CN103716208A - Network management method and system supporting elephant flow, interchanger and network - Google Patents
Network management method and system supporting elephant flow, interchanger and network Download PDFInfo
- Publication number
- CN103716208A CN103716208A CN201310752282.6A CN201310752282A CN103716208A CN 103716208 A CN103716208 A CN 103716208A CN 201310752282 A CN201310752282 A CN 201310752282A CN 103716208 A CN103716208 A CN 103716208A
- Authority
- CN
- China
- Prior art keywords
- network
- path
- elephant
- weight
- network switch
- 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
- 241000406668 Loxodonta cyclotis Species 0.000 title claims abstract description 138
- 238000007726 management method Methods 0.000 title abstract description 45
- 238000000034 method Methods 0.000 claims abstract description 16
- 238000004891 communication Methods 0.000 claims abstract description 14
- 230000009471 action Effects 0.000 claims description 17
- 230000000875 corresponding effect Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 6
- 230000000737 periodic effect Effects 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 10
- 238000006116 polymerization reaction Methods 0.000 description 4
- 239000012792 core layer Substances 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010000234 Abortion spontaneous Diseases 0.000 description 1
- 101100448334 Arabidopsis thaliana GG1 gene Proteins 0.000 description 1
- 101100495256 Caenorhabditis elegans mat-3 gene Proteins 0.000 description 1
- 241000283080 Proboscidea <mammal> Species 0.000 description 1
- 102100033121 Transcription factor 21 Human genes 0.000 description 1
- 101710119687 Transcription factor 21 Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000010410 layer Substances 0.000 description 1
- 208000015994 miscarriage Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 208000000995 spontaneous abortion Diseases 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
本发明公开了一种支持大象流的网络管理方法、网络管理系统、网络交换机和通信网络，其中支持大象流的网络管理方法包括：在检测到网络中的大象流时，提取大象流的源地址和目的地址，获取源地址和目的地址之间的多个路径，利用所述路径的网络负载和网络拓扑结构计算支持大象流的对应路径的权重；网络管理系统生成网络交换机的转发表，并将转发表发送至网络交换机，所述转发表包括支持大象流的对应路径的权重和网络交换机的操作指令，供网络交换机执行所述转发表，将大象流的流量按照路径的权重分发。采用上述方法与系统装置能够使存在大象流的通信网络保持流量均衡，有效提高网络性能与资源利用率。
The invention discloses a network management method supporting elephant flow, a network management system, a network switch and a communication network, wherein the network management method supporting elephant flow includes: when detecting an elephant flow in the network, extracting the elephant flow The source address and destination address of the flow, obtain multiple paths between the source address and the destination address, use the network load and network topology of the path to calculate the weight of the corresponding path supporting the elephant flow; the network management system generates the network switch forwarding table, and sending the forwarding table to the network switch, the forwarding table includes the weight of the corresponding path that supports the elephant flow and the operation instruction of the network switch, for the network switch to execute the forwarding table, and the flow of the elephant flow according to the path weight distribution. The above method and system device can maintain traffic balance in a communication network with elephant flow, and effectively improve network performance and resource utilization.
Description
Technical field
The present invention relates to network communication field, particularly a kind of network management, network management system, the network switch and communication network of supporting elephant stream.
Background technology
Existing multi-path routing technology, as equal cost multipath (Equal-Cost Multi path Routing, ECMP) etc., is used the static Hashing mapping techniques based on data flow stream to be divided on many available equative routes at random.Yet no matter all data flow, be their size (byte number), or their duration be all not identical, so this simple strategy is not enough to prevent the generation of network focus.This strategy based on Hash mapping does not only have elephant stream (data flow that byte number is large) just effective in the situation that only there is mouse stream (data flow that byte number is little) in network.Particularly, two long-standing elephant streams may all be mapped to same path within a very long time, therefore the congested meeting causing declines network performance, and meanwhile in network, may also have the idle capacity on backup path available, therefore use existing multi-path routing technology, in the situation that there is elephant stream, easily cause the decline of network performance.
Summary of the invention
The invention provides a kind of network management, network management system, the network switch and communication network of supporting elephant stream, when elephant stream being detected, utilize the offered load in path and the weight that network topology structure is calculated the respective path of supporting elephant stream, and form corresponding transmitting, the network switch can be distributed according to weight according to the flow of transmitting elephant stream, has guaranteed the performance of network.
The invention provides a kind of network management of supporting elephant stream, the method comprises:
When the elephant in network being detected is flowed, extract source address and the destination address of elephant stream, obtain a plurality of paths between source address and destination address, utilize the offered load in described path and the weight that network topology structure is calculated the respective path of supporting that elephant is flowed;
Transmitting of network management system generating network switch, and will transmit and be sent to the network switch, described transmitting comprises the weight of respective path and the operational order of the network switch of supporting elephant stream, described in carrying out for the network switch, transmit, the flow of elephant stream is distributed according to the weight in path.
Preferably, described stream table comprises destination address and corresponding action fields, and described action fields is pointed to respectively a group table; Described group of table comprises respectively a plurality of action items, and described action items comprises respectively the weight in path and the operational order of the network switch.
Preferably, the method also comprises: generate LLDP message and be regularly issued to the network switch; Receive the LLDP message that the network switch is uploaded, by the LLDP message receiving, obtain network topology structure; The statistics of the periodic polling network switch, to obtain the load of the network switch and the load of network link.
