, in the routing node, each interface records the average sending rate I (t) of the Intrest packet and the average sending rate D (t) of the Data packet,
the direction in which an Interest packet is sent from a consumer to a producer is referred to as the upstream direction, the direction in which a Data packet is returned from a producer to a consumer is referred to as the downstream direction, when an interface receives a Data packet at a permissible sending rate B (t) from an upstream routing node, a (t) is updated to a smaller value between I (t) x s and B (t),
step 3.2, when the routing node forwards the Data packet, the routing node calculates and informs the downstream routing node of the allowable sending rate of the Data flow f to which the Data packet belongs：/>，
Wherein L is a forwarding interface available list of the Interest packet in a routing node FIB table entry,indicating the transmission rate allowed by the Interest packet of the forwarding interface available list item j in the FIB list item,
when forwarding Data packetsThe routing node replaces the Data packet allowed sending rate B (t) with the Data flow f allowed sending rate R to which the Data packet belongs f (t), further, updating B (t) toA smaller value between D (t), taking B (t) as the feedback value, a Data packet with feedback value B (t) will be forwarded to the corresponding interface according to the record of the PIT entry,
step 3.3, when the routing node forwards an Interest packet to an interface or when the interface receives an Interest packet of a downstream routing node, the interface updates the feedback value F (t) of the Interest packet to a smaller value between F (t) and I (t) x s,
step 3.4, the routing node maintains an input interface list for each Interest flow, records all interfaces receiving the Interest packet belonging to the Interest flow, reserves the weight W of each interface, and calculatesDelta is a weighting factor, F (t) is a feedback value of the t-time Intrest packet, W t The weight W is taken as the value at the time t, W t-1 For the weight W to take the value at the time t-1,
when receiving a Data packet, a routing node obtains a corresponding input interface list of the Interest packet according to the Data content name of the Data packetNote the Interest packet input interface list +.>M interfaces, the allowable transmission rate R of the Data stream to which the Data packet belongs is determined based on the following formula f (t) assigned to each interface:
，
，
wherein M is a tableInput interface list showing the corresponding Interest packet of the Data packetNumber of middle interface (L/L)>The weight of the p-th interface at time t is represented, and (2)>Assigning a value to the allowed sending rate of the Data packets of the interface, for>The subscripts m and p of (2) are interface sequence numbers, < >>Is the intermediate weight of the p-th interface,
when a Data packet is forwarded to interface k, the allowable transmission rate B (t) of the Data packet is set to the allowable transmission rate allocation value of the Data packet。
Step 4 as described above comprises the steps of:
when a Data packet arrives at the consumer, the consumer extracts a Data packet transmission permission rate B (t) from the Data packet, and updates the transmission permission rate a (t) of the packet to a smaller value between I (t) ×s and B (t).
Compared with the prior art, the invention has the following beneficial effects:
due to the problems of multisource, multipath, PIT aggregation and the like in the NDN, the RTT value of the Intrest packet is difficult to calculate accurately, and the traditional end-to-end congestion control algorithm is not suitable for the NDN. The invention updates the sending rate of the routing node by a congestion control mechanism based on the rate, thereby avoiding the problems that RTT is difficult to estimate and unordered data packets arrive and cannot be used as congestion signals. The current routing node updates the average sending rate of the Intrest packet and the Data packet through different prefixes of the Data content names, selects an Intrest packet forwarding interface according to the congestion degree, calculates the sending rate of the feedback Data packet through the average sending rate of the Intrest packet and the Data packet, and finally updates the sending rate of the Intrest packet, so that congestion control is realized, and an experimental result in ndnSIM2.0 shows that the method is effective and fair in an NDN network, and the proposed multipath forwarding strategy can fully utilize the available link bandwidth to realize traffic load balance.
Detailed Description
The present invention will be further described in detail below in conjunction with the following examples, for the purpose of facilitating understanding and practicing the present invention by those of ordinary skill in the art, it being understood that the examples described herein are for the purpose of illustration and explanation only and are not intended to limit the invention.
