CN113162863B - Fair multipath routing method based on wheel disc algorithm - Google Patents
Fair multipath routing method based on wheel disc algorithm Download PDFInfo
- Publication number
- CN113162863B CN113162863B CN202110390824.4A CN202110390824A CN113162863B CN 113162863 B CN113162863 B CN 113162863B CN 202110390824 A CN202110390824 A CN 202110390824A CN 113162863 B CN113162863 B CN 113162863B
- Authority
- CN
- China
- Prior art keywords
- port
- forwarding
- path
- interest packet
- probability
- 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.)
- Active
Links
Images
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
- H04L45/12—Shortest path evaluation
-
- 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
- H04L45/14—Routing performance; Theoretical aspects
-
- 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
- H04L45/24—Multipath
Abstract
The fair multi-path routing strategy based on the roulette algorithm comprises the following steps of: step one, obtaining each port capable of forwarding the interest packet and the distributed weight; step two, when the interest packet reaches the routing node and no relevant record of the interest packet exists in the CS and PIT of the routing node, selecting a forwarding path in the eFIB, and triggering a path selection strategy based on a wheel disc algorithm; and step three, calculating the selected probability of each port capable of forwarding the interest packet according to the distributed weight, then accumulating the probability of each port to obtain the accumulated probability of the port, wherein the accumulated probability is represented by line segments with different lengths, a straight line with the length of 1 is formed by combining the line segments, then a number is randomly generated in an interval [0,1], and the interval where the number falls is the forwarding port selected by the strategy, and data is forwarded through the forwarding port. The routing strategy can effectively avoid the problems of data redundancy and unfair flow distribution in the network.
Description
Technical Field
The invention relates to the technical field of computer networks, in particular to a fair multi-path routing strategy based on a roulette algorithm.
Background
In order to ensure the stability of the NDN and the traffic load balance, a research on the forwarding strategy of the NDN is necessary. The NDN project group realizes a flooding route and an optimal path route strategy, the flooding route can forward the interest packet to all available forwarding interfaces, and although the requested data can be quickly obtained, a large amount of redundancy can be brought to the network, and the network load is increased. The optimal path strategy can reduce the requests of interest packets by selecting a shortest path to forward the interest packets, but brings the problem of unfair flow distribution, when a certain forwarding path is frequently selected, the path becomes congested, and other idle paths are not fully used. The named data network internally supports multi-path forwarding, and how to utilize multi-path cooperative forwarding becomes an important issue in NDN research.
Disclosure of Invention
In order to solve the technical problems, the invention provides a fair multi-path routing strategy based on a wheel disc algorithm, which can effectively avoid the problems of data redundancy and unfair flow distribution in a network.
In order to realize the technical purpose, the adopted technical scheme is as follows: the fair multi-path routing strategy based on the roulette algorithm comprises the following steps of:
step one, when a routing node receives a NACK packet fed back by an upstream node, updating a field value in an eFIB table according to link information carried by the NACK packet to obtain each port capable of forwarding the interest packet and an allocated weight;
step two, when the interest packet reaches the routing node and no relevant record of the interest packet exists in the CS and PIT of the routing node, selecting a forwarding path in the eFIB, and triggering a path selection strategy based on a wheel disc algorithm;
step three, the path selection strategy of the wheel disc algorithm is specifically realized by the following steps: calculating the probability of selecting each port capable of forwarding the interest packet according to the distributed weight, then accumulating the probability of each port to obtain the accumulated probability of the port, wherein the accumulated probability is represented by line segments with different lengths, the line segments are combined into a straight line, the length of the straight line is 1, the longest line segment represents the maximum probability of selecting the port, then randomly generating a number in an interval [0,1], and the interval where the number falls is the forwarding port selected by the strategy, and forwarding data through the forwarding port.
The eFIB table comprises a Face field, a State field, an UpdateTime field and a W field;
the Face field indicates which ports the interest packet can be forwarded by, and is determined by the path discovery phase;
the State field is used for recording the State of a path link and is maintained by link information carried by a NACK packet fed back by an upstream node;
the UpdateTime field indicates the survival time of the entry for guaranteeing the validity of the entry;
the W field represents a weight value, which is determined by the link status.
