The dispatching method and device of aggregated links in a kind of packet network
Technical field
The present invention relates to a kind of dispatching parties of aggregated links in packet network technical field more particularly to packet network
Method and device.
Background technology
Growth with data business volume and the raising to quality of service requirement, high bandwidth offer have become packet switching network
The most important feature of network.Link aggregation is an important technology for meeting network high bandwidth requirements.
Link aggregation technology (having regulation description in IEEE 802.3ad) is by several physical link groups of two equipment rooms
A link (an also referred to as aggregated links) on combinator, which is logically an entirety, internal composition
Details with transmission data is transparent to top service.
In the prior art, it can be assigned to when the LSP (label switched path) for creating or breaking down recoveries to allow
Resource is to carry out seizing for resource according to the height of priority, to be provided when the newly-built or LSP that breaks down being made to restore
Source.Defect there are two doing so:First, the resource for finding low priority is just seized, the cost being preempted is not weighed;Its
Two, the processing that the LSP (s) being preempted is not got well.
In order to efficiently use aggregated link bandwidth and avoid service disconnection loss of data, Chinese patent as far as possible
CN200810025657.8 " a kind of method and link port for realizing that link is seized " proposes following scheme:It is same when detecting
On aggregated links port the high link of priority will preempting priority low link when, be in using the high link of high priority
Be delayed race to control state;After the high link of priority, which meets link state, to be stablized, the chain that keeps the priority of delay race to control state high
Road is placed in the link low by chain road preempting priority is robbed.The advantages of program is the high priority link that can make in aggregated links
Business data flow is avoided to lose when seizing, but since it does not account for seizing the cost and robbed that low priority link is spent
Whether the low priority link accounted for can seize or obtain again resource situation, but it is low only to consider how high priority is seized
Priority link and data are not lost, therefore the cost that can still result in cost is larger and the user of other chain roads being preempted
Loss of data.
In packet network, the technologies such as MPLS, MPLS-TP, PBB, VLAN exchange is used to realize TE flow works at present
Journey.According to current TE Technology of Traffic Engineering, control plane and data planar separation, before transmitting user data, service carries
Donor requires to formulate SLA (service level agreement) first according to user, and SLA includes to meet user QoS (service quality) requirements
Different service flows bandwidth parameter, and control plane is according to the bandwidth parameter and current network of pre-determined different service flows
Resource behaviour in service configures the forward-path of service flow to data plane, and changes in the resource using status of network
When (deletion or increase of such as LSP, network link increase or decrease), further modification has the forward-path of service flow and leads to
Primary data plane carries out data forwarding when data plane receives service flow data packet according to the forward-path that control plane configures,
To can guarantee that the qos requirement of user service stream can optimize the use to Internet resources again, so TE Technology of Traffic Engineering mesh
The preceding application in carrier network is more and more extensive.
Since ISP needs to be charged for the different grades of service and different bandwidth that different user provides according to it,
Therefore, how to ensure the service quality of high priority and one of the important feature of TE Technology of Traffic Engineering.Further, such as
Fruit control plane can make full use of this feature, ensure that the business of high priority can be obtained when spending less cost as possible
Link aggregation function is realized in quality services, and the business acquisition for solving lower priority cannot be filled less than resource and bandwidth
Divide the shortcomings that utilizing, is also very necessary in packet network.
Invention content
The technical problem to be solved by the present invention is to provide the dispatching method and device of aggregated links in a kind of packet network, energy
Enough make preemption cost when scheduling minimum, and attempt to reschedule resource to the LSP being preempted, to effectively realize that link is poly-
High efficient and reliable is closed, ensures the quality requirement of each LSP.
In order to solve the above technical problems, the present invention provides a kind of dispatching method of aggregated links in packet network,
Currently wait for that the label switched path (LSP) of preempting resources judges whether each Member Link meets and seizes item one by one
Part, and select from the one or more Member Links for meeting preemptive condition preemption cost it is minimum as the LSP groups being preempted.
Further, the method further includes:
After selecting the LSP groups being preempted, the LSP additions in the LSP groups being preempted are waited for into preempting resources
In LSP queues;
Wherein, the LSP groups being preempted include one or more LSP.
Further, the preemptive condition includes:The available bandwidth of the Member Link currently waits seizing more than or equal to described
The guaranteed bandwidth of the LSP of resource.
Further, the available bandwidth of the Member Link, including:The idle bandwidth of the Member Link and the member
The summation of the not reserved bandwidth of available LSP in link.