A rate-based multipath aware congestion control method in a named data network, comprising the steps of:
step 1, a routing node divides an Interest packet and a Data packet into different Interest flows and Data flows according to different Data content name prefixes, and calculates an average sending rate of the Interest packetAnd an average transmission rate D (t) of Data packets;
the step 1 specifically comprises the following steps:
step 1.1, dividing an Interest packet and a Data packet into different Interest flows and Data flows according to different Data content name prefixes;
in step 1.2, in the NDN network, the routing node forwards the interference packet and the Data packet to its interface, resulting in that the interference packet and the Data packet simultaneously flow out of the interface, which causes the method to consider the influence of the interference packet when allocating bandwidth for the Data packet. If the interfaces cannot forward the Data packets in time, the Data packets are queued at the routing nodes, and in order to empty the queuing queues of the Data packets and allocate the available link bandwidth to the Data packets of each interface, the following formula is used to calculate the average sending rate of the Intrest packets at the time t corresponding to each interface：
（1）
In the formula, I (t) is the average sending rate of the Intrest packet at the t moment corresponding to each interface, I (t-d) is the average sending rate of the Intrest packet at the t-d moment corresponding to each interface, eta is the target link utilization rate, and alpha and beta are parameters for controlling stability and performance respectively.Indicating the number of packets of Interest through the interface at time t,/->The number of Data packets passing through the interface at time t. Wherein C is the available bandwidth of the link, y (t) is the interface's +.>Bandwidth already used. q (t) represents the length of the buffer queue at time t, d t The average response time between the consumer sending the Interest packet to the matching Data packet reaching the current routing node at time t is the average value of the size ratio of the Data packet to the Interest packet.
The average response time was calculated using an exponentially weighted moving average (Exponential Weighted Moving Average, EWMA):wherein d is t For a moving average predicted value of the average response time d at time t, θ t For the true value of the average response time d at time t,/->As a weighting factor, d t-1 And sending the Interest packet to the average response time between the arrival of the matched Data packet at the current routing node for the consumer at the time t-1.
Step 1.3, calculating an average sending rate D (t) of the Data packets of the same interface by using the average sending rate I (t) of the Interest packets:
（2）
in step 1.4, when the interface receives a new packet, since the above formula already allocates the link bandwidth to all the passed packets, there is no available bandwidth for the new packet, which may be significantly out of range and cause packet loss. To avoid this problem, it is necessary to reduce the average transmission rate of other packets of Interest. When a new Interest packet arrives, the method updates the average sending rate of the Interest packet by the following formula:
（3）
wherein I is new (t) represents the average transmission rate of the post-update Interest packet, and is different from the pre-update I (t). N (N) i (t) represents the number of packets passing through the interface at time t, N d (t) indicates the number of Data packets passing through the interface at time t.
Step 1.5, when the interface receives a new Data packet, updating the average sending rate of the Data packet by using the following formula:
（4）
wherein D is new (t) represents the average transmission rate of Data packets after update, and is different from D (t) before update. N (N) i (t) represents the number of packets passing through the interface at time t, N d (t) indicates the number of Data packets passing through the interface at time t.
And 2, in the NDN, the data content routing node can be retrieved from multiple paths, and the method determines a final forwarding interface through a wheel disc selection algorithm according to the transmission rate A (t) allowed by the Intrest packet recorded by the FIB table entry, and fully utilizes network resources to realize higher total throughput.
The step 2 specifically comprises the following steps:
and 2.1, recording an allowable transmission rate A (t) of the Interest packet according to the FIB table entry of the Interest packet, and setting the A (t) as a forwarding weight of the Interest packet. When the routing node receives a request of an Interest packet, if the FIB table entry has N available forwarding interfaces, selecting the available forwarding interfaces through a wheel disc selection algorithm, and selecting the interfacesProbability of hThe method can be calculated by the following formula:
（5）
wherein, the liquid crystal display device comprises a liquid crystal display device,allowing a transmission rate for an Interest packet of the interface h at the time t; />The transmission rate is allowed for the interval packet of interface i at time t.
Step 2.2, after obtaining the selected probability of each interface, the cumulative probability of each interface hThe method can be calculated by the following formula:
（6）
wherein the cumulative probability of each interface h is calculatedIt is understood that when h is 1,2,3 … N,is a function of the sum of the numbers.
The forwarding interface of the Interest packet can be determined by the cumulative probability. The method uses a random function to generate a random value r between 0 and 1 ifThe first interface is selected, otherwise, when +.>And when k is more than or equal to 2 and less than or equal to N, selecting a kth interface to forward the Intrest packet.