The update method of the UpdateTime field is that each port ensures validity through a time-to-live, before the time-to-live expires, if the port is a valid port, the time-to-live is extended, otherwise, the port is deleted, and the extended time is:
T update indicates the survival time after update, T now Representing the current update time, t before The forwarding time of the interest packet is last time, t is the forwarding time of the current interest packet, and lambda is an adjustment parameter.
The weight assigned to each port is assigned according to 4 link states, and the weight assignment strategy is as follows:
w s1 ＝αw s2 ＝βw s3 ＝γw s4 (2)
w s1 、w s2 、w s3 、w s4 representing the weight values of the idle, normal, busy and congested link states; alpha, beta and gamma are ratio parameters of four weight values; f. of n Representing the number of ports of the selectable path;
representing the weight of the j port, and the sum of the weight values of the ports is 1.
Calculating the probability of selecting each port capable of forwarding the interest packet according to the assigned weight as
cumulative probability of a port is
Wherein the content of the first and second substances,
is the probability that port k is selected.
The invention has the beneficial effects that: the equitable multi-path strategy based on the wheel disc algorithm is superior to a congestion control strategy of an NDN (named data networking) network in overall performance, compared with a flooding route strategy and an optimal path route strategy which are already realized by an NDN project group, because the equitable multi-path strategy based on the wheel disc algorithm is used for discriminating the link state of the network and adopting a corresponding path selection strategy, the interest packet can be effectively prevented from being forwarded to a network congested link, network resources are utilized to the maximum extent, the overall performance of the network is improved, and the problems of data redundancy and unfairness in flow distribution in the network can be effectively solved.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a diagram of a bottleneck topology;
FIG. 3 is a link utilization diagram for paths R1-R2;
FIG. 4 is a graph of link utilization for paths R1-R3-R2;
FIG. 5 is a diagram of a complex network topology;
FIG. 6 is a graph comparing the amount of data received by a consumer per second;
fig. 7 is a graph comparing the timeout data amount of the interest packet.
Detailed Description
In order to fully improve the influence of a link state on the routing of an interest packet, a fair multi-path strategy based on a wheel disc algorithm is provided, an FIB field is expanded on the basis that an original NDN architecture is not changed, and an expanded eFIB structure is shown in table 1. The multipath implementation comprises 3 steps: path discovery, congestion control, and multi-path forwarding strategies. The main task of the path discovery stage is to confirm an effective path, the detection process is to send a route-exploring interest packet at an edge node, the intra-network node broadcasts the special interest packet after receiving the special interest packet, under the condition that the intra-network node does not have any cache content, the interest packet request reaches a content producer, the content producer returns a corresponding data packet, and the data packet carries basic information such as round-trip delay, hop count and the like. After the intermediate routing node receives the special data packet for the first time, marking a receiving port, and marking all ports of the edge node receiving the return data packet as effective ports; the research on the multipath congestion control is mainly combined with a forwarding strategy to complete the distribution of data streams on a plurality of transmission paths. And under the condition that the multi-path discovery is completed, acquiring the congestion state of each link through congestion monitoring, and then distributing corresponding weight to the congestion states of different links to update the expanded forwarding information table eFIB field. When the interest packet needs to be forwarded by the routing node, a wheel algorithm is used for calculating the weight field of the forwarding information table to determine a forwarding path, each available port is selected by the algorithm at an opportunity and is not required to be forwarded by each available port, the defect of high cost of a flooding strategy is avoided, and meanwhile the problem of unfair flow distribution caused by an optimal path routing strategy can be well solved.
Table 1 improved forwarding information table ebb
The invention provides a fair multi-path strategy based on a wheel disc algorithm, which mainly comprises two parts of updating and maintaining a field of an extended forwarding information table and a wheel disc selection path algorithm.
As shown in fig. 1, when the routing node receives a NACK packet signal fed back by an upstream node, the value of the ebbs table field is updated according to the link state information carried by the NACK packet. When an interest package arrives, if no record about the interest package exists in a CS (cache storage) and a PIT (unsatisfied interest package table), a routing strategy based on a roulette algorithm is executed.
