CN106330730A - Multicast control method and apparatus - Google Patents
Multicast control method and apparatus Download PDFInfo
- Publication number
- CN106330730A CN106330730A CN201510374309.1A CN201510374309A CN106330730A CN 106330730 A CN106330730 A CN 106330730A CN 201510374309 A CN201510374309 A CN 201510374309A CN 106330730 A CN106330730 A CN 106330730A
- Authority
- CN
- China
- Prior art keywords
- multicast tree
- weight
- multicast
- tree
- 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.)
- Withdrawn
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/20—Support for services
- H04L49/201—Multicast operation; Broadcast operation
-
- 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
- H04L45/48—Routing tree calculation
-
- 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
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Small-Scale Networks (AREA)
Abstract
The present invention provides a multicast control method and apparatus. The method comprises the steps of detecting network topology and link information; generating a plurality of multicast trees according to the network topology; calculating a weight of each multicast tree of the plurality of multicast trees according to the network topology and link information, wherein the weight is used for indicating a switch to select a priority of transmitting data of a corresponding multicast tree in a polling manner; and delivering the weight of each multicast tree to the switch of the corresponding multicast tree. When a multicast source forwards data by using multiple multicast trees, the load of the switches is unbalanced, so that the forwarding efficiency is low. The method improves the forward efficiency.
Description
Technical field
The present invention relates to the communications field, in particular to a kind of multicast control method and device.
Background technology
Along with developing rapidly of data center, data center network is disposed multicast service, is possible not only to reduce identical number
According to the quantity sent, save unnecessary network traffics, and the bandwidth demand to transmitting terminal can be reduced, can be more effectively
Utilize bandwidth resources.Especially, particularity based on tree data center network, the multicast between multicast member
Tree can be determined by some core layer switch, and between multiple multicast tree that multiple core layer switches determine
Possible non-overlapping link, or there may be the link that partly overlaps.
Research process finds, just there may be, due to multicast tree, the link that partly overlaps, in multicast source is by overlay links
Switch forward data time, carry out data forwarding respectively discounting for the link information of each multicast tree, then may
Cause some follow-up switch load excessive, and the problem that some switch load is too small, thus have impact on turning of data
Send out efficiency.
The switch load that forwards data to exist by multiple multicast tree for multicast source in correlation technique is unbalanced to be caused
The problem that forward efficiency is low, the most not yet proposes effective solution.
Summary of the invention
The invention provides a kind of multicast control method and device, at least to solve in correlation technique multicast source by multiple groups
Broadcast the problem that the unbalanced forward efficiency caused of switch load of tree forwarding data existence is low.
According to an aspect of the invention, it is provided a kind of multicast control method, including: detection network topology and link letter
Breath；Multiple multicast tree are generated according to described network topology；According to described network topology and described link information, calculate respectively
The weight of each multicast tree in the plurality of multicast tree, described weight is used for indicating switch polls to select corresponding multicast tree to pass
The priority of transmission of data；Issue the described weight of described each multicast tree to the switch in corresponding multicast tree.
Preferably, according to described network topology and described link information, calculate in the plurality of multicast tree each group respectively
After broadcasting the described weight of tree, described method also includes: detect link information in real time；According to described network topology with in real time
The link information of detection, updates the described weight of described each multicast tree respectively；Issue the described weight of renewal to respective sets
Broadcast the described switch in tree.
Preferably, according to described network topology and the link information of detection in real time, the institute of described each multicast tree is updated respectively
State weight to include: in the case of determining link failure by the link information of detection in real time, by corresponding for the link of fault
The described weight zero setting of multicast tree.
Preferably, according to described network topology and described link information, calculate each multicast in the plurality of multicast tree respectively
The described weight of tree includes: according to described link information and default multicast tree weight rule, calculate described each multicast respectively
The weight of tree；According to described network topology, it is judged that whether the plurality of multicast tree exists overlay path；Determining existence
In the case of overlay path, according to the weight of multiple multicast tree of described overlay path, calculate the power of overlay links multicast tree
Weight, and using the weight of described overlay links multicast tree as the weight of multiple multicast tree of described overlay path.
Preferably, issue the described weight of described each multicast tree to include to the described switch in corresponding multicast tree: issue
Select type group table, to the switch in corresponding multicast tree, wherein, carries respective sets in described select type group table
That broadcasts tree goes out port and the described weight of corresponding multicast tree.
Preferably, the described weight of described each multicast tree is being issued to after the described switch in corresponding multicast tree, institute
Method of stating also includes: described switch receives the described weight of at least one multicast tree；Described switch is according to described weight
Poll selects multicast tree transmission data at least one multicast tree described.
According to another aspect of the present invention, additionally provide a kind of multicast and control device, including: detection module, it is used for examining
Survey grid network topological sum link information；Generation module, for generating multiple multicast tree according to described network topology；Computing module,
For according to described network topology and described link information, calculate the weight of each multicast tree in the plurality of multicast tree respectively,
Described weight selects the priority of corresponding multicast tree transmission data for indicating switch polls；Issue module, be used for issuing
The described weight of described each multicast tree is to the switch in corresponding multicast tree.
Preferably, described detection module, it is additionally operable to detect in real time link information；Described computing module, is additionally operable to according to institute
State network topology and the link information of detection in real time, update the described weight of described each multicast tree respectively；Described issue mould
Block, is additionally operable to the described weight issuing renewal to the described switch in corresponding multicast tree.
Preferably, described computing module, it is additionally operable in the situation being determined link failure by the link information of detection in real time
Under, by the described weight zero setting of corresponding for the link of fault multicast tree.