Preferably, the described offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream utilized, comprising:
Obtain the set that represents all available paths between source address and destination address, the number that makes described path is I; To each path i(0≤i≤I-1) the following operation of execution:
The number that makes it subpath comprising is J
i; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); Make λ
kji(0≤k≤K
ji-1) be the load of link k; Make c
kji(0≤k≤K
ji-1) be the capacity of link k;
Calculate the link utilization of each link:
lU
kjilink utilization for link k;
The weight of calculating path:
w
i∈ 0,1 ..., 9} is the weight of path i.
The present invention also provides a kind of network management system, and described network management system comprises:
Elephant flow path computing unit, for when the elephant stream of network being detected, extract source address and the destination address of elephant stream, obtain a plurality of paths between source address and destination address, utilize the offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream;
Path management unit, for transmitting of generating network switch, and will transmit and be sent to the network switch, described transmitting comprises the weight of respective path and the operational order of the network switch of supporting elephant stream, described in carrying out for the network switch, transmit, the flow of elephant stream, according to the weight distribution in path, preferably, is distributed to next-hop network node corresponding to described path.
Preferably, described network management system also comprises: network topology unit, for generating LLDP message and being regularly issued to the network switch, and receive the LLDP message that the network switch is uploaded, and by the LLDP message receiving, obtain network topology structure; Offered load monitoring means, for the statistics of the periodic polling network switch, to obtain the load of the network switch and the load of network link.
Preferably, described elephant flow path computing unit, for utilizing the weight of offered load and the network topology structure calculating path in described path;
Wherein, described elephant flow path computing unit, for obtaining the set that represents all available paths between source address and destination address, the number that makes described path is I; Described elephant flow path computing unit, is further used for each path i(0≤i≤I-1) carry out following operation, with the weight of calculating path:
The number that makes it subpath comprising is J
i; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); The load that makes link k is λ
kji(0≤k≤K
ji-1); The capacity that makes link k is c
kji(0≤k≤K
ji-1);
Calculate the link utilization of each link:
lU
kjilink utilization for link k;
The weight of calculating path:
w
i∈ 0,1 ..., 9} is the weight of path i.
The present invention also provides a kind of network switch, comprising: elephant stream reports unit, for choosing from the data flow receiving, is labeled as the data flow of elephant stream and selected data flow is reported to network management system; Retransmission unit, for receiving and carry out transmitting of network management system transmission, according to the weight distribution in transmitting, preferably, is distributed to next-hop network node corresponding to path by the flow of elephant stream.
The present invention also provides a kind of communication network, comprising: network management system as above; And signal is connected in the network switch as above of described network management system.
The beneficial effect of the embodiment of the present invention is: by according to the load information in each path of network, calculate the weight of the respective path of supporting elephant stream, then switch, according to the weight in each path by the traffic distribution of elephant stream, can improve network performance.By detecting elephant stream and carrying out the processing of the weight of calculating path for elephant stream, for mouse stream, use basic routing algorithm, can when improving network performance, control resource overhead.
Accompanying drawing explanation
Fig. 1 is a kind of flow chart of supporting the network management of elephant stream of one embodiment of the invention;
Fig. 2 is the detailed maps of transmitting of the present invention;
Fig. 3 is the structured flowchart of a kind of communication network of one embodiment of the invention;
Fig. 4 is the structured flowchart of a kind of network switch of one embodiment of the invention;
Fig. 5 is the structured flowchart of a kind of network management system of one embodiment of the invention;
Fig. 6 (a) is for being used the schematic diagram of a kind of application scenarios of equal cost multipath routing algorithm;
The schematic diagram of a kind of application scenarios that Fig. 6 (b) is one embodiment of the invention;
Fig. 7 is for being used the present invention and the contrast schematic diagram that uses the network delay of prior art;
Fig. 8 is for being used the present invention and the contrast schematic diagram that uses the network throughput of prior art;
Fig. 9 is for being used the present invention and the contrast schematic diagram that uses the link utilization of prior art.
Embodiment
For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing, embodiment of the present invention is described further in detail.
Technical conceive of the present invention is mainly, supports the weight (load light path weight value higher) of the respective path of elephant stream by calculating, and elephant stream can be distributed according to the load of network path.The path weight value light due to load is high, and the flow therefore flowing to the elephant of the light path distribution of load is large, thereby reaches load balancing and the technique effect that improves the overall performance of network.
Fig. 1 is a kind of flow chart of supporting the network management of elephant stream of one embodiment of the invention.As shown in Figure 1, the method comprises the following steps:
When S100, the stream of the elephant in network being detected, extract source address and the destination address of elephant stream, obtain a plurality of paths between source address and destination address, utilize the offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream.
Transmitting of S200, generating network switch, and will transmit and be sent to the network switch, described transmitting comprises the weight of respective path and the operational order of the network switch of supporting elephant stream, described in carrying out for the network switch, transmit, the flow of elephant stream is distributed according to the weight in path, preferably, the flow of elephant stream is distributed to next-hop network node corresponding to described path according to the weight in path.