1. Updating and maintaining fields in an eFIB table
The Face field indicates which ports the packet of interest for that prefix may be forwarded by, as determined by the path discovery phase.
The State field is used for recording the path link State and is maintained by NACK packet information fed back by the upstream node.
The UpdateTime field indicates the time-to-live of the entry for guaranteeing the validity of the entry.
The W field represents a weight value, which is determined by the link status.
The edge node periodically maintains the fields of the eFIB, each port ensures validity through a lifetime, before the lifetime expires, if the port is used (namely, the path discovery phase is marked as a valid port), the lifetime is prolonged, otherwise, the entry where the port is located is deleted, and the prolonged time is as follows:
T update indicates the survival time after update, T now Representing the current update time, t before The forwarding time of the interest packet is last time, t is the forwarding time of the current interest packet, and lambda is an adjustment parameter. t-t before The smaller the entry is, the slower the survival time of the entry increases, which is to prevent the port from frequently sending the interest packet for a certain period of time, so that the survival time of the entry is too long, and then for a certain period of time, the entry rarely sends the interest packet or even the interest packet is not sent from the port, but because the survival time of the entry is too long, the entry is not deleted for a certain period of time, so that the forwarding table space is occupied.
The link state of the forwarding path corresponding to each table entry in the extended forwarding information table has 4 types: idle, normal, busy and congested, and the w-field information is distributed according to the above 4 link states, and the weight distribution strategy is as follows:
w s1 ＝αw s2 ＝βw s3 ＝γw s4 (2)
w s1 、w s2 、w s3 、w s4 representing the weight values of the idle, normal, busy and congestion four link states; alpha, beta and gamma are ratio parameters of four weight values; f. of n Representing the number of interfaces of the selectable path;
and representing the weight of the j port, wherein the sum of the weight values of the ports meeting the forwarding condition under the prefix interest packet is 1.
2. Roulette selection strategy
After fields of the FIB table entry are filled, when a received interest packet needs to be forwarded by a routing node, a wheel selection strategy is triggered. The idea of the roulette algorithm is to give each individual the opportunity to be selected, preventing individuals with a lower probability from being eliminated. The wheel disc selection strategy is characterized in that according to a weight field value distributed by a FIB table link state, an interface is selected to be forwarded by using a wheel disc algorithm, the specific execution process is that the probability of each port being selected is calculated according to the distributed weight, then the probability of each port is accumulated to obtain the accumulated probability of the port, the accumulated probability passes through line segments with different lengths, the line segments are combined into a straight line, the length of the straight line is 1, in the straight line, the longest line segment represents the maximum probability of port being selected, then a digit is randomly generated in a range [0,1], and the range where the digit falls is the forwarding port selected by the strategy. The specific calculation formula is as follows:
P fi probability of being selected for each port; q fi The cumulative probability of being selected for port i.
3. Algorithm steps
1) When a routing node receives a NACK packet fed back by an upstream node, updating a field value in an eFIB table according to link information carried by the NACK to obtain each port capable of forwarding the interest packet and distributed weight, wherein the weights can be realized by formulas (1), (2) and (2);
2) when an interest packet reaches a routing node and a relevant record of the interest packet does not exist in CS and PIT of the routing node, a forwarding path needs to be selected in FIB, and a path selection strategy based on a wheel disc is triggered;
3) the path selection strategy of the wheel disc algorithm is specifically realized by the following steps: calculating the probability of selecting each port capable of forwarding the interest packet according to the distributed weight, then accumulating the probability of each port to obtain the accumulated probability of the port, wherein the accumulated probability is represented by line segments with different lengths, the line segments are combined into a straight line, the length of the straight line is 1, the longest line segment represents the maximum probability of selecting the port, then randomly generating a number in an interval [0,1], and the interval where the number falls is the forwarding port selected by the strategy, and forwarding data through the forwarding port.
A path within the survival time can be found according to the w field value and the selection probability calculation in the eFIB by using formula 4 and formula 5, and data is forwarded through the port.
4) Finishing the algorithm;
5) the next routing node proceeds to step 1)2) 3).