Step 3, calculating Data packageAllowed sending rate of the belonging Data stream fAnd the allowed transmission rate allocation value of the Data packet +.>When receiving a Data packet, a routing node transmits an allowable transmission rate R of a Data stream to which the Data packet belongs f (t) assigned to each interface, and when the Data packet is forwarded to interface k, setting the allowable transmission rate B (t) of the Data packet to be the allowable transmission rate assignment value +.>And assigned to each interface.
The step 3 specifically comprises the following steps:
step 3.1, in NDN, the Data packet sizes of the different Data streams may not be uniform. Feeding back the allowed transmission rate of the Interest packet may result in that the Interest packet of the same transmission rate may pull back the Data packet of a different transmission rate. Therefore, the method feeds back the allowable sending rate of the Data packets so as to ensure fairness among consumers with different Data packet sizes.
In the routing node, each interface calculates and records the average sending rate I (t) of the Interest packets and the average sending rate D (t) of the Data packets through the step 1, and simultaneously records the allowable sending rate A (t) of each Interest packet according to the FIB table entry.
The direction in which the Intest packet is sent from the consumer to the producer is referred to as the upstream direction, and the direction in which the Data packet is returned from the producer to the consumer is referred to as the downstream direction. When the interface receives a Data packet with a transmission permission rate of B (t) from the upstream routing node, in order to avoid the problem of packet loss, the interface needs to reduce the transmission permission rate of the packet of the Interest a (t), that is, update a (t) to a smaller value between I (t) ×s and B (t), that is:
（7）
step 3.2, when the routing node forwards the Data packet, the routing node needs to calculate and inform the downstream routing node of the Datallowable transmission rate of Data stream f to which packet a belongs：
（8）
Wherein L is a forwarding interface available list of the Interest packet in a routing node FIB table entry,and the transmission rate allowed by the Interest packet of the available list item j of the forwarding interface in the FIB list item is indicated.
When forwarding a Data packet, a routing node replaces the Data packet allowed sending rate B (t) with the Data flow f allowed sending rate R to which the Data packet belongs f (t), further, since the Data packets all have their own average sending rate D (t), in order to ensure that the receiving rate of the Data packets is not greater than D (t), the interface checks B (t) and updates B (t) to beThe smaller value between D (t) is:
（9）
with B (t) as the feedback value, a Data packet with feedback value B (t) will be forwarded to the corresponding interface according to the record of the PIT entry.
In step 3.3, in NDN, the interference packets requesting the same Data content may reach different interfaces, resulting in that the Data packet allowed sending rate (feedback value) must be fed back to multiple interfaces, which may lead to congestion. To avoid this problem, the present method proposes a multipath feedback mechanism: the Interest packet will also carry a feedback value F (t), and on the premise of guaranteeing fairness of Data transmission, data packet allowed sending rate is allocated to multiple interfaces.
When a routing node forwards an Interest packet to an interface or when the interface receives an Interest packet of a downstream routing node, the interface updates the feedback value F (t) of the Interest packet to a smaller value between F (t) and I (t) ×s, namely:
（10）
step 3.4, the routing node maintains an input interface list for each Interest flow, records all interfaces receiving the Interest packet belonging to the Interest flow, reserves the weight W of each interface, and calculates by EWMA (exponentially weighted moving average method):. Wherein delta is a weighting factor, F (t) is a feedback value of the t-time Intrest packet, W t The weight W is taken as the value at the time t, W t-1 The weight W is taken as a value at the time t-1.
In NDN, when receiving a Data packet, a routing node obtains a corresponding input interface list of an Interest packet according to the Data content name of the Data packet. Input interface list of if Intrest packet +.>M interfaces are provided, a multi-path feedback mechanism in the method adopts a maximum and minimum fair allocation principle, and the allowable sending rate R of the Data flow to which the Data packet belongs is determined according to the weight W of each interface f (t) assigned to each interface, the specific formula is as follows:
（11）
（12）
wherein M represents an input interface list of an Interest packet corresponding to the Data packetNumber of middle interface (L/L)>Weight value of p-th interface at t moment, < ->Assigning a value to the allowed sending rate of the Data packets of the interface, for>The subscripts m and p of (2) are interface sequence numbers, < >>Is the intermediate weight of the p-th interface.
When a Data packet is forwarded to interface k, the allowable transmission rate B (t) of the Data packet is set to the allowable transmission rate allocation value of the Data packet。
When the packets of Interest with the same data content name are aggregated into one PIT entry, the mechanism cannot immediately make full use of the link bandwidth, but since step 1 is performed periodically, the available bandwidth can be utilized after a few time intervals d. Without PIT aggregation, this mechanism can well prevent upstream link congestion.