Preferably, described computing module includes: the first computing unit, for according to described link information and default multicast tree
Weight rule, calculates the weight of described each multicast tree respectively；Judging unit, for according to described network topology, it is judged that
Whether the plurality of multicast tree exists overlay path；Second computing unit, for determining the situation that there is overlay path
Under, according to the weight of multiple multicast tree of described overlay path, calculate the weight of overlay links multicast tree, and by described heavy
The weight of folded link multicast tree is as the weight of multiple multicast tree of described overlay path.
Preferably, described in issue module, be additionally operable to issue select type group table to the switch in corresponding multicast tree, its
In, carry corresponding multicast tree in described select type group table goes out port and the described weight of corresponding multicast tree.
By the present invention, use detection network topology and link information；Multiple multicast tree are generated according to network topology；According to
Network topology and link information, calculate the weight of each multicast tree in multiple multicast tree respectively, and weight is used for indicating switch
Poll selects the priority of corresponding multicast tree transmission data；Issue the weight of each multicast tree to the exchange in corresponding multicast tree
The mode of machine, solves the unbalanced forwarding caused of switch load that multicast source forwards data to exist by multiple multicast tree
Inefficient problem, improves forward efficiency.
Accompanying drawing explanation
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, the present invention
Schematic description and description be used for explaining the present invention, be not intended that inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart of multicast control method according to embodiments of the present invention；
Fig. 2 is the structured flowchart that multicast according to embodiments of the present invention controls device；
Fig. 3 is the preferred structure block diagram that multicast according to embodiments of the present invention controls device；
Fig. 4 is the structural representation of the data center network according to the preferred embodiment of the present invention one；
Fig. 5 is the flow chart of the multicast tree weighing computation method according to the preferred embodiment of the present invention two；
Fig. 6 is the flow chart of the method for multicasting according to the preferred embodiment of the present invention three；
Fig. 7 is structure and the schematic diagram of multicast tree weight of the data center network according to the preferred embodiment of the present invention three；
Fig. 8 is the false code of the weight poll selection algorithm according to the preferred embodiment of the present invention three；
Fig. 9 is the flow chart of the method for multicasting according to the preferred embodiment of the present invention four；
Figure 10 is structure and the schematic diagram of multicast tree weight of the data center network according to the preferred embodiment of the present invention four.
Detailed description of the invention
Below with reference to accompanying drawing and describe the present invention in detail in conjunction with the embodiments.It should be noted that in the feelings do not conflicted
Under condition, the embodiment in the application and the feature in embodiment can be mutually combined.
It should be noted that term " first " in description and claims of this specification and above-mentioned accompanying drawing, " second "
Etc. being for distinguishing similar object, without being used for describing specific order or precedence.
Providing a kind of multicast control method in the present embodiment, Fig. 1 is multicast control method according to embodiments of the present invention
Flow chart, as it is shown in figure 1, this flow process comprises the steps:
Step S102, detection network topology and link information；
Step S104, generates multiple multicast tree according to network topology；
Step S106, according to network topology and link information, calculates the weight of each multicast tree in multiple multicast tree respectively,
Weight selects the priority of corresponding multicast tree transmission data for indicating switch polls；
Step S108, issues the weight of each multicast tree to the switch in corresponding multicast tree.
By above-mentioned steps, indicate switch by the weight having network topology and the calculated multicast tree of link information
Poll selects the priority of corresponding multicast tree transmission data, thus solves multicast source and forward data to deposit by multiple multicast tree
The low problem of the unbalanced forward efficiency caused of switch load, improve forward efficiency.
Preferably, said method can be applied on OpenFlow controller, will be with OpenFlow controller below
Example is described and illustrates.
Preferably, after the said method of the application embodiment of the present invention, it is also possible to according to the link information of detection in real time, real
Time adjust multicast tree weight, such as, after step s 106, OpenFlow controller can also detect link in real time
Information；According to network topology and the link information of detection in real time, update the weight of each multicast tree respectively；And issue renewal
Weight to the switch in corresponding multicast tree.
Preferably, when detection link information in real time, if be detected that a certain link failure, then can be by the link of fault
The weight zero setting of corresponding multicast tree；After the weight zero setting of multicast tree, issue the weight after zero setting to switch, friendship
Change planes and then no longer can transmit data by this multicast tree.After the fault restoration of link being detected in real time, it is also possible to by weight
Newly issue the weight of multicast tree to switch, so that switch can continue through this multicast tree and other multicast tree by wheel
The mode multicast packet ask.
Preferably, in step s 106, if a certain multicast tree and the non-overlapping link of other multicast tree, then could basis
Link information and default multicast tree weight rule, directly calculate the weight of this multicast tree；But, sentencing according to network topology
Breaking in the case of there is overlay path to multiple multicast tree, the weight of these multiple multicast tree then needs secondary calculating, such as:
The weight of the multiple multicast tree according to overlay path, calculates the weight of overlay links multicast tree, and by overlay links multicast tree
Weight as the weight of multiple multicast tree of overlay path.So, there are multiple multicast tree of overlay path at overlapping chain
There is on road identical weight.
Preferably, issue the weight of each multicast tree to include to the switch in corresponding multicast tree: issue OpenFlow association
Select type group table defined in view, to the switch in corresponding multicast tree, wherein, carries in select type group table
Corresponding multicast tree go out port and the weight of corresponding multicast tree.
Preferably, issuing the weight of each multicast tree to after the switch in corresponding multicast tree, switch receives at least
The weight of one multicast tree；Switch selects multicast tree transmission data at least one multicast tree according to weight poll.