4. Results of the experiment
The experiment uses the ndnSIM simulator to evaluate the path selection strategy and compare it with the congestion control strategy of the NDN itself. For the purpose of this experiment, a bottleneck link is preset, which cannot provide enough bandwidth. In order to verify that the algorithm can show self-adaptive capacity to network congestion, an in-network cache function is not enabled in the experiment. When the simulation is started, the network is allowed to normally operate for a period of time, and each routing node is allowed to obtain initial parameters.
The simulation experiment is completed in two different network topological structures, namely a bottleneck type network topological structure and a complex type network topological structure. The bottleneck network verifies the working mode difference between the strategy provided by the bottleneck network and the NDN congestion control strategy; the complex network topology verifies the difference between the two in network throughput and the number of request timeouts.
1. Bottleneck network
The bottleneck network topology structure adopts 11 network nodes, and the topology structure is shown in fig. 2.
Nodes C1, C2, C3 and C4 are content consumer nodes in the network, P1, P2, P3 and P4 are content producer nodes, and R1, R2 and R3 represent routing nodes in the network. The number next to a node represents the port ID of that node.
The link bandwidth between the content consumer node and the routing node in the topology and between the routing node and the content producer node is set to be 10Mbps, the link bandwidth between the routing nodes is 1Mbps, the queue length of a buffer area of the routing node is 25, the size of a data packet is 1024bytes, the size of an interest packet is 16bytes, and a fair multi-path strategy based on a wheel algorithm is used for a forwarding strategy of the routing node.
Experiments mainly compare the link utilization of R1, R2, and R1, R3, R2, as shown in fig. 3. The abscissa represents the sending time of the interest packet, the ordinate represents the link utilization rate, and the link utilization rate is calculated by dividing the flow size of the current link by the link bandwidth.
As can be seen from fig. 3, the NDN congestion control and the roulette-based algorithm both pre-use the path between R1 and R2 in the simulation because the path between R1 and R2 is higher in priority. As can be seen from fig. 4, the transmission path through R3 is used earlier based on the roulette algorithm, because the roulette algorithm can select an appropriate link for forwarding in advance based on the network link state, and therefore the path between R1, R3 and R2 is used earlier, which shows that the roulette algorithm has higher sensitivity for the selection of the path. The link utilization jitter of the NDN congestion strategy is because the AIMD algorithm adopted by NDN reduces the transmission window of the interest packet by half when the link is congested.
2. Complex network topology
A complex network topology is shown in fig. 5. The topological structure comprises 22 nodes and 36 links, the bandwidth among the links is set to be 1Mbps, the time delay is set to be 10ms, the survival time of an interest packet is 3s, the retransmission time is 50ms, the size of a buffer queue of the routing node is set to be 20, each consumer node downloads required data to a corresponding producer node, the content generated by the producer node has a common prefix, and the simulation result is as follows:
as can be seen from fig. 6, using a round-robin based algorithm fair multi-path policy performs better in terms of the number of packets received per second by the consumer than using NDN's own congestion control. As can be seen from fig. 7, the congestion control using the carousel-based algorithm is smaller than using NDN itself, in the number of interest packet timeouts, and remains a smoother state over time. The NDN congestion control generates several large fluctuations, because when the link is congested, the node will rapidly reduce the sending rate of the interest packet, and then the timeout number of the interest packet in a period of time is rapidly reduced. When the node receives the corresponding data packet, the sending rate is rapidly increased, and the overtime quantity of the interest packet is increased. And the equitable multi-path strategy based on the wheel disc algorithm is based on the congestion state of the link, other suitable paths are selected in time by utilizing the wheel disc algorithm, and the use of the congested link is avoided, so that the overtime quantity of the interest packets is lower than that of NDN congestion control.
From simulation results, the equitable multi-path strategy based on the roulette algorithm is superior to the congestion control strategy of the NDN network in overall performance. The link state of the network is discriminated based on the wheel disc algorithm fair multi-path strategy, and a corresponding path selection strategy is adopted, so that the interest packet can be effectively prevented from being forwarded to the congested link of the network, network resources are utilized to the maximum extent, and the overall performance of the network is improved.