Step 4, in NDN, the communication is initiated by the consumer. In the method, the consumer updates the allowed sending rate A (t) of the Intrest packet by feeding back the allowed sending rate B (t) of the Data packet.
In the initial stage, the consumer sends an Interest packet. Once a Data packet arrives at the consumer, the consumer extracts the Data packet allowable transmission rate B (t) from the Data packet and feeds back, and updates the allowable transmission rate a (t) of the Interest packet according to the Data packet allowable transmission rate B (t) and based on equation (7).
When the routing node loses a packet due to buffer overflow, it sends a Data packet without the requested Data content to the consumer, which also contains the feedback value B (t). The routing node also feeds back the transmission rate to the consumer via a Data packet that does not contain the requested Data content, as in the feedback mode described above in the Data packet. When receiving a Data packet containing no request Data content, the consumer needs to retransmit the lost Interest packet, and updates the allowed transmission rate a (t) of the Interest packet based on formula (7) according to the feedback value B (t) in the Data packet containing no request Data content.
The consumer will also maintain a timeout retransmission (Retransmission Timeout, RTO) mechanism for the Interest packet. Due to multi-source, multi-path and routing node buffering, measurement of Round Trip Time (RTT) for data transmission in NDN is not accurate. In the method, a relatively large timeout retransmission time (200 ms) is set for each Interest packet, and although the response of a consumer to packet loss is slow, the queue length can be kept small and even zero, so that the packet loss is avoided, and the performance of an algorithm is not affected.
Finally, performance simulation analysis will be performed on the inventive method (RMCM). The adopted simulation platform is ndnSIM2.0, ndnSIM2.0 which is an NDN simulation module for realizing the CCNx protocol by C++ writing based on ns-3 network simulation software, so that the functions of basic network protocol, routing forwarding strategy, data packet node cache and the like are realized, various deployment scenes can be simulated, and a large-scale NDN experiment is supported; in step 1 there are three parameters, η is the target link utilization, which is set to 0.9 in all experiments. Alpha and beta are parameters controlling stability and performance, respectively, and it has been found through experimentation that the present invention (RMCM) works well when the alpha value is less than 0.3 and the beta value is between 0.5 and 1.0. Therefore, in the following experiments, α was set to 0.2 and β was set to 0.9.
1. Validity of the invention (RMCM)
(1) Single path scenario: the present invention (RMCM) first uses a dumbbell topology to evaluate the effect of considering the Interest flow in equation (1). Consumer1 (Consumer 1) retrieves the data content from Producer1 (Producer 1) and Consumer2 (Consumer 2) retrieves the data content from Producer2 (Producer 2) beginning with a delay of 5 s.
Experimental results show that in a scenario where the Interest flow is not considered, queues in Router1 (routing node 1) and Router2 (routing node 2) are established when Consumer2 (Consumer 2) starts requesting Data packets at a time point of 5s, and the length of the average 100 Data packets is maintained in Router2 (routing node 2) before Consumer2 (Consumer 2) stops at a time point of 20 s. This illustrates the necessity to take into account the effects of the Interest flow in step 1.
(2) Multipath scenario: next, the effectiveness of the present invention (RMCM) in a multi-source multi-path scenario was experimentally evaluated. Two data sources, producer1 and Producer2, were used in the experiment to store the same data content. Consumer retrieves data content from both data sources.
According to the experimental results, the utilization rate of different links and the packet loss conditions in different nodes are shown, and the average rate of data streams in three bottleneck links is larger than 4Mbps and is close to an optimal value of 4.5Mbps (eta is 0.9). During the experiment, no packet loss occurred. Experimental results show that the invention (RMCM) can effectively support multipath scenes.
(3) Buffering and multipath feedback scenarios: in this scenario, the effectiveness of the buffering and the performance of the multipath feedback mechanism are evaluated, using C1 (Consumer 1) and C2 (Consumer 2) to retrieve the same data content stored in P1 (Producer 1). C2 (Consumer 2) starting at the 8s time point and stopping at the 25s time point. The cache size of R1 (routing node 1) is 5000 and the caching policy adopted is the least recently used (LeastRecently Used, LRU) algorithm. The buffer size of each interface is 100. Experiments with and without multipath feedback mechanisms were performed, respectively.