Further, it when judging whether each Member Link meets preemptive condition, specifically includes:
If the LSP of the lowest priority of the Member Link is built at or above the current LSP's for waiting for preempting resources
Vertical priority, then judge that the Member Link is unsatisfactory for preemptive condition;
If the LSP of the lowest priority of the Member Link establishes priority less than the current LSP's for waiting for preempting resources,
It is then successively that the LSP of lowest priority in the Member Link to priority is excellent less than establishing for the LSP for currently waiting for preempting resources
The not reserved bandwidth of the LSP of first grade is set to available, the sky up to being set to available total not reserved bandwidth and the Member Link
The summation of not busy bandwidth is more than or equal to the guaranteed bandwidth of the current LSP for waiting for preempting resources, then judges that Member Link satisfaction is robbed
Account for condition;If all idle bandwidths for being set to available total not reserved bandwidth and the Member Link in the Member Link
Summation still less than the guaranteed bandwidth of the current LSP for waiting for preempting resources, then judge that the Member Link is unsatisfactory for preemptive condition.
Further, the LSP groups of the preemption cost minimum refer to:The number for the LSP being preempted is minimum;Alternatively, by robbing
The priority of the LSP accounted for is low;Alternatively, it is most short to seize the time.
The present invention also provides a kind of dispatching device of aggregated links in packet network, described device includes:Resource is looked into
Module and schedule process module are ask, wherein
The resource query module is used for, and is inquired and is judged whether each Member Link meets the LSP's for waiting for preempting resources
Preemptive condition, and the one or more Member Links for meeting preemptive condition are returned into the schedule process module;
The schedule process module is used for, and preemption cost is selected from the one or more Member Links for meeting preemptive condition
It is minimum as the LSP groups being preempted.
Further, the schedule process module is additionally operable to, and after selecting the LSP groups being preempted, is robbed described
LSP in the LSP groups accounted for is added in the LSP queues for waiting for preempting resources;
Wherein, the LSP groups being preempted include one or more LSP.
Further, the resource query module is used for, when judging whether each Member Link meets preemptive condition, tool
Body includes:
If the LSP of the lowest priority of the Member Link is built at or above the current LSP's for waiting for preempting resources
Vertical priority, then judge that the Member Link is unsatisfactory for preemptive condition;
If the LSP of the lowest priority of the Member Link establishes priority less than the current LSP's for waiting for preempting resources,
It is then successively that the LSP of lowest priority in the Member Link to priority is excellent less than establishing for the LSP for currently waiting for preempting resources
The not reserved bandwidth of the LSP of first grade is set to available, the sky up to being set to available total not reserved bandwidth and the Member Link
The summation of not busy bandwidth is more than or equal to the guaranteed bandwidth of the current LSP for waiting for preempting resources, then judges that Member Link satisfaction is robbed
Account for condition, and to the schedule process module return reserved bandwidth in the Member Link, and/or the Member Link be set to it is available
LSP groups;If all idle bands for being set to available total not reserved bandwidth and the Member Link in the Member Link
Wide summation then judges that the Member Link is unsatisfactory for seizing item still less than the guaranteed bandwidth of the current LSP for waiting for preempting resources
Part.
Further, the LSP groups of the preemption cost minimum of the schedule process module selection being preempted, including:It is robbed
The number of the LSP accounted for is minimum;The priority for the LSP being preempted is low;Alternatively, it is most short to seize the time.
It cannot effectively solve to spend for existing link aggregation FF (fixed filtering) resource reservation mode dispatching method less
The shortcomings that processing that cost and the LSP being preempted are not got well, the present invention proposes a kind of new scheduling scheme, according to every
The LSP situations of Member Link find a kind of dispatching method of preemption cost minimum, and attempt to adjust again to the LSP being preempted
Resource is spent, to effectively realize link aggregation high efficient and reliable, ensures the quality requirement of each LSP.
Description of the drawings
Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, this hair
Bright illustrative embodiments and their description are not constituted improper limitations of the present invention for explaining the present invention.In the accompanying drawings:
Fig. 1 is the network diagram of Member Link and LSP in an aggregated links;
Fig. 2 is the flow chart of link aggregation dispatching method during the LSP of the embodiment of the present invention is created.
Specific implementation mode
Present embodiment provides a kind of dispatching method of aggregated links in packet network, using following scheme:
It currently waits for that the LSP of preempting resources judges whether each Member Link meets preemptive condition one by one, and is seized from satisfaction
Selected in one or more Member Links of condition preemption cost it is minimum as the LSP groups being preempted.