Claims (4)
1. The equitable multipath routing method based on the wheel disc algorithm is characterized in that:
step one, when a routing node receives a NACK packet fed back by an upstream node, updating a field value in an eFIB table according to link information carried by the NACK packet to obtain each port capable of forwarding an interest packet and an allocated weight;
the eFIB table comprises a Face field, a State field, an UpdateTime field and a W field;
the Face field indicates which ports the interest packet can be forwarded by, and is determined by the path discovery phase;
the State field is used for recording the State of a path link and is maintained by link information carried by a NACK packet fed back by an upstream node;
the UpdateTime field indicates the survival time of the entry for guaranteeing the validity of the entry;
the W field represents weight, and the weight value is determined by the link state;
step two, when the interest packet reaches the routing node and the related record of the interest packet does not exist in the CS and the PIT of the routing node, selecting a forwarding path in the eFIB and triggering a path selection strategy based on a wheel disc algorithm;
step three, the path selection strategy of the wheel disc algorithm is specifically realized by the following steps: calculating the probability of selecting each port capable of forwarding the interest packet according to the distributed weight, accumulating the probability of each port to obtain the accumulated probability of the port, wherein the accumulated probability is represented by line segments with different lengths, the line segments are combined into a straight line, the length of the straight line is 1, the longest line segment represents the maximum probability of selecting the port, then randomly generating a number in an interval [0,1], and the interval where the number falls is the forwarding port selected by the strategy at this time, and forwarding data through the forwarding port; the weight assigned to each port is assigned according to 4 link states, and the weight assignment strategy is as follows:
w s1 ＝αw s2 ＝βw s3 ＝γw s4 (2)
w s1 、w s2 、w s3 、w s4 representing the weight values of the idle, normal, busy and congested link states; alpha, beta and gamma are ratio parameters of four weight values; f. of n Representing the number of ports of the selectable path;
representing the weight of the j port, and the sum of the weight values of the ports is 1.
2. The round-robin based fair multi-path routing method of claim 1 wherein: the update method of the UpdateTime field is that each port ensures validity through a time-to-live, before the time-to-live expires, if the port is a valid port, the time-to-live is extended, otherwise, the port is deleted, and the extended time is:
T update indicating the time-to-live after the update,T now representing the current update time, t before The forwarding time of the interest packet is last time, t is the forwarding time of the current interest packet, and lambda is an adjustment parameter.
3. The round-robin based fair multi-path routing method of claim 1 wherein: calculating the probability of selecting each port capable of forwarding the interest packet according to the assigned weight as
4. The round-robin based fair multipath routing method of claim 3 wherein: cumulative probability of a port is
Wherein the content of the first and second substances,
is the probability that port k is selected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110390824.4A CN113162863B (en) | 2021-04-12 | 2021-04-12 | Fair multipath routing method based on wheel disc algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110390824.4A CN113162863B (en) | 2021-04-12 | 2021-04-12 | Fair multipath routing method based on wheel disc algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113162863A CN113162863A (en) | 2021-07-23 |
| CN113162863B true CN113162863B (en) | 2022-08-02 |
Family
ID=76890061
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110390824.4A Active CN113162863B (en) | 2021-04-12 | 2021-04-12 | Fair multipath routing method based on wheel disc algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113162863B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116455821A (en) * | 2023-06-19 | 2023-07-18 | 中南民族大学 | Rate-based multipath perceived congestion control method in named data network |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103841018A (en) * | 2012-11-26 | 2014-06-04 | 华为技术有限公司 | Content centric network multiport forwarding method and router |
| CN106332211A (en) * | 2016-08-18 | 2017-01-11 | 北京邮电大学 | Routing selection method and device |