By calculating, support the weight of the respective path of elephant stream, generation comprise support the weight of respective path of elephant stream and the operational order of the network switch transmit and be sent to switch, make the switch can be according to transmitting elephant stream according to the load distribution of network path.Owing to transmitting the operational order that comprises the network switch, so the network switch can be determined next-hop network node corresponding to path (switch or all the other network equipments) and data flow is sent to this network node according to operational order; And owing to transmitting the weight that comprises the respective path of supporting elephant stream, so the network switch can be according to the weight distribution elephant stream in path.Preferably, the flow of the elephant of distribution stream is directly proportional to weight.
Further, because accounting for the very large and above-mentioned offered load in described path and the weight that network topology structure is calculated each path utilized of the ratio of network traffics, the mouse stream in network needs certain calculation resources expense, if adopt said method of the present invention to process easily to cause resource overhead excessive for mouse stream is same.Therefore the present invention does not carry out aforesaid operations for mouse stream.Preferably, for mouse stream, use the mode of equal cost multipath (ECMP) to carry out multipath route.
Further, the method for the elephant stream in Sampling network can have multiple, for example, at network edge switch, detect also mark etc.
Further, the method also comprises: generate LLDP(Link Layer Discovery Protocol, Link Layer Discovery Protocol) message be regularly issued to the network switch; Receive the LLDP message that the network switch is uploaded, by the LLDP message receiving, obtain network topology structure.Particularly, every a predetermined time out period, will call maker function, during each calls, network management device generates LLDP message, and LLDP message is sent to the network switch of its management, the network switch that receives LLDP message forwards the packet to the adjacent network switch, and the network switch that receives the LLDP message being forwarded by all the other network switchs is uploaded to network management device by this LLDP message.Network management device is inferred link level connection by the LLDP message receiving, and then obtains network topology structure.Preferably, network topology structure is stored in an adjacency list, and network topology structure can be expressed as connected graph G=(V, E), wherein V is the set of node, E is the set of directed edge.
Further, the method also comprises: the statistics of the periodic polling network switch, and to obtain the load of the network switch and the load of network link.Particularly, can inquire about termly, consolidate and store the statistics of all switches, these data, can be used in the load of switch and the load of link of calculating communication network, and then for the calculating of elephant flow path.Preferably, from network on all switch, the statistical information of collecting based on each table, each stream and each port, is stored in internal memory with the form of snapshot object.Each snapshot is usingd a time series number as sign, and it increases 1 after each time interval, thereby can identify the network condition of special time.In order to save memory space, only have 2 nearest snapshots to be maintained in internal memory.
In the present embodiment, the described offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream utilized, is used following weighting routing algorithm, specifically comprises:
Obtain the set that represents all available paths between source address and destination address, the number that makes described path is I, particularly, the data flow that is marked as elephant stream that can detect from switch, obtain source address and the destination address of data flow, and then use various network topologies algorithm to obtain available path.Further, to each path i(0≤i≤I-1) the following operation of execution:
The number that makes it subpath comprising is J
i; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); Make λ
kji(0≤k≤K
ji-1) be the load of link k; Make c
kji(0≤k≤K
ji-1) be the capacity of link k.
So, can by following formula (1), calculate the link utilization of each link, wherein, LU
kjilink utilization for link k.
Can be by the weight of following formula (2) calculating path, wherein, w
i∈ 0,1 ..., 9} is the weight of path i.
In the present embodiment, the assignment of traffic ratio of the respective path of the support elephant stream flowing out from some switches, is directly proportional to the weight in described path.For example, the path of the elephant stream of support flowing out from some switches has four, and weight is respectively 4,6,7,3, the flow of elephant stream is distributed according to 20%, 30%, 35%, 15% ratio.
In the present embodiment, the path load that weight is high is light, the path load weight that weight is low, the flow flowing due to the elephant of distribution is directly proportional to weight, if be the flow that elephant stream is not distributed in 0 Ze Duigai path, path so there is weight, all the other paths are according to the flow of the ratio distribution elephant stream being directly proportional to weight.For example, suppose to exist four paths, weight is respectively 0,2,3,5, the flow of elephant stream is distributed according to 0,20%, 30%, 50% ratio, certainly, in the present embodiment, also can directly get rid of weight and be 0 path (consulting the Exclude the overutilized path in false code), make not utilize the path of load too high to forward the flow of elephant stream, thereby the performance in the path of load too high is improved faster.
The path weight value light due to load is high, and the flow therefore flowing to the elephant of the light path distribution of load is large, thereby reaches load balancing, has improved the overall performance of network.
For weighting routing algorithm of the present invention is described better, provide the false code of weighting routing algorithm, false code is as follows.
Fig. 2 is the detailed view of transmitting of the present invention.Described transmitting comprises stream table (Flow Table) and group table (Group Table).
Described stream table (Flow Table) comprises destination address (IP Dst) and corresponding action fields (Instructions), and described action fields is pointed to respectively a group table, particularly, by group Table I D(Group ID) group table of sensing.
Described group of table (Group Table) comprises respectively a plurality of action items (Action Buckets), and described action items comprises respectively the weight (Weight) in path and the operational order (Action) of the network switch.
More specifically, the present embodiment is shown by stream and being used in conjunction with of group table, to support multipath route.Multipath route is to realize by the action fields in stream table is pointed to a group table.Each group table is comprised of a set of group of set of actions, and each group set of actions comprises a set of by the action being applied in the data flow that the match is successful.Each set has a weight field, and it has defined by this gathers the share that handled flow accounts for whole group of flow that table is processed.Action fields in the stream table of the elephant stream that each is detected can be pointed to a group table.Group table comprises set of actions, and each set is relevant to a paths, and elephant stream can arrive destination address via these paths.