Experimental results show that C1 (consumer 1) is the only consumer in the network topology before the 8s point in time, retrieving data content at a rate approaching 4.5 Mbps. At the 8s point in time, C2 (Consumer 2) begins to request data, the request hits the cache at R1 (routing node 1), resulting in a higher bandwidth of 18 Mbps. At about the point in time of 9s, C2 runs out of the cache at R1 (routing node 1) and joins C1 (consumer 1) to retrieve the data content from P1 (producer 1), causing a large number of requests to enter P1 (producer 1), which results in an increasing queue in P1 (producer 1).
The same results can also be found between the time points of experiments 12s to 14 s. In the experimental result using the multipath feedback mechanism, after the time point of 14s, C1 (consumer 1) and C2 (consumer 2) retrieve the data content from P1 (producer 1) at the same transmission rate, and the transmission rate of the links R1-P1 is close to the optimal value of 9Mbps. During this time, the queue lengths of P1 (producer 1) and R1 (routing node 1) are zero. In experimental results without the multipath feedback mechanism, between the 13s and 18s time points, C1 (consumer 1) retrieves the data content from the cache of R1 (routing node 1), resulting in C2 (consumer 2) retrieving the data content from P1 (producer 1) at a transmission rate of 9Mbps. But after 19s C1 (consumer 1) joins C2 (consumer 2) to retrieve the data content from P1 (producer 1). This is where no problem of the multipath feedback mechanism starts to manifest, R1 (routing node 1) feeds back a transmission rate of 9Mbps to C1 (consumer 1) and C2 (consumer 2). Because the bottleneck link C1-R1 has a bandwidth of 5Mbps, C1 (consumer 1) retrieves data content at a rate of 4.5Mbps, while C2 (consumer 2) retrieves data content at a rate of 9Mbps. But on the R1-P1 link, the bandwidth is only 10Mbps, less than 13.5Mbps, resulting in a large amount of packet loss. Based on the above analysis, it can be found that the present invention (RMCM) can adapt to the in-network cache in NDN, and the multipath feedback mechanism enables the present invention (RMCM) to maintain fairness, so as to avoid congestion.
2. Fairness of RMCM
Experiments the fairness of the present invention (RMCM) was evaluated by comparison with two other congestion control algorithms IRCF and IRPI. IRCF is a multipath congestion control algorithm that combines remote adaptive active queue management RAAQM and multipath forwarding policy CF. IRPI is a combination of RAAQM and another multipath forwarding strategy PI. The experiment uses the following three scenarios:
equal: four consumers have the same Round Trip Time (RTT) for Data transmission and the same size Data packets.
Diff_rtt: four consumers have different RTTs and Data packets of the same size. The RTTs from C1 (consumer 1), C2 (consumer 2), C3 (consumer 3) and C4 (consumer 4) to R1 (routing node 1) are 1ms, 10ms, 20ms and 40ms, respectively.
Diff_ds: four consumers have the same RTT, but the Data packets are different in size, 1KB, 2KB, 3KB, and 4KB, respectively.
It can be obviously observed from the experimental results that compared with the other two algorithms, the invention (RMCM) can ensure fairness of consumers in different scenes. In equal and different RTT scenarios, the Data packets received at the same rate by different consumers are almost identical. The difference is also not large in the diff_ds scenario, the rate of the consumer with larger Data packet size is slightly larger, but the rate difference between the four consumers is negligible compared to the other two algorithms. In addition, the invention (RMCM) can be found to have better stability than the other two algorithms.
3. Performance of multipath aware forwarding
The above simulation results have demonstrated the effectiveness and fairness of the present invention (RMCM). Next, the performance of the multipath forwarding in the present invention (RMCM) will be verified. The invention selects three paths of R2 (routing node 2) -P1 (producer 1), R2 (routing node 2) -P2 (producer 2) and R2 (routing node 2) -P3 (producer 3) to realize three different scenes, wherein R2 (routing node 2) serves as a consumer to acquire data content from the three producers. Three cases are as follows:
equal: the three paths R2 (routing node 2) -P1 (producer 1), R2 (routing node 2) -P2 (producer 2) and R2 (routing node 2) -P3 (producer 3) have the same bandwidth, are 10Mbps, and RTT is 10ms.
Diff_rtt: the three paths have the same bandwidth, all 10Mbps, but differ in RTT, 10ms (port 257), 50ms (port 258), and 100ms (port 259), respectively.