Preferably, switch have multiple multicast tree available in the case of, switch can select in these multiple multicast tree
Several multicast tree that weight is higher, and pass through several multicast tree poll group multicast datas that weight is higher.
Through the above description of the embodiments, those skilled in the art is it can be understood that arrive according to above-described embodiment
Method can add the mode of required general hardware platform by software and realize, naturally it is also possible to by hardware, but a lot
In the case of the former is more preferably embodiment.Based on such understanding, technical scheme is the most in other words to existing
The part having technology to contribute can embody with the form of software product, and this computer software product is stored in one
In storage medium (such as ROM/RAM, magnetic disc, CD), including some instructions with so that a station terminal equipment (can
To be mobile phone, computer, server, or the network equipment etc.) method that performs each embodiment of the present invention.
Additionally providing a kind of multicast in the present embodiment and control device, this device is used for realizing above-described embodiment and being preferable to carry out
Mode, had carried out repeating no more of explanation.As used below, term " module " can realize predetermined function
Software and/or the combination of hardware.Although the device described by following example preferably realizes with software, but hardware,
Or the realization of the combination of software and hardware also may and be contemplated.
Fig. 2 is the structured flowchart that multicast according to embodiments of the present invention controls device, as in figure 2 it is shown, this device includes:
Detection module 22, generation module 24, computing module 26 and issue module 28, wherein, detection module 22, it is used for examining
Survey grid network topological sum link information；Generation module 24, coupled to detection module 22, for generating many according to network topology
Individual multicast tree；Computing module 26, coupled to generation module 24, for according to network topology and link information, counts respectively
Calculating the weight of each multicast tree in multiple multicast tree, weight is used for indicating switch polls to select corresponding multicast tree transmission data
Priority；Issue module 28, coupled to computing module 26, for issuing the weight of each multicast tree to corresponding multicast
Switch in tree.
By the interaction of above-mentioned module, solve multicast source and forward the switch of data existence to bear by multiple multicast tree
Carry the problem that the unbalanced forward efficiency caused is low, improve forward efficiency.
Preferably, detection module 22, it is additionally operable to detect in real time link information；Computing module 26, is additionally operable to according to network
The link information that topological sum detects in real time, updates the weight of each multicast tree respectively；Issue module 28, be additionally operable to issue more
New weight is to the switch in corresponding multicast tree.
Preferably, computing module 26, it is additionally operable in the case of determining link failure by the link information of detection in real time,
Weight zero setting by corresponding for the link of fault multicast tree.
Fig. 3 is the preferred structure block diagram that multicast according to embodiments of the present invention controls device, as shown in Figure 3, it is preferable that
Computing module 26 includes: the first computing unit 262, for according to link information and default multicast tree weight rule, difference
Calculate the weight of each multicast tree；Judging unit 264, for according to network topology, it is judged that whether multiple multicast tree exist
Overlay path；Second computing unit 266, in the case of determining and there is overlay path, according to overlay path
The weight of multiple multicast tree, calculates the weight of overlay links multicast tree, and using the weight of overlay links multicast tree as overlap
The weight of multiple multicast tree in path.
Preferably, issue module 28, be additionally operable to issue select type group table to the switch in corresponding multicast tree, wherein,
Carry corresponding multicast tree in select type group table goes out port and the weight of corresponding multicast tree.
It should be noted that above-mentioned modules can be by software or hardware realizes, for the latter, Ke Yitong
Cross in the following manner to realize, but be not limited to this: above-mentioned module is respectively positioned in same processor；Or, above-mentioned module position respectively
In multiple processors.
Embodiments of the invention additionally provide a kind of software, and this software is used for performing in above-described embodiment and preferred implementation
The technical scheme described.
Embodiments of the invention additionally provide a kind of storage medium.In the present embodiment, above-mentioned storage medium can be set
For storage for performing the program code of following steps:
Step S1, detection network topology and link information；
Step S2, generates multiple multicast tree according to network topology；
Step S3, according to network topology and link information, calculates the weight of each multicast tree in multiple multicast tree respectively,
Weight selects the priority of corresponding multicast tree transmission data for indicating switch polls.
Step S4, issues the weight of each multicast tree to the switch in corresponding multicast tree.
Alternatively, in the present embodiment, above-mentioned storage medium can include but not limited to: USB flash disk, read only memory (ROM,
Read-Only Memory), random access memory (RAM, Random Access Memory), portable hard drive,
The various medium that can store program code such as magnetic disc or CD.
In order to the description making the embodiment of the present invention is clearer, it is described below in conjunction with preferred embodiment and illustrates.
The preferred embodiment of the present invention provides reliable efficient many multicast tree polling mechanism of a kind of software definition data center, relates to
And one utilizes OpenFlow group table (group table) to realize reliable efficient many multicast tree poll in data center network
Mechanism.
In multicast service, data are sent to multiple recipient along a multicast tree from source, and the quality of multicast tree is shadow
Ring a key factor of multicast service.Existing substantial amounts of multicast tree generating algorithm, such as shortest path first, minimum generation
Tree algorithm and Steiner tree algorithm.Traditionally, between multicast member, set up a multicast tree for multicast data transmission,
If in the case of there is a plurality of feasible path or multiple feasible multicast tree from source to recipient, having an opportunity, it is multiple superfluous to utilize
Remaining multicast tree multicast data transmission.As in point to point network transmitting procedure equal cost multipath (Equal-Cost Multipath,
Referred to as ECMP) and weight multipath (Weight-Cost Multipath, referred to as WCMP) agreement and make full use of
Redundancy multipath, if multicast service also can so make full use of multipath, it is achieved transport multicast while many multicast tree
Data, thus make multicast service efficiency be greatly improved, and take full advantage of the bandwidth of network, it is achieved multicast service
High efficiency.