Fig. 3 is the structured flowchart of a kind of communication network of one embodiment of the invention.Communication network comprises network management system 3, a plurality of network switch 2, only schematically shows a network switch 2 among Fig. 3, and Practical Project is not as limit.
Fig. 4 is the structured flowchart of a kind of network switch of one embodiment of the invention.The network switch 2, comprises that elephant stream reports unit 21 and retransmission unit 22.
Elephant stream reports unit 21, for choosing from the data flow receiving, is labeled as the data flow of elephant stream and selected data flow is reported to network management system 3.Particularly, be by selecting elephant stream packets (Marked Packets) and selected elephant stream packets being reported to network management system 3, for network management system 3 according to packet and the weight of calculating path.In the present embodiment, the unmarked packet (Unmarked Packets as shown in the figure) for elephant stream and the packet that to be labeled as elephant stream packets (Marked Packets as shown in the figure) be the network switch 2 receives from the network terminal or remaining network switch 2.
Fig. 5 is the structured flowchart of a kind of network management system of one embodiment of the invention.Network management system 3 comprises elephant flow path computing unit 31 and path management unit 32.
Elephant flow path computing unit 31, when the elephant in network being detected is flowed, the source address and the destination address that extract elephant stream, obtain a plurality of paths between source address and destination address, utilizes the offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream.
The present invention is by the weight of elephant flow path computing unit 31 calculating paths, path management unit 32 generate comprise the weight in path and the operational order of the network switch transmit and be sent to the network switch 2, make the network switch 2 can be according to transmitting elephant stream according to the load distribution of network path.Owing to transmitting the operational order that comprises the network switch, so the network switch 2 can be determined next-hop network node corresponding to path (switch or all the other network equipments) and data flow is sent to this network node according to operational order; And owing to transmitting the weight that comprises path, so the network switch 2 can distribute elephant stream according to weight, the data volume of distribution is directly proportional to weight.
Further, because the mouse stream in network accounts for the very large and above-mentioned offered load in described path and the weight that network topology structure is calculated each path utilized of the ratio of network traffics, need certain calculation resources expense, if carry out above-mentioned processing by elephant flow path computing unit 31 equally for mouse stream, easily cause resource overhead excessive.Therefore the present invention does not carry out the operation of above-mentioned calculating weight by elephant flow path computing unit 31 for mouse stream, preferably, for mouse stream, uses the mode of equal cost multipath (ECMP) to carry out multipath route.
The network management system 3 of the present embodiment is not provided for the device of the elephant stream in Sampling network.It will be understood by a person skilled in the art that, in all the other embodiment, network management system 3 can be set up the device for detection of the elephant stream in network.
Further, network management system 3 also comprises network topology unit 33 and offered load monitoring means 34.
Offered load monitoring means 34, for the statistics of the periodic polling network switch, to obtain the load of the network switch and the load of network link.Particularly, the statistics of all network switchs 2 can be inquired about, consolidates and be stored to offered load monitoring means 34 termly, these data, can be used in the load of the network switch 2 and the load of link of calculating communication network, and then for the calculating of elephant flow path.Preferably, from network on all network switch 2, the statistical information of collecting based on each table, each stream and each port, is stored in internal memory with the form of snapshot object.Each snapshot is usingd a time series number as sign, and it increases 1 after each time interval, thereby can identify the network condition of special time.In order to save memory space, only have 2 nearest snapshots to be maintained in internal memory.
In the present embodiment, elephant flow path computing unit 31, utilizes the offered load in path and the weight that network topology structure is calculated the respective path of supporting elephant stream, comprising:
Obtain the set that represents all available paths between source address and destination address, the number that makes described path is I, particularly, can be from being marked as the packet (Marked Packets as shown in Figure 3) of elephant stream to obtain source address and the destination address of data flow, and then use various network topologies algorithm to obtain available path.Further, to each path i(0≤i≤I-1) carry out following operation, with the weight of calculating path:
The number that makes it subpath comprising is Ji; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); Make λ
kji(0≤k≤K
ji-1) be the load of link k; Make c
kji(0≤k≤K
ji-1) be the capacity of link k.
So, can by following formula (1), calculate the link utilization of each link, wherein, LU
kjilink utilization for link k.
Can be by the weight of following formula (2) calculating path, wherein, w
i∈ 0,1 ..., 9} is the weight of path i.
In the present embodiment, through the calculating of elephant flow path computing unit 31, the assignment of traffic ratio of the respective path of the support elephant of flowing out from some switches stream, is directly proportional to the weight in described path.For example, the path of the elephant stream of support flowing out from some switches has four, and weight is respectively 4,6,7,3, the flow of elephant stream is distributed according to 20%, 30%, 35%, 15% ratio.
The path weight value light due to load is high, and the flow therefore flowing to the elephant of the light path distribution of load is large, thereby reaches load balancing, has improved the overall performance of network.