Diff_bw: the three paths have the same RTT, all 10ms, but different bandwidths, 10Mbps (port 257), 20 Mbps (port 258), 40 Mbps (port 259), respectively.
The present invention (RMCM) is compared with two other multipath forwarding strategies CF and PI. The PI preferentially selects the interface with the least interest packet to forward, and the CF calculates the weight of each interface on the basis of the PI and selects the forwarding interface by using a weighted polling algorithm. And under the condition that each algorithm is stable, counting the transmission traffic of each port.
The experimental result shows that under the Equal scene, the three forwarding strategies have good effects, namely the traffic received by the three ports is basically the same. In the diff_rtt scenario, both PI and CF prefer ports with a lower RTT. In the Diff BW scenario, PI and CF cannot effectively use higher bandwidth ports. The ideal optimum ratio is 14%:28%:57%, but only up to 18%:36%:45% is actually achieved, thus reducing the overall throughput. In both diff_rtt and diff_bw scenarios, the forwarding policies of CF and PI are more prone to paths with lower RTT or higher bandwidth. The multipath forwarding of the present invention directly responds to the available bandwidth, and in any event the desired forwarding split ratio can be achieved, which enables the present invention (RMCM) to quickly achieve higher overall throughput.
It should be noted that the specific embodiments described in this application are merely illustrative of the spirit of the invention. Those skilled in the art may make various modifications or additions to the described embodiments or substitutions thereof without departing from the spirit of the invention or its scope as defined in the accompanying claims.
Claims (6)
1. A rate-based multipath perceived congestion control method in a named data network, comprising the steps of:
step 1, a routing node divides an Intrest packet and a Data packet into different Intrest flows and Data flows according to different name prefixes of a named Data network, and calculates an average sending rate of the Intrest packetAnd an average transmission rate D (t) of Data packets;
step 2, determining a final forwarding interface through a wheel disc selection algorithm according to the allowable transmission rate A (t) of an Interest packet recorded by the FIB table entry;
step 3 calculating the allowable sending rate of the Data flow f to which the Data packet belongsAnd the allowed sending rate allocation value of the Data packetWhen receiving a Data packet, a routing node transmits an allowable transmission rate R of a Data stream to which the Data packet belongs f (t) assigned to each interface, and when the Data packet is forwarded to interface k, setting the allowable transmission rate B (t) of the Data packet to be the allowable transmission rate assignment value +.>；
And 4, updating the allowable sending rate A (t) of the Intrest packet by the consumer through feeding back the allowable sending rate B (t) of the Data packet.
2. The method for rate-based multipath perceived congestion control in a named data network of claim 1, wherein said step 1 comprises the steps of:
step 1.1, dividing an Interest packet and a Data packet into different Interest flows and Data flows according to different name prefixes of named Data networks;
step 1.2, calculating an average sending rate of an Interest packet at a time t corresponding to each interface:
，
wherein I (t) is the average sending rate of the Intrest packet at the t moment corresponding to each interface, I (t-d) is the average sending rate of the Intrest packet at the t-d moment corresponding to each interface, eta is the target link utilization, alpha and beta are parameters for controlling stability and performance respectively,for the number of packets that pass through the interface at time t, and (2)>For the number of Data packets passing through the interface at time t, C is the available bandwidth of the link, y (t) is the interface's +.>The used bandwidth, q (t) represents the length of the buffer queue at time t, d t Is the average response time between the consumer sending the Interest packet to the matching Data packet reaching the current routing node at time t, s is the average of the size ratio of the Data packet to the Interest packet,
step 1.3, calculating an average sending rate D (t) of the Data packets of the same interface:。
3. the method for rate-based multipath perceived congestion control in a named data network of claim 2, wherein said step 1 further comprises the steps of:
step 1.4, when the interface receives a new Interest packet, updating the average sending rate of the Interest packet according to the following formula:，
wherein I is new (t) represents the average transmission rate of the updated Intrest packets, N i (t) represents the number of packets passing through the interface at time t, N d (t) represents the number of Data packets passing through the interface at time t,
step 1.5, when the interface receives a new Data packet, updating the average sending rate of the Data packet by using the following formula:
，
wherein D is new (t) represents the average transmission rate of the updated Data packets, N i (t) represents the number of packets passing through the interface at time t, N d (t) indicates the number of Data packets passing through the interface at time t.