Meanwhile, software defined network based on OpenFlow (Software Defined Network, referred to as SDN)
Technology, adds motility and the extensibility of network management.Its core concept is network-based control plane and data to be put down
Face separates.OpenFlow controller operates by OpenFlow switch carries out write stream table and group table etc., and it is right to complete
Whole network-based control.In OpenFlow v1.1 agreement, add group table especially, provide the biggest side to multicast transmission
Just.Multicast packet, after coupling stream table, performs entrance group table after stream table instruction, then proceedes to execution group table action bucket collection
(Action Buckets).The structure of group table includes group table id, group table type, enumerator, action bucket collection.Each type
Group table perform action the most different, everything bucket in all type group table then execution group tables；Select type group table is then
Selecting an action bucket to perform, wherein selection algorithm can freely configure.
Combine OpenFlow agreement in a preferred embodiment of the invention to control to be optimized to multicast, thus improve in data
The reliability of heart multicast service and high efficiency.Based on OpenFlow many multicast tree wheel that the preferred embodiment of the present invention proposes
The mechanism of inquiry have employed following scheme:
OpenFlow controller collection network topological sum link information, generates multiple multicast tree for multicast member.By controlling
Multicast tree weight computation module in device processed is that each multicast tree calculates weight, and weight then determines this multicast tree for forwarding
The priority of multicast packet.According to the weight of each multicast tree, the weight poll selection algorithm on switch realize many groups
Broadcast tree poll multicast data transmission, improve efficiency of transmission.
In such scheme, by controller periodic contribution link information, not only it is adapted dynamically multicast tree weight,
Realize the load balancing of network, and can also protect when certain link failure by the weight of zero setting fault multicast tree
The reliability of barrier multicast service.
Preferably, in the data center network of tree, when choosing a core layer switch as multicast tree path
During node, it is possible to generate multicast tree for multicast member.If there are multiple feelings in the multicast tree passing by a core layer switch
Condition, randomly selects a multicast tree corresponding with this core layer switch.So choosing multiple core layer switch, this multicast
Group just can have multiple multicast tree, non-overlapping link or the link that partly overlaps between these multicast tree.Tree data
The point-to-points multicast tree of spy of central site network multicast data transmission simultaneously brings great convenience.Based on OpenFlow technology,
Controller collection network topological sum link information, is set up multiple multicast tree by many multicast tree generating algorithm.
Preferably, controller, after collection network link information sets up many multicast tree, utilizes multicast tree weight rule to count respectively
Calculating the weight of each multicast tree, weight represents the priority that this multicast tree is selected, and multicast packet is according to this weight poll
Multiple multicast tree is selected to be transmitted.At many multicast tree overlay path, overlapping many multicast tree weight needs secondary calculating.
Then the select type group table (based on OpenFlow v1.1 agreement) configured by multicast tree weight is handed down to by controller
Switch.Configuration weight polling algorithm on switches, by according to the weight position in select type group table, from multiple
In multicast tree, poll selects a multicast tree for multicast data forwarding.The number of multicast tree also must be limited, and one
As choose the higher several multicast tree of priority for transmitting data.
Preferably due to the mutability of the flow of data center network and randomness, based on OpenFlow technology,
The collection network link information in OpenFlow controller cycle also carries out decision-making, it is possible to well realize traffic engineering.This
In utilize the link information collected to calculate the weight of multicast tree in real time, adjust dynamically according to calculated multicast tree weight
Weight position in whole select type group table.Can more preferably realize dynamic load leveling and the considerable controllability to multicast service,
Ensure the efficiency of transmission of data.
Preferably, at the link that breaks down, by the weight zero setting of its corresponding multicast tree, only between residue multicast tree
Poll selects, and reduces packet loss, thus ensures the reliability of data.
Below in conjunction with accompanying drawing the preferred embodiments of the present invention it is described and illustrates.
Wherein it should be noted that in following preferred embodiment, with based under OpenFlow environment, stack of cabinets (point
Of delivery, referred to as pod) number be 4 simple fat tree (Fat-tree) data center network as a example by illustrate.
But, in actual use, may be applied in pod number more Fat-tree network or in other data
In heart network, it is also possible to be applied in other software defined network, but, these different application scenarios are the most applicable
Method in the present invention.
Preferred embodiment one:
Fig. 4 show in the preferred embodiment of the present invention in the Fat-tree data center network that pod number is 4, for
Multicast member sets up the algorithm of many multicast tree.Controller obtains whole network topological information, only in conjunction with Fat-tree network
Special level, is calculated by the many multicast tree generation module in controller and generates multicast tree, comprise the steps:
Step S401: according to network topological information, randomly selects 4 core layer switches；
Wherein, in Fig. 4, shallow dashed rectangle represents a pod, has pod0 ... pod3, T1 are marginal layer switches, A1,
A2 represents convergence-level switch, and C1, C2, C3, C4 represent core layer switch.S represents source, R1, R2,
R3, R4, R5 are recipients.In the diagram, four multicast tree chosen are respectively with short chain-dotted line, double dot dash line, thin
Line and thick line represent, in addition it is also necessary to explanation, in pod0～pod3 convergence-level switch and edge switch it
Between link fall within aforementioned four multicast tree, such as, the link between A1 and T1 belongs simultaneously to the multicast at C1 place
Tree (multicast tree that chain-dotted line represents) and the multicast tree (two dashdotted multicast tree) at C2 place；In view of lines in accompanying drawing
Restriction, the most do not distinguish and illustrate.