Known from the above mentioned, in the present embodiment, network switch 2(for example, OpenFlow switch) elephant stream report unit 21 by the package forward being labeled to network management system 3.The elephant flow path computing unit 31 of network management system 3 receives after packet, source address and the destination address of data flow will be extracted, then by consultation network topology unit 33, to obtain the topology information between source host and destination host, and calculate all possible shortest path according to these information.Once after all shortest paths are calculated, the statistical information being provided by offered load monitoring means 34 will be used to calculate the load in each path.Then elephant flow path computing unit 31 is by according to the loading condition in current each path, use weighting multi-path routing algorithms, calculate the weight of each path candidate, and the path and the weight information thereof that after calculating, obtain are passed to path management unit 32, the stream entry of being responsible for these paths to convert to network of relation switch 2 by it, is finally issued on the network switch 2.
Fig. 6 (a) is for being used the schematic diagram of a kind of application scenarios of equal cost multipath routing algorithm.The schematic diagram of a kind of application scenarios that Fig. 6 (b) is one embodiment of the invention.Fig. 6 (a) and Fig. 6 (b) have described an application scenarios of data center, and two servers that are arranged in Pod1 transmit two elephant streams to same destination Pod4 simultaneously, and transfer rate is respectively 500Mbps and 300Mbps.This communication pattern is very common in data center, and for example distributed large data handling utility will produce this communication pattern.Suppose port one in core layer network switch 1 treated large batch stream, port utilization ratio has reached 80%, and the port one of the network switch 2 is in light condition, port utilization ratio only has 40%.If adopt equal cost multipath routing algorithm (ECMP), may there is Hash collision and finally all move towards the port one in the very high network switch of load 1 in the large stream that these two speed are large, life period is long.Suppose that link bandwidth is 1Gbps, after the polymerization of two elephant stream, required speed is 800Mbps, and it has surpassed the residual available bandwidth (200Mbps) of the network switch 1 port one.So now will occur congestedly, the service quality that causes producing the application of this elephant stream declines.In this example, if use scheme disclosed by the invention, these two elephant streams just can all reach their required transmission rates.Because the port one of core layer switch 1 can only transmit extra 200Mbps at most, this even can not meet the elephant stream (required speed is 300Mbps) that bandwidth requirement is less.Therefore, best forwarding scheme is, at the network switch AGG1 place of network polymerization layer, two elephants are flowed to polymerization, then the elephant stream of polymerization is cut apart to the port one place of the port one and the switch 2 that are forwarded to core layer switch 1 according to the ratio being inversely proportional to link load, for example, the flow rate of two link transmission is respectively 200Mbps, 600Mbps.Solve like this congestion problems, improved network utilization, can not produce infringement to producing the service quality of the application of elephant stream.
Fig. 7 is for being used the present invention and the contrast schematic diagram that uses the network delay of prior art.
In the present embodiment, use OpenFlow network simulator mininet to build a test environment, topological structure is as shown in Fig. 6 (a), Fig. 6 (b), the network switch is all the OpenFlow switch based on software, its use be Openflow agreement 1.3 editions, the network management of the support elephant stream of this embodiment, especially weighting routing algorithm operates on NOX controller.Meanwhile, equal cost multipath (ECMP) routing algorithm and single path routing algorithm on NOX controller, have been realized in contrast.
Further, simulation produces the flow rate mode that an elephant is flowed and mouse stream coexists.Each data flow is comprised of the sequence of data packet that has identical TCP five-tuple (source IP, source port, object IP address, destination interface, protocol type).The destination address of data flow is evenly distributed in network.Flow rate mode meets Poisson distribution, obtains the test result as shown in Fig. 7, Fig. 8, Fig. 9.
The transverse axis of Fig. 7 is that offered load (unit the is Mbps) longitudinal axis is time delay (unit is microsecond), and the ratio (for example, proportion is 95.6%-99.2%) that the present embodiment can account for by increasing the raw flow of elephant miscarriage all flows increases overall load.Along with the increase of offered load, use the time delay of single path route because serious data flow collision increases very fast.On the contrary, use ECMP and use weighting routing algorithm delay of control just occur the decline of performance in the situation that load is very high effectively.When elephant stream ratio is enough high, use the time delay of ECMP also to start to increase substantially, and add, use the time delay of weighting routing algorithm still to show well, be better than the time delay of using ECMP to obtain.
Fig. 8 is for being used the present invention and the contrast schematic diagram that uses the network throughput of prior art.The transverse axis of Fig. 8 is that offered load (unit the is Mbps) longitudinal axis is throughput (unit is Mbps), uses single path route and uses ECMP to reach rapidly bottleneck, because occur when network in after focus that mass data is coated with, is blocked in network.Use weighting routing algorithm to compare and use ECMP throughput to improve about 10%.
Fig. 9 is for being used the present invention and the contrast schematic diagram that uses the link utilization of prior art.The transverse axis of Fig. 9 is that offered load (unit the is Mbps) longitudinal axis is link utilization.When offered load reaches 100%, using the average link utilance of single path route is 70%, and using the average link utilance of ECMP is 84%, and use the average link utilance of weighting routing algorithm, is 92%.
The foregoing is only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.All any modifications of doing within the spirit and principles in the present invention, be equal to replacement, improvement etc., be all included in protection scope of the present invention.