4. A method for rate-based multipath perceived congestion control in a named data network according to claim 3, wherein said step 2 comprises the steps of:
step 2.1, recording an allowable sending rate A (t) of each Interest packet by an FIB table entry, setting the allowable sending rate A (t) of the Interest packet as a forwarding weight of the Interest packet, recording N available forwarding interfaces in the FIB table entry when a routing node receives a request of the Interest packet, and selecting the probability of an interface hThe method is calculated by the following formula: />，
Wherein, the liquid crystal display device comprises a liquid crystal display device,allowing a transmission rate for an Interest packet of the interface h at the time t; />The transmission rate is allowed for the interval packet of interface i at time t,
step 2.2 cumulative probability per interface hThe method is calculated by the following formula: />，
Generating a random value r between 0 and 1 using a random function ifThe first interface is selected, otherwise, when，/>And selecting a kth interface to forward the Interest packet.
5. The method for rate-based multipath perceived congestion control in a named data network of claim 4, wherein said step 3 comprises the steps of:
step 3.1, in the routing node, each interface records the average sending rate I (t) of the Intrest packet and the average sending rate D (t) of the Data packet,
the direction in which an Interest packet is sent from a consumer to a producer is referred to as the upstream direction, the direction in which a Data packet is returned from a producer to a consumer is referred to as the downstream direction, when an interface receives a Data packet at a permissible sending rate B (t) from an upstream routing node, a (t) is updated to a smaller value between I (t) x s and B (t),
step 3.2, when the routing node forwards the Data packet, the routing node calculates and informs the downstream routing node of the allowable sending rate of the Data flow f to which the Data packet belongs：/>，
Wherein L is a forwarding interface available list of the Interest packet in a routing node FIB table entry,indicating the transmission rate allowed by the Interest packet of the forwarding interface available list item j in the FIB list item,
when forwarding a Data packet, a routing node replaces the Data packet allowed sending rate B (t) with the Data flow f allowed sending rate R to which the Data packet belongs f (t) furtherStep B (t) is updated toA smaller value between D (t), taking B (t) as the feedback value, a Data packet with feedback value B (t) will be forwarded to the corresponding interface according to the record of the PIT entry,
step 3.3, when the routing node forwards an Interest packet to an interface or when the interface receives an Interest packet of a downstream routing node, the interface updates the feedback value F (t) of the Interest packet to a smaller value between F (t) and I (t) x s,
step 3.4, the routing node maintains an input interface list for each Interest flow, records all interfaces receiving the Interest packet belonging to the Interest flow, reserves the weight W of each interface, and calculatesDelta is a weighting factor, F (t) is a feedback value of the t-time Intrest packet, W t The weight W is taken as the value at the time t, W t-1 For the weight W to take the value at the time t-1,
when receiving a Data packet, a routing node obtains a corresponding input interface list of the Interest packet according to the Data content name of the Data packetInput interface list of the Interest packet is recorded +.>M interfaces, the allowable transmission rate R of the Data stream to which the Data packet belongs is determined based on the following formula f (t) assigned to each interface:
，
，
wherein M represents the corresponding Data packetInput interface list of Intrest packetNumber of middle interface (L/L)>Weight value of p-th interface at t moment, < ->Assigning a value to the allowed sending rate of the Data packets of the interface, for>The subscripts m and p of (2) are interface sequence numbers, < >>Is the intermediate weight of the p-th interface,
when a Data packet is forwarded to interface k, the allowable transmission rate B (t) of the Data packet is set to the allowable transmission rate allocation value of the Data packet。
6. The method for rate-based multipath perceived congestion control in a named data network of claim 5, wherein said step 4 comprises the steps of:
when a Data packet arrives at the consumer, the consumer extracts a Data packet transmission permission rate B (t) from the Data packet, and updates the transmission permission rate a (t) of the packet to a smaller value between I (t) ×s and B (t).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310721754.5A CN116455821A (en) | 2023-06-19 | 2023-06-19 | Rate-based multipath perceived congestion control method in named data network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310721754.5A CN116455821A (en) | 2023-06-19 | 2023-06-19 | Rate-based multipath perceived congestion control method in named data network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116455821A true CN116455821A (en) | 2023-07-18 |