Step S402: determine the source exclusive path to 4 core layer switches, convergence-level switch and marginal layer are deposited
When node is shared by branch, it is marked, respectively T1 and A1, A2；
Step S403:4 core layer switch uniquely determines to the path of the recipient of multicast member；
Preferred embodiment two:
Fig. 5 is after preferred embodiment one establishes multiple multicast tree, the weight calculation of multicast tree and the collocation method of group table,
It is calculated according to the link-state information collected by controller, comprises the steps:
Step S501: the network link information that controller is collected, such as bandwidth b of each link, time delay d, loads v；
Step S502: calculated the weight of each multicast tree by the multicast tree weight computation module in controlling
, such as: weight W=Φ (b, d, v)；
Wherein it is possible to using the minimum bandwidth of links all in multicast tree as weight；Or the minimal time delay of multicast tree is made
Weight for multicast tree；Or using the algebraical sum of average bandwidth and time delay as weight.For different business demands, power
The computation rule of weight has multiformity.Weight WiUniquely represent the priority of the i-th multicast tree.
Notice the feature of many multicast tree, all of branch node all in source with in pod.So having only at this pod
In marginal layer switch and convergence-level switch at selection of configuration algorithm and issue group table and can be achieved with many multicast tree poll
Transmission data.The multicast tree choosing that convergence-level switch the marginal layer switch being joined directly together in source with pod occurs
Select, be called one-level and select and two grades of selections.Two grades of selections are compared one-level and are selected simple, and two grades of convergence-level switches are every
Individual output port only represents a multicast tree, and the output port link of one-level marginal layer switch is probably multiple multicast tree
Overlay links, it will represent multiple multicast tree.When selecting many multicast tree overlay links for one-level, it is necessary to individually
Calculate overlay links multicast tree weight and configure group table.
Step S503: selecting for two grades when, each output port of convergence-level switch only represents a multicast tree,
Now, the action bucket of each group of table defines a port output, and the weight position of action bucket is exactly the weight of this multicast tree；
Step S504: the when of there is one-level selection, the output port link of marginal layer switch is probably multiple multicast
The overlay links of tree, when now organizing the action bucket output of table, it is also desirable to carry out weight poll selection, the weight of this action bucket
(weight) must carry out integrating computing according to the weight of the multiple multicast tree through this link
W=Ψ (W1, W2...Wn, b, d, v), common integration rules takes weight limit W=max (W1,W2...Wn) or
Average weight W=Avr (W1, W2...Wn)。
Preferred embodiment three:
Fig. 6 show after preferred embodiment one sets up multiple multicast tree and example two calculates this many multicast tree weight, it is achieved many
Multicast tree high efficiency method, comprises the steps:
Step S601: by link discovery module collection network topology and the link information of controller, many multicast tree generate mould
Tuber generates multiple multicast tree according to the topology information collected and multicast member relation, then is calculated by multicast tree weight computation module
The weight of each multicast tree.As it is shown in fig. 7, short chain-dotted line, double dot dash line, fine rule and thick line represent four multicasts respectively
Tree, P3 and P4 of A1 represents 3 ports and 4 ports of A1 respectively, and P1, P2, P3, P4 of C1 represent respectively
The 1 of C1,2,3,4 ports.In addition it is also necessary to explanation, convergence-level switch and edge in pod0～pod3
Link between switch falls within aforementioned four multicast tree, and such as, the link between A1 and T1 belongs simultaneously to C1 institute
Multicast tree (multicast tree that chain-dotted line represents) and the multicast tree (two dashdotted multicast tree) at C2 place；In view of attached
In figure, the restriction of lines, does not the most distinguish and illustrates.
Wherein, the weight calculating each multicast tree is respectively W1=7, W2=9, W3=4, W4=1, particularly exist
Weight at many multicast tree overlay links is W12=9, W34=4:
Step S602: one-level selects, when sender S starts to send the edge switch T1 that multicast packet is connected to it,
Multicast packet arrives edge switch, and the stream table and the group table that issue according to controller forward.Stream table coupling multicast packet
Bag, proceeds to corresponding select type group table by the packet of coupling.Selected different by select type group table weight poll
Action bucket carries out exporting (i.e. representing different multicast tree), and the action that action bucket herein performs is port output；
Wherein, having group table id=1, group entry type is select type, action bucket bucket=output:3weight=4,
Bucket=output:4weight=9.Being performed the two buckets by the circulation of weight poll selection algorithm, weight poll selects
Algorithm is as shown in Figure 8.Under normal circumstances, selection algorithm circulation 3 ports and the output of 4 ports are performed.
Step S603: two grades of selections, the multicast packet of edge switch is sent to convergence-level switch A 1 and A2, stream
Table coupling multicast packet, proceeds to corresponding select type group table by the packet of coupling；
Wherein, the group table that A1, A2 switch issues is as follows:
A1: having group table id=1, group entry type is select type, action bucket bucket=output:3weight=9,
Bucket=output:4weight=7.The same with edge switch, circulation performs two buckets, and poll selects A1 to hand over
3 ports changed planes and 4 ports forward, and multicast packet arrives separately at C1 and C2 switch.
A2: having group table id=1, group entry type is select type, action bucket bucket=output:3weight=1,
Bucket=output:4weight=4.The same with edge switch, circulation performs two buckets, and poll selects A2 to hand over
3 ports changed planes and 4 ports forward, and multicast packet arrives separately at C3 and C4 switch.