Claims (9)
1. a network management of supporting elephant stream, is characterized in that, the method comprises:
When the elephant in network being detected is flowed, extract source address and the destination address of elephant stream, obtain a plurality of paths between source address and destination address, utilize the offered load in described path and the weight that network topology structure is calculated the respective path of supporting that elephant is flowed;
Transmitting of network management system generating network switch, and will transmit and be sent to the network switch, described transmitting comprises the weight of respective path and the operational order of the network switch of supporting elephant stream, described in carrying out for the network switch, transmit, the flow of elephant stream is distributed according to the weight in path.
2. the method for claim 1, is characterized in that, described in transmit and comprise stream table and group table;
Described stream table comprises destination address and corresponding action fields, and described action fields is pointed to respectively a group table;
Described group of table comprises respectively a plurality of action items, and described action items comprises respectively the weight in path and the operational order of the network switch.
3. the method for claim 1, is characterized in that, the method also comprises:
Generate LLDP message and be regularly issued to the network switch; Receive the LLDP message that the network switch is uploaded, by the LLDP message receiving, obtain network topology structure;
The statistics of the periodic polling network switch, to obtain the load of the network switch and the load of network link.
4. the method for claim 1, is characterized in that,
The described offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream utilized, comprising:
Obtain the set that represents all available paths between source address and destination address, the number that makes described path is I; To each path i(0≤i≤I-1) the following operation of execution:
The number that makes it subpath comprising is J
i; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); The load that makes link k is λ
kji(0≤k≤K
ji-1); The capacity that makes link k is c
kji(0≤k≤K
ji-1);
Calculate the link utilization of each link:
lU
kjilink utilization for link k;
The weight of calculating path:
w
i∈ 0,1 ..., 9} is the weight of path i.
5. a network management system, is characterized in that, described network management system comprises:
Elephant flow path computing unit, for when the elephant stream of network being detected, extract source address and the destination address of elephant stream, obtain a plurality of paths between source address and destination address, utilize the offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream;
Path management unit, for transmitting of generating network switch, and will transmit and be sent to the network switch, described transmitting comprises the weight of respective path and the operational order of the network switch of supporting elephant stream, described in carrying out for the network switch, transmit, the flow of elephant stream is distributed according to the weight in path.
6. network management system as claimed in claim 5, is characterized in that, described network management system also comprises:
Network topology unit, for generating LLDP message and being regularly issued to the network switch, and receives the LLDP message that the network switch is uploaded, and by the LLDP message receiving, obtains network topology structure;
Offered load monitoring means, for the statistics of the periodic polling network switch, to obtain the load of the network switch and the load of network link.
7. network management system as claimed in claim 5, is characterized in that,
Described elephant flow path computing unit, for utilizing the offered load in described path and the weight that network topology structure is calculated the respective path of supporting elephant stream;
Wherein, described elephant flow path computing unit, for obtaining the set that represents all available paths between source address and destination address, the number that makes described path is I;
Described elephant flow path computing unit, is further used for each path i(0≤i≤I-1) carry out following operation, with the weight of calculating path:
The number that makes it subpath comprising is J
i; Making the number of the link that subpath j comprises is K
ji(0≤j≤J
i-1); The load that makes link k is λ
kji(0≤k≤K
ji-1); The capacity that makes link k is c
kji(0≤k≤K
ji-1);
Calculate the link utilization of each link:
lUkji is the link utilization of link k;
The weight of calculating path:
w
i∈ 0,1 ..., 9} is the weight of path i.
8. a network switch, comprising:
Elephant stream reports unit, for choosing from the data flow receiving, is labeled as the data flow of elephant stream and selected data flow is reported to network management system;
Retransmission unit, for receiving and carry out transmitting of network management system transmission, distributes the flow of elephant stream according to the weight in transmitting.
9. a communication network, comprising:
Network management system as described in claim 5 to 7 any one; And