Family
ID=87130620
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310721754.5A Pending CN116455821A (en) | 2023-06-19 | 2023-06-19 | Rate-based multipath perceived congestion control method in named data network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116455821A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015048773A2 (en) * | 2013-09-30 | 2015-04-02 | Northeastern University | System and method for joint dynamic forwarding and caching in content distribution networks |
| CN105245448A (en) * | 2015-10-08 | 2016-01-13 | 中南民族大学 | Multi-service and multi-path routing optimization method in named data network |
| US9270598B1 (en) * | 2013-12-13 | 2016-02-23 | Cisco Technology, Inc. | Congestion control using congestion prefix information in a named data networking environment |
| CN113162863A (en) * | 2021-04-12 | 2021-07-23 | 河南科技大学 | Fair multipath routing strategy based on wheel disc algorithm |
| CN113746748A (en) * | 2021-09-10 | 2021-12-03 | 中南民族大学 | Explicit congestion control method in named data network |
-
2023
- 2023-06-19 CN CN202310721754.5A patent/CN116455821A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015048773A2 (en) * | 2013-09-30 | 2015-04-02 | Northeastern University | System and method for joint dynamic forwarding and caching in content distribution networks |
| US9270598B1 (en) * | 2013-12-13 | 2016-02-23 | Cisco Technology, Inc. | Congestion control using congestion prefix information in a named data networking environment |
| CN105245448A (en) * | 2015-10-08 | 2016-01-13 | 中南民族大学 | Multi-service and multi-path routing optimization method in named data network |
| CN113162863A (en) * | 2021-04-12 | 2021-07-23 | 河南科技大学 | Fair multipath routing strategy based on wheel disc algorithm |
| CN113746748A (en) * | 2021-09-10 | 2021-12-03 | 中南民族大学 | Explicit congestion control method in named data network |
Non-Patent Citations (2)
| Title |
|---|
| SHANGRU ZHONG ET AL.: "A Rate-based Multipath-aware Congestion Control Mechanism in Named Data Networking", 2017 IEEE INTERNATIONAL SYMPOSIUM ON PARALLEL AND DISTRIBUTED PROCESSING WITH APPLICATIONS AND 2017 IEEE INTERNATIONAL CONFERENCE ON UBIQUITOUS COMPUTING AND COMMUNICATIONS (ISPA/IUCC), pages 3 - 4 * |
| 姜一鸣;雒江涛;: "命名数据网络中的一种冗余控制策略", 计算机应用研究, no. 10 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP2540042B1 (en) | Communication transport optimized for data center environment | |
| US6078564A (en) | System for improving data throughput of a TCP/IP network connection with slow return channel | |
| CN109120544B (en) | Transmission control method based on host end flow scheduling in data center network | |
| US8873385B2 (en) | Incast congestion control in a network | |
| WO2020244448A1 (en) | Congestion control method and device, and computer-readable medium | |
| US20110013512A1 (en) | Transmission control protocol (tcp) congestion control using transmission delay components | |
| CA2732258A1 (en) | Dynamic setting of optimal buffer sizes in ip networks | |
| CN113746748B (en) | Explicit congestion control method in named data network | |
| CN109873773B (en) | Congestion control method for data center | |
| CN112737964B (en) | Transmission control method and system integrating push-pull semantics | |
| Lu | SED: An SDN-based explicit-deadline-aware TCP for cloud Data Center Networks | |
| CN113141314B (en) | Congestion control method and equipment | |
| CN115460156B (en) | Method, device, equipment and medium for controlling lossless network congestion of data center | |
| CN110138678A (en) | Data transfer control method and device and web-transporting device and storage medium | |
| Zhong et al. | A rate-based multipath-aware congestion control mechanism in named data networking | |
| CN116455821A (en) | Rate-based multipath perceived congestion control method in named data network | |
| EP3108631B1 (en) | Buffer bloat control | |
| CN107770082B (en) | Transmission control method based on task flow characteristics in data center network | |
| CN115665060A (en) | Multi-path transmission scheduling method and device for heterogeneous network | |
| CN113162863B (en) | Fair multipath routing method based on wheel disc algorithm | |
| Asmaa et al. | EC-elastic an explicit congestion control mechanism for named data networking | |
| CN107302504A (en) | A kind of multiplexing dispatching method and system based on virtual transmit queue | |
| US10833999B2 (en) | Active request management apparatus in stateful forwarding networks and method thereof | |
| Fang et al. | Prompt congestion reaction scheme for data center network using multiple congestion points | |
| Chen et al. | On meeting deadlines in datacenter networks |
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 |