Step S604: what the arrival core layer switch having from convergence-level switch data out had be directly forwarded to
The recipient of pod.Adding, at core layer switch, become a mandarin table and group table, the packet of coupling is proceeded to group table by stream table.Now
All core layer group of switches table types are all types, action bucket bucket=output:1, bucket=output:2,
Bucket=output:3, bucket=output:4.Now multicast packet is respectively from 1, and 2,3,4 ports forward；
Step S605: the path from core layer to recipient is well-determined.Switch in residue multicast tree, only
Need execution group table.Group table id, group table type is indirect type, bucket=output:X.Multicast packet is not through
It is sent to each recipient, it is achieved pass altogether, improve the efficiency of transmission with multicast tree.
Preferred embodiment four:
After the high efficiency realizing preferred embodiment three, it is also possible to improve further its multicast service dynamic load leveling and
Fault recovery.Due to the randomness of load in data center network, the situation of link changes the most therewith, to multicast service
There is certain impact, according to the real time status of network, it is achieved dynamic load leveling, comprise the steps:
Step S901: by controller cycle contribution link information, in real time according to calculating groups such as the load of multicast tree and time delays
Broadcast the weight of tree；
Wherein, as shown in Figure 10, short chain-dotted line, double dot dash line, fine rule and thick line represent four generated respectively
Multicast tree.It should be noted that the link between convergence-level switch and edge switch falls within pod0～pod3
Aforementioned four multicast tree, such as, the link between A1 and T1 belongs simultaneously to the multicast tree at C1 place, and (chain-dotted line represents
Multicast tree) and the multicast tree (two dashdotted multicast tree) at C2 place；In view of the restriction of lines in accompanying drawing, at figure
Do not distinguish in 10 and illustrate.
Wherein, the weight calculating each multicast tree is respectively W1=10, W2=5, W3=4, W4=1, particularly exist
Weight at many multicast tree overlay links is W12=10, W34=4.
Step S902: according to the multicast tree weight recalculated, to needing one-level, two grades of marginal layers selected and convergence
The group table of layer switch reconfigures and issues；
T1: having group table id=1, group entry type is select type, action bucket bucket=output:3weight=4,
Bucket=output:4weight=10；
A1: having group table id=1, group entry type is select type, action bucket bucket=output:3weight=5,
Bucket=output:4weight=10；
A2: having group table id=1, group entry type is select type, action bucket bucket=output:3weight=1,
Bucket=output:4weight=4；
Step after step S902 is identical with step S604 shown in Fig. 6～step S605, will not be described in great detail at this.
Wherein, if link failure, the higher level multicast tree branch broken down is shared and points out, amendment group table, will
The weight of corresponding bucket is set to zero, and this is the special circumstances that dynamic load leveling adjusts.In select type group
In table, if the weight of bucket is zero, the most do not perform this bucket.Such that it is able to avoid non-working port multicast tree,
Proceed weight poll in remaining good ports to forward.
In sum, the above embodiment of the present invention and preferred embodiment, in data center environment, utilize OpenFlow
Technology, sets up multiple multicast tree and is simultaneously used for multicast service, and real time monitoring network link condition, and dynamically regulation link is born
Carry, it is achieved equally loaded, reached multicast service reliably efficient target.The above embodiment of the present invention and being preferable to carry out
Example takes full advantage of the bandwidth of network；Many multicast tree are utilized to pass the efficiency of transmission that improve multicast altogether；The many multicasts of period modulation
The priority of tree, it is achieved that dynamic load leveling；When breaking down, by fault multicast tree priority zero setting, it is achieved that
The reliability of multicast service.
Obviously, those skilled in the art should be understood that each module of the above-mentioned present invention or each step can be with general
Calculating device to realize, they can concentrate on single calculating device, or be distributed in multiple calculating device and formed
Network on, alternatively, they can realize, it is thus possible to by them with calculating the executable program code of device
Storage is performed by calculating device in the storage device, and in some cases, can hold with the order being different from herein
Step shown or described by row, or they are fabricated to respectively each integrated circuit modules, or by many in them
Individual module or step are fabricated to single integrated circuit module and realize.So, the present invention is not restricted to any specific hardware
Combine with software.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for the technology of this area
For personnel, the present invention can have various modifications and variations.All within the spirit and principles in the present invention, that is made is any
Amendment, equivalent, improvement etc., should be included within the scope of the present invention.
Claims (11)
1. a multicast control method, it is characterised in that including:
Detection network topology and link information；
Multiple multicast tree are generated according to described network topology；
According to described network topology and described link information, calculate each multicast tree in the plurality of multicast tree respectively
Weight, described weight selects the priority of corresponding multicast tree transmission data for indicating switch polls；
Issue the described weight of described each multicast tree to the switch in corresponding multicast tree.
Method the most according to claim 1, it is characterised in that according to described network topology and described link information,
Calculating respectively in the plurality of multicast tree after the described weight of each multicast tree, described method also includes:
Detect link information in real time；
According to described network topology and the link information of detection in real time, update the described power of described each multicast tree respectively
Weight；
Issue the described weight of renewal to the described switch in corresponding multicast tree.
Method the most according to claim 2, it is characterised in that according to the link letter of described network topology and detection in real time
Breath, the described weight updating described each multicast tree respectively includes:
In the case of determining link failure by the link information of detection in real time, by corresponding for the link of fault group
Broadcast the described weight zero setting of tree.