| CN108924051A (en) * | 2018-07-04 | 2018-11-30 | 河南科技大学 | Information centre's network interest package transmitting method based on enhancing study |
| CN112039781A (en) * | 2020-09-09 | 2020-12-04 | 北京同创神州航天科技有限公司 | Named data network forwarding method based on flow control |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2460046C (en) * | 2001-09-10 | 2014-06-10 | Igt | Method for developing gaming programs compatible with a computerized gaming operating system and apparatus |
| US9729616B2 (en) * | 2014-07-18 | 2017-08-08 | Cisco Technology, Inc. | Reputation-based strategy for forwarding and responding to interests over a content centric network |
| US9577949B2 (en) * | 2014-08-11 | 2017-02-21 | Cisco Technology, Inc. | Maintaining named data networking (NDN) flow balance with highly variable data object sizes |
| CN104301305B (en) * | 2014-09-17 | 2017-10-24 | 北京大学深圳研究生院 | Interest bag is forwarded under information centre's network method and forwarding terminal |
| CN108881048B (en) * | 2018-08-23 | 2019-06-14 | 北京理工大学 | A kind of name data network congestion control method based on intensified learning |
| CN111770024B (en) * | 2020-07-03 | 2022-09-23 | 重庆邮电大学 | Interest packet forwarding interface selection method in named data network |
-
2021
- 2021-04-12 CN CN202110390824.4A patent/CN113162863B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103841018A (en) * | 2012-11-26 | 2014-06-04 | 华为技术有限公司 | Content centric network multiport forwarding method and router |
| CN106332211A (en) * | 2016-08-18 | 2017-01-11 | 北京邮电大学 | Routing selection method and device |
| CN108924051A (en) * | 2018-07-04 | 2018-11-30 | 河南科技大学 | Information centre's network interest package transmitting method based on enhancing study |
| CN112039781A (en) * | 2020-09-09 | 2020-12-04 | 北京同创神州航天科技有限公司 | Named data network forwarding method based on flow control |
Non-Patent Citations (3)
| Title |
|---|
| "A Rate-Based Multipath-Aware Congestion Control Mechanism in Named Data Networking";Shangru Zhong等;《2017ISPA/IUCC》;20180528;全文 * |
| "命名数据网络中自适应转发策略的研究";胡涛;《南京邮电大学硕士学位论文》;20210215;全文 * |
| 基于蚁群算法的ICN路由算法;龚娉婷等;《计算机工程与设计》;20200916(第09期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113162863A (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8000239B2 (en) | Method and system for bandwidth allocation using router feedback | |
| CN106027416B (en) | A kind of data center network traffic scheduling method and system based on space-time combination | |
| US8897130B2 (en) | Network traffic management | |
| US8644164B2 (en) | Flow-based adaptive private network with multiple WAN-paths | |
| Lee et al. | A workload-based adaptive load-balancing technique for mobile ad hoc networks | |
| CN106533960A (en) | Data center network routing method based on Fat-Tree structure | |
| CN103685054A (en) | Multipath load balancing method based on service awareness | |
| WO2000024164A1 (en) | Method and apparatus for network control | |
| CN107070794A (en) | A kind of low rail information network optimal network benefit delay constraint method for routing | |
| CN105245448A (en) | Multi-service and multi-path routing optimization method in named data network | |
| US11863322B2 (en) | Communication method and apparatus | |
| CN110351187A (en) | Data center network Road diameter switches the adaptive load-balancing method of granularity | |
| Nelakuditi et al. | A localized adaptive proportioning approach to QoS routing | |
| CN113162863B (en) | Fair multipath routing method based on wheel disc algorithm | |
| CN109088822A (en) | Data traffic retransmission method, device, system, computer equipment and storage medium | |
| CN111585911A (en) | Method for balancing network traffic load of data center | |
| US8289851B2 (en) | Lightweight bandwidth-management scheme for elastic traffic | |
| Hertiana et al. | A joint approach to multipath routing and rate adaptation for congestion control in openflow software defined network | |
| AU2006324005B2 (en) | Remotely control method and device for congesting in a packet mode a flow of links in a telecommunication network | |
| CN105072046A (en) | Delay tolerant network congestion prevention method based on data concurrence and forwarding by token control node | |
| Dimitriou et al. | Effective buffer and storage management in DTN nodes | |
| JP2933905B1 (en) | Multicast communication system | |
| Curran et al. | The effects of badly behaved routers on Internet congestion | |
| CN117081984B (en) | Route adjustment method and device and electronic equipment | |
| CN114553783B (en) | Load balancing method for self-adaptive regulation of cell granularity of data center network |
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 |