Signal is connected in the network switch as claimed in claim 8 of described network management system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310752282.6A CN103716208B (en) | 2013-12-31 | 2013-12-31 | Support network management, system, interchanger and the network of elephant stream |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310752282.6A CN103716208B (en) | 2013-12-31 | 2013-12-31 | Support network management, system, interchanger and the network of elephant stream |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103716208A true CN103716208A (en) | 2014-04-09 |
| CN103716208B CN103716208B (en) | 2017-06-30 |
Family
ID=50408813
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310752282.6A Expired - Fee Related CN103716208B (en) | 2013-12-31 | 2013-12-31 | Support network management, system, interchanger and the network of elephant stream |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103716208B (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104468353A (en) * | 2014-12-26 | 2015-03-25 | 深圳市新格林耐特通信技术有限公司 | SDN based data center network flow management method |
| CN104767694A (en) * | 2015-04-08 | 2015-07-08 | 大连理工大学 | A Data Stream Forwarding Method Oriented to Fat-Tree Data Center Network Architecture |
| CN105282044A (en) * | 2014-07-01 | 2016-01-27 | 上海宽带技术及应用工程研究中心 | Network multi-path implementation method and system based on OpenFlow protocol |
| CN105376178A (en) * | 2015-12-07 | 2016-03-02 | 北京邮电大学 | Data transmission method and device |
| CN106209669A (en) * | 2016-06-30 | 2016-12-07 | 中国人民解放军国防科学技术大学 | Towards SDN data center network maximum of probability path stream scheduling method and device |
| CN106301921A (en) * | 2016-08-16 | 2017-01-04 | 清华大学 | Elephant flow transmission dispatching method based on tunnel and system |
| WO2017000243A1 (en) * | 2015-06-30 | 2017-01-05 | 华为技术有限公司 | Routing method, and relevant device and system |
| CN106330730A (en) * | 2015-06-30 | 2017-01-11 | 中兴通讯股份有限公司 | Multicast control method and apparatus |
| CN106506395A (en) * | 2016-11-28 | 2017-03-15 | 迈普通信技术股份有限公司 | A kind of business stream scheduling method and device |
| CN107592272A (en) * | 2017-10-17 | 2018-01-16 | 盛科网络（苏州）有限公司 | A kind of method and device that load balancing is carried out based on source IP address |
| CN107637030A (en) * | 2015-05-19 | 2018-01-26 | 阿尔卡特朗讯公司 | Method and apparatus for self-tuning adaptive routing |
| CN107682258A (en) * | 2017-09-27 | 2018-02-09 | 北京邮电大学 | A kind of multi-path network transmission method and device based on virtualization |
| US10263883B2 (en) | 2016-12-14 | 2019-04-16 | International Business Machines Corporation | Data flow configuration in hybrid system of silicon and micro-electro-mechanical-switch (MEMS) elements |
| CN110505148A (en) * | 2019-07-24 | 2019-11-26 | 烽火通信科技股份有限公司 | Method and system for implementing weighted multipath |
| CN110572332A (en) * | 2019-08-30 | 2019-12-13 | 西南交通大学 | Task division method of network equipment packet observation data collection |
| CN112260950A (en) * | 2020-10-19 | 2021-01-22 | 桂林电子科技大学 | Cloud storage network multicast stream scheduling method based on service priority differentiation |
| CN112311674A (en) * | 2019-07-31 | 2021-02-02 | 北京华为数字技术有限公司 | Message sending method, device and storage medium |
| CN112751781A (en) * | 2019-10-31 | 2021-05-04 | 阿里巴巴集团控股有限公司 | Method, device and equipment for processing flow data and computer storage medium |
| CN114884883A (en) * | 2022-06-16 | 2022-08-09 | 深圳星云智联科技有限公司 | Flow forwarding method, device, equipment and storage medium |
| WO2023169407A1 (en) * | 2022-03-08 | 2023-09-14 | 华为技术有限公司 | Method for issuing routing table and related apparatus |
| US12015539B2 (en) | 2022-04-21 | 2024-06-18 | VMware LLC | Mitigating oversubscription of traffic to edge nodes in a data center |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102004003548B3 (en) * | 2004-01-23 | 2005-06-30 | Siemens Ag | Traffic loading optimization method for packet-based communications network with multi-path routing using adaption of distribution weightings of outgoing communications links dependent on traffic loadings |
| CN102984064A (en) * | 2012-12-28 | 2013-03-20 | 盛科网络（苏州）有限公司 | Method and system for distinguishing and transmitting elephant flow |
-
2013
- 2013-12-31 CN CN201310752282.6A patent/CN103716208B/en not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| 张进 等: "加权多路径路由在LEO卫星星座组网中的仿真应用", 《电讯技术》 * |
Cited By (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105282044A (en) * | 2014-07-01 | 2016-01-27 | 上海宽带技术及应用工程研究中心 | Network multi-path implementation method and system based on OpenFlow protocol |
| CN105282044B (en) * | 2014-07-01 | 2018-08-07 | 上海宽带技术及应用工程研究中心 | Network multipath diameter realization method and system based on OpenFlow agreements |
| CN104468353A (en) * | 2014-12-26 | 2015-03-25 | 深圳市新格林耐特通信技术有限公司 | SDN based data center network flow management method |
| CN104767694A (en) * | 2015-04-08 | 2015-07-08 | 大连理工大学 | A Data Stream Forwarding Method Oriented to Fat-Tree Data Center Network Architecture |
| CN104767694B (en) * | 2015-04-08 | 2017-10-13 | 大连理工大学 | A kind of stream compression forwarding method towards Fat Tree data center network architectures |
| CN107637030A (en) * | 2015-05-19 | 2018-01-26 | 阿尔卡特朗讯公司 | Method and apparatus for self-tuning adaptive routing |
| US10630592B2 (en) | 2015-06-30 | 2020-04-21 | Huawei Technologies Co., Ltd. | Routing method, related device, and system |
| CN106330730A (en) * | 2015-06-30 | 2017-01-11 | 中兴通讯股份有限公司 | Multicast control method and apparatus |
| WO2017000243A1 (en) * | 2015-06-30 | 2017-01-05 | 华为技术有限公司 | Routing method, and relevant device and system |
| US11509584B2 (en) | 2015-06-30 | 2022-11-22 | Huawei Technologies Co., Ltd. | Routing method, related device, and system |