Method the most according to claim 1, it is characterised in that according to described network topology and described link information, point
Do not calculate the described weight of each multicast tree in the plurality of multicast tree to include:
According to described link information and default multicast tree weight rule, calculate the weight of described each multicast tree respectively；
According to described network topology, it is judged that whether the plurality of multicast tree exists overlay path；
In the case of determining and there is overlay path, according to the weight of multiple multicast tree of described overlay path, meter
Calculate the weight of overlay links multicast tree, and using many as described overlay path for the weight of described overlay links multicast tree
The weight of individual multicast tree.
Method the most according to claim 1, it is characterised in that issue the described weight of described each multicast tree to accordingly
Described switch in multicast tree includes:
Issue select type group table to the switch in corresponding multicast tree, wherein, in described select type group table
Carry corresponding multicast tree goes out port and the described weight of corresponding multicast tree.
Method the most according to claim 1, it is characterised in that issuing the described weight of described each multicast tree to phase
After answering the described switch in multicast tree, described method also includes:
Described switch receives the described weight of at least one multicast tree；
Described switch selects a multicast tree transmission at least one multicast tree described according to described weight poll
Data.
7. a multicast controls device, it is characterised in that including:
Detection module, is used for detecting network topology and link information；
Generation module, for generating multiple multicast tree according to described network topology；
Computing module, for according to described network topology and described link information, calculates the plurality of multicast tree respectively
In the weight of each multicast tree, described weight selects the excellent of corresponding multicast tree transmission data for indicating switch polls
First level；
Issue module, for issuing the described weight of described each multicast tree to the switch in corresponding multicast tree.
Device the most according to claim 7, it is characterised in that
Described detection module, is additionally operable to detect in real time link information；
Described computing module, is additionally operable to, according to described network topology and the link information of detection in real time, update institute respectively
State the described weight of each multicast tree；
Described issue module, be additionally operable to the described weight issuing renewal to the described switch in corresponding multicast tree.
Device the most according to claim 8, it is characterised in that
Described computing module, is additionally operable in the case of determining link failure by the link information of detection in real time,
By the described weight zero setting of corresponding for the link of fault multicast tree.
Device the most according to claim 7, it is characterised in that described computing module includes:
First computing unit, for according to described link information and default multicast tree weight rule, calculates described respectively
The weight of each multicast tree；
Judging unit, for according to described network topology, it is judged that whether the plurality of multicast tree exists overlay path；
Second computing unit is in the case of determining and there is overlay path, many according to described overlay path
The weight of individual multicast tree, calculates the weight of overlay links multicast tree, and the weight of described overlay links multicast tree is made
Weight for multiple multicast tree of described overlay path.
11. devices according to claim 7, it is characterised in that
Described issue module, be additionally operable to issue select type group table to the switch in corresponding multicast tree, wherein,
Carry corresponding multicast tree in described select type group table goes out port and the described weight of corresponding multicast tree.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510374309.1A CN106330730A (en) | 2015-06-30 | 2015-06-30 | Multicast control method and apparatus |
| PCT/CN2016/081758 WO2017000677A1 (en) | 2015-06-30 | 2016-05-11 | Multicast control method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510374309.1A CN106330730A (en) | 2015-06-30 | 2015-06-30 | Multicast control method and apparatus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106330730A true CN106330730A (en) | 2017-01-11 |
Family
ID=57607778
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510374309.1A Withdrawn CN106330730A (en) | 2015-06-30 | 2015-06-30 | Multicast control method and apparatus |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106330730A (en) |
| WO (1) | WO2017000677A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120537A (en) * | 2017-06-23 | 2019-01-01 | 迈普通信技术股份有限公司 | Multicast tree calculation method and device |
| CN109218180A (en) * | 2018-08-24 | 2019-01-15 | 刘昱 | Multicast control method and device based on Local Area Network |
| US11196588B2 (en) | 2019-02-12 | 2021-12-07 | Cisco Technology, Inc. | Redundant multicast tree in a fat tree network topology |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107645647A (en) * | 2017-09-21 | 2018-01-30 | 京信通信系统（中国）有限公司 | A kind of multichannel audio-video frequency transmission method and device |
| CN110278158B (en) * | 2019-08-09 | 2024-01-30 | 京信网络系统股份有限公司 | Multicast data packet transmission method, computer device, and storage medium |
| CN112996045B (en) * | 2021-02-04 | 2022-11-29 | 深圳市吉祥腾达科技有限公司 | Method and system for forwarding dimension link packet of mesh multi-frequency networking path |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299671A (en) * | 2007-04-30 | 2008-11-05 | 上海贝尔阿尔卡特股份有限公司 | Method and device for transmitting and receiving multicast data packet |
| CN101453345A (en) * | 2007-12-07 | 2009-06-10 | 上海贝尔阿尔卡特股份有限公司 | Multicast access control method based on service provider and access equipment thereof |