Wherein, the LSP groups being preempted include one or more LSP (s), and further, the method further includes:
After selecting the one or more LSP (s) being preempted, the LSP (s) being preempted is added and waits for preempting resources
In LSP lists.
Further, the preemptive condition includes:The available bandwidth of the Member Link currently waits seizing more than or equal to described
The guaranteed bandwidth of the LSP of resource.
Further, the available bandwidth, specially:It can in the idle bandwidth of the Member Link and the Member Link
The summation of the not reserved bandwidth of LSP.
Further, it when judging whether each Member Link meets preemptive condition, specifically includes:
If the LSP of the lowest priority of the Member Link is built at or above the current LSP's for waiting for preempting resources
Vertical priority, then judge that the Member Link is unsatisfactory for preemptive condition;
If the LSP of the lowest priority of the Member Link establishes priority less than the current LSP's for waiting for preempting resources,
The LSP of lowest priority in the Member Link to priority is then less than building for the current LSP for waiting for preempting resources successively
The not reserved bandwidth of the LSP of vertical priority be set to it is available, until being set to available total not reserved bandwidth and the Member Link
The summation of idle bandwidth be more than or equal to the guaranteed bandwidth of the current LSP for waiting for preempting resources, then judge that the Member Link is full
Sufficient preemptive condition;If all in the Member Link be set to available total not reserved bandwidth (i.e. all priority be less than institute
State the total of the LSP for establishing priority of the current LSP for waiting for preempting resources and do not reserve bandwidth) idle band with the Member Link
Wide summation then judges that the Member Link is unsatisfactory for seizing item still less than the guaranteed bandwidth of the current LSP for waiting for preempting resources
Part.
Further, the preemption cost is minimum, including but not limited to:The number for the LSP being preempted is minimum;Alternatively,
The bandwidth seized in the case that the number of the LSP being preempted is minimum is minimum;The priority for the LSP being preempted is minimum;Alternatively, seizing
Resource utilization highest ensures that resource can be accessed by being preempted LSPs as far as possible;Spend the time minimum alternatively, seizing, most soon
Find resource, etc..
As shown in Fig. 2, present embodiment provide packet network in aggregated links dispatching method, specific steps are such as
Under:
Step 1, according to load balancing, idling-resource is inquired.N is entered step if inquiry returns successfully;If
Inquiry returns to failure and then enters step 2.
Step 2, the LSP of the lowest priority bandwidth occupied is set as " can be used ", then judges each Member Link
Available bandwidth whether meet condition.
If there is a Member Link meets, then N is entered step.
If there is a plurality of Member Link meets, then a Member Link is selected, the LSP that the Member Link is preempted is met
(s) number is minimum, enters step N.
If met without Member Link, 3 are entered step.
Step 3, the bandwidth LSP of next lowest priority occupied, is arranged to " can be used ", calculating process is the same as step 2.
Preferential level height (that is, being set to available LSP priority equal to or higher than before, but is less than and is established by step N-1
Priority) LSP occupy bandwidth, be arranged to " can be used " temporarily, calculating process is with step 2, if at this time still without member's chain
Road meets condition, then entirely seizes scheduling process failure.
Step N is seized and is dispatched successfully, the LSP (s) for returning to Member Link and being preempted.If there is the LSP (s) being preempted,
It is then added into the LSP lists (also referred to as pending LSP lists) for waiting for preempting resources, waits pending.
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention
Embodiment be described in detail.It should be noted that in the absence of conflict, in the embodiment and embodiment in the application
Feature mutually can arbitrarily combine.
In following all implementations are described in detail, there is following agreement describing mode:
1) Linka-b indicates the Member Link of an aggregated links (also referred to as logical links) Linka, which is also one
A physical link;
2) BW1 in Linka-b (BW1, BW2) indicates that bandwidth, BW2 can be used to indicate the also not pre- of link for the physical link
The usable bandwidth stayed does not reserve bandwidth also referred to as.
3) MaxLspBW in Linka (MaxLspBW) indicates the not reserved maximum LSP bandwidth of logical links Linka.For
Incorrect order occurs in avoiding user data package from transmitting, provides not allow a LSP that can transmit across multiple Member Links simultaneously in the present invention,
Therefore MaxLspBW should be the maximum value in the not reserved bandwidth of all Member Links.If the guaranteed bandwidth of a LSP is BW0, when
When BW0 <=MaxLspBW, which could pass through the logical links.Logical links Linka also has other describable bandwidth ginsengs
" maximum link bandwidth (bandwidth that can be used), maximum reservable bandwidth, not reserved bandwidth be not (pre- on each Member Link for number
Stay the sum of bandwidth) " etc..
4) LSP (Prior, CIR, link ID) indicates some key feature value informations of a LSP, wherein:Prior tables
Show foundation/holding priority level (being holding priority after foundation to establish priority when establishing);CIR indicates guaranteed bandwidth;
Link ID indicate Member Link marks of the LSP Jing Guo aggregated links.
Embodiment one
This embodiment describes the dispatching method of the link aggregation " when creating unidirectional LSP by an aggregated links "
Realization process.
It is assumed that in the present embodiment, the current principal states information of aggregated links is made of following three class value:
1) Link1-1 (100M, 0M), Link1-2 (100M, 0M);
2)Link1(0M);
3) LSP1 (1,20M, Link1-1), LSP2 (3,20M, Link1-1), LSP3 (5,30M, Link1-1), LSP4 (5,
30M, Link1-1), LSP5 (1,30M, Link1-2), LSP6 (3,30M, Link1-2), LSP7 (5,10M, Link1-2), LSP8
(5,10M, Link1-2), LSP9 (5,20M, Link1-2).
LSP10 (it is 4, CIR 30M that it, which establishes priority level) is established by distributed RSVP-TE signalings in present network,
It is needed by Link1 by router-level topology LSP10, during establishing LSP10 using distributed signaling mode, when LSP believes
When order reaches aggregated links endpoint, refering to what is shown in Fig. 2, the step of LSP implements link aggregation control technology function during creating
Flow chart is as follows:
Step 1, LSP10 (4,30M, Link1) establishes signaling and reaches node A, according to the scheduling request information, determines LSP
Link Link1 need to be passed through, and (and LSP is in local outbound.).
It inquires Link1 (0M), and it is 30M that LSP10, which needs CIR, without idling-resource, needs to seize and enter step 2.
Step 2, for two LSP of lowest priority in Link1-1 (it is 5 to keep priority):LSP3 (5,30M,
The LSP4 bandwidth occupied is temporarily set as " can be used ", then Link1-1 by Link1-1), LSP4 (5,30M, Link1-1) first
The available bandwidth for having 30M in Member Link, it is the condition that 4, CIR is 30M to meet LSP10 to establish priority;Others are checked again
Whether Member Link meets preemptive condition;
For three LSP of lowest priority in Link1-2 (keep priority be 5), LSP7 (5,10M, Link1-2),
The bandwidth of LSP9 is temporarily set as " can be used " by LSP8 (5,10M, Link1-2), LSP9 (5,20M, Link1-2) first, then
The available bandwidth for having 20M in Link1-1 Member Links is unsatisfactory for LSP10 and establishes priority to be the condition that 4, CIR is 30M;Continue
The bandwidth of LSP8 is temporarily set as " can be used ", then there is the available bandwidth of 30M in Link1-2 Member Links, meets LSP10's
Preemptive condition;
Comparing the cost that Member Link Link1-1 and Link1-2 are spent, the LSP numbers that Link1-1 is seized are minimum, because
This, enters step N.
Step N is seized and is dispatched successfully, the LSP (s) for returning to Member Link and being preempted:
Link1-1 (LSP1 (1,20M, Link1-1), LSP2 (3,20M, Link1-1), LSP10 (4,30M, Link1-1),
LSP3 (5,30M, Link1-1));
Pending LSP lists are added in the LSP4 (5,30M, --) being preempted, and wait pending.
Embodiment two
This embodiment describes what the link aggregation " when creating unidirectional LSP by an aggregated links " was preempted
LSP dispatches realization process again.
It is assumed that the current principal states information of the present embodiment aggregated links is made of following three class value:
1) Link1-1 (50M, 0M), Link1-2 (50M, 10M);
2)Link1(10M);
3) LSP1 (3,10M, Link1-1), LSP2 (4,20M, Link1-1), LSP3 (4,10M, Link1-1), LSP4 (5,
10M, Link1-1), LSP5 (2,20M, Link1-2), LSP6 (3,10M, Link1-2), LSP7 (6,10M, Link1-2).
LSP8 (it is 3, CIR 30M that it, which establishes priority level) is established by distributed RSVP-TE signalings in present network,
It is needed by Link1 by