| CN105376178A (en) * | 2015-12-07 | 2016-03-02 | 北京邮电大学 | Data transmission method and device |
| CN105376178B (en) * | 2015-12-07 | 2019-03-15 | 北京邮电大学 | A method and device for data transmission |
| CN106209669A (en) * | 2016-06-30 | 2016-12-07 | 中国人民解放军国防科学技术大学 | Towards SDN data center network maximum of probability path stream scheduling method and device |
| CN106209669B (en) * | 2016-06-30 | 2019-06-07 | 中国人民解放军国防科学技术大学 | Towards SDN data center network maximum probability path stream scheduling method and device |
| CN106301921A (en) * | 2016-08-16 | 2017-01-04 | 清华大学 | Elephant flow transmission dispatching method based on tunnel and system |
| CN106301921B (en) * | 2016-08-16 | 2019-05-14 | 清华大学 | Elephant flow transmission dispatching method and system based on tunnel |
| CN106506395A (en) * | 2016-11-28 | 2017-03-15 | 迈普通信技术股份有限公司 | A kind of business stream scheduling method and device |
| US10263883B2 (en) | 2016-12-14 | 2019-04-16 | International Business Machines Corporation | Data flow configuration in hybrid system of silicon and micro-electro-mechanical-switch (MEMS) elements |
| CN107682258A (en) * | 2017-09-27 | 2018-02-09 | 北京邮电大学 | A kind of multi-path network transmission method and device based on virtualization |
| CN107592272A (en) * | 2017-10-17 | 2018-01-16 | 盛科网络（苏州）有限公司 | A kind of method and device that load balancing is carried out based on source IP address |
| CN107592272B (en) * | 2017-10-17 | 2020-10-27 | 盛科网络（苏州）有限公司 | Method and device for load balancing based on source IP address |
| CN110505148A (en) * | 2019-07-24 | 2019-11-26 | 烽火通信科技股份有限公司 | Method and system for implementing weighted multipath |
| CN112311674A (en) * | 2019-07-31 | 2021-02-02 | 北京华为数字技术有限公司 | Message sending method, device and storage medium |
| CN112311674B (en) * | 2019-07-31 | 2022-02-25 | 北京华为数字技术有限公司 | Message sending method, device and storage medium |
| CN110572332B (en) * | 2019-08-30 | 2022-08-26 | 西南交通大学 | Network equipment message observation data acquisition task dividing method |
| CN110572332A (en) * | 2019-08-30 | 2019-12-13 | 西南交通大学 | Task division method of network equipment packet observation data collection |
| CN112751781A (en) * | 2019-10-31 | 2021-05-04 | 阿里巴巴集团控股有限公司 | Method, device and equipment for processing flow data and computer storage medium |
| CN112260950A (en) * | 2020-10-19 | 2021-01-22 | 桂林电子科技大学 | Cloud storage network multicast stream scheduling method based on service priority differentiation |
| WO2023169407A1 (en) * | 2022-03-08 | 2023-09-14 | 华为技术有限公司 | Method for issuing routing table and related apparatus |
| US12015539B2 (en) | 2022-04-21 | 2024-06-18 | VMware LLC | Mitigating oversubscription of traffic to edge nodes in a data center |
| CN114884883A (en) * | 2022-06-16 | 2022-08-09 | 深圳星云智联科技有限公司 | Flow forwarding method, device, equipment and storage medium |
| CN114884883B (en) * | 2022-06-16 | 2024-01-30 | 深圳星云智联科技有限公司 | Traffic forwarding method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103716208B (en) | 2017-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103716208A (en) | Network management method and system supporting elephant flow, interchanger and network | |
| US10771389B2 (en) | Virtual tunnel endpoints for congestion-aware load balancing | |
| US20220217078A1 (en) | System and method for facilitating tracer packets in a data-driven intelligent network | |
| Liu et al. | SDN based load balancing mechanism for elephant flow in data center networks | |
| Gelenbe et al. | Adaptive qos routing for significant events in wireless sensor networks | |
| CN104579961B (en) | The dispatching method and device of data message | |
| CN102780637B (en) | Routing method for data transmission in space delay/disruption tolerant network | |
| CN110351286B (en) | Link flooding attack detection response mechanism in software defined network | |
| CN102123088B (en) | Set up the method and apparatus of traffic engineering tunnel | |
| Trestian et al. | OFLoad: An OpenFlow-based dynamic load balancing strategy for datacenter networks | |
| Cui et al. | Difs: Distributed flow scheduling for data center networks | |
| CN102263697A (en) | Method and device for sharing aggregated link traffic | |
| CN106961387A (en) | A kind of link type DDoS defence methods migrated certainly based on forward-path and system | |
| CN107786440A (en) | A kind of method and device of data message forwarding | |
| CN105915467A (en) | Data center network flow balancing method and device oriented to software definition | |
| CN101958847A (en) | A Selection Method of Distributed QOS Routing | |
| CN101018203A (en) | Dynamic load allocation method of the aviation telecom network | |
| CN109039941B (en) | Adaptive Packet Scattering Method Based on Path Classification in Data Center Networks | |
| CN104917677A (en) | Data stream forwarding control method and system | |
| Alanazi et al. | CAFT: Congestion-aware fault-tolerant load balancing for three-tier clos data centers | |
| Chang et al. | Using sdn technology to mitigate congestion in the openstack data center network | |
| Huang et al. | An enhanced scheduling framework for elephant flows in SDN-based data center networks | |
| Wang et al. | Spray and Wait routing algorithm based on Transfer Utility of Node in DTN | |
| CN112910795A (en) | Edge load balancing method and system based on many sources | |
| Balakiruthiga et al. | A simple congestion avoidance mechanism for opendaylight (odl)-multipath tcp (mptcp) network structure in software defined data center (sddc) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170630 |