| CN101965031A (en) * | 2010-05-19 | 2011-02-02 | 北京科技大学 | Maximum probability-based cognitive radio multi-path multicast routing method |
| CN101986628A (en) * | 2010-12-15 | 2011-03-16 | 山东大学 | Method for realizing multisource multicast traffic balance based on ant colony algorithm |
| CN102938920A (en) * | 2012-11-21 | 2013-02-20 | 山东大学 | Movable multicast routing method based on cognitive Ad hoc network |
| CN103716208A (en) * | 2013-12-31 | 2014-04-09 | 北京邮电大学 | Network management method and system supporting elephant flow, interchanger and network |
| CN104683244A (en) * | 2015-03-09 | 2015-06-03 | 东北大学 | Multicast routing method based on route node driving strategy |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101436945B (en) * | 2008-12-25 | 2012-05-23 | 中兴通讯股份有限公司 | Method, system and apparatus for protecting multicast business fault |
| JP2013098647A (en) * | 2011-10-28 | 2013-05-20 | Nippon Telegr & Teleph Corp <Ntt> | Routing method in multicast communication |
| CN103139265B (en) * | 2011-12-01 | 2016-06-08 | 国际商业机器公司 | Network adaptation transmitter optimization method in massive parallel processing and system |
| US9338096B2 (en) * | 2013-08-30 | 2016-05-10 | Futurewei Technologies, Inc. | Multicast tree packing for multi-party video conferencing under SDN environment |
| CN104038946B (en) * | 2014-07-08 | 2017-05-10 | 吉林大学 | Partially-overlapped channel allocation method for multicast of wireless mesh network |
-
2015
- 2015-06-30 CN CN201510374309.1A patent/CN106330730A/en not_active Withdrawn
-
2016
- 2016-05-11 WO PCT/CN2016/081758 patent/WO2017000677A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299671A (en) * | 2007-04-30 | 2008-11-05 | 上海贝尔阿尔卡特股份有限公司 | Method and device for transmitting and receiving multicast data packet |
| CN101453345A (en) * | 2007-12-07 | 2009-06-10 | 上海贝尔阿尔卡特股份有限公司 | Multicast access control method based on service provider and access equipment thereof |
| CN101965031A (en) * | 2010-05-19 | 2011-02-02 | 北京科技大学 | Maximum probability-based cognitive radio multi-path multicast routing method |
| CN101986628A (en) * | 2010-12-15 | 2011-03-16 | 山东大学 | Method for realizing multisource multicast traffic balance based on ant colony algorithm |
| CN102938920A (en) * | 2012-11-21 | 2013-02-20 | 山东大学 | Movable multicast routing method based on cognitive Ad hoc network |
| CN103716208A (en) * | 2013-12-31 | 2014-04-09 | 北京邮电大学 | Network management method and system supporting elephant flow, interchanger and network |
| CN104683244A (en) * | 2015-03-09 | 2015-06-03 | 东北大学 | Multicast routing method based on route node driving strategy |
Non-Patent Citations (1)
| Title |
|---|
| 王另秀，曹叶文: ""基于蚁群算法的多源组播流量均衡的研究"", 《山东大学学报（理学版）》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120537A (en) * | 2017-06-23 | 2019-01-01 | 迈普通信技术股份有限公司 | Multicast tree calculation method and device |
| CN109120537B (en) * | 2017-06-23 | 2020-10-16 | 迈普通信技术股份有限公司 | Multicast tree calculation method and device |
| CN109218180A (en) * | 2018-08-24 | 2019-01-15 | 刘昱 | Multicast control method and device based on Local Area Network |
| US11196588B2 (en) | 2019-02-12 | 2021-12-07 | Cisco Technology, Inc. | Redundant multicast tree in a fat tree network topology |
| US11271774B2 (en) | 2019-02-12 | 2022-03-08 | Cisco Technology, Inc. | Redundant multicast tree in a fat tree network topology with anisotropic routing |
| US11290302B2 (en) | 2019-02-12 | 2022-03-29 | Cisco Technology, Inc. | Distribution of data packet via overlay tunnel connecting multicast trees utilizing a fat tree network topology |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2017000677A1 (en) | 2017-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11178001B2 (en) | Multi-stage switch fabric fault detection and handling | |
| CN106330730A (en) | Multicast control method and apparatus | |
| CN104104718B (en) | User autonomous routing customization system based on software-defined network | |
| CN105959232B (en) | A kind of satellite network method for routing based on software defined network control points optimization | |
| US8040906B2 (en) | Utilizing betweenness to determine forwarding state in a routed network | |
| EP2911348A1 (en) | Control device discovery in networks having separate control and forwarding devices | |
| EP2608459B1 (en) | Router, virtual cluster router system and establishing method thereof | |
| GB2483563A (en) | Dual homing connection of access network to transport network | |
| EP3029893A1 (en) | Method, controller, device and system for protecting service path | |
| CN108809847B (en) | Method, device and network system for realizing load balance | |
| EP2883334B1 (en) | Techniques for flooding optimization for link state protocols in a network topology | |
| CN103795805A (en) | Distributed server load balancing method based on SDN | |
| CN104335537A (en) | System and method for layer-2 multicast multipathing | |
| CN103891216A (en) | Fhrp optimizations for n-way gateway load balancing in fabric path switching networks | |
| CN102347905A (en) | Network equipment and forwarded information updating method | |
| CN106452958A (en) | Flow control method and system, and centralized controller | |
| CN110324159B (en) | Link configuration method, controller and storage medium | |
| US20140036726A1 (en) | Network, data forwarding node, communication method, and program | |
| CN104871490A (en) | Multi-path communication device capable of improving energy use efficiency and traffic distribution method for improving energy use efficiency thereof | |
| CN102413059A (en) | Multicast forwarding method based on SPB (Shortest Path Bridging) network and SPBM (MAC-in-MAC SPB) bridge | |
| CN108337179A (en) | Link flow control method and device | |
| CN102158422B (en) | Message forwarding method and equipment for layer 2 ring network | |
| CN103532840A (en) | Link switching method and link switching device | |
| CN104935509A (en) | Method and apparatus of avoiding loop generation in vertical stack system | |
| CN107682265B (en) | Message routing method and device of payment system |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20170111 |
|
| WW01 | Invention patent application withdrawn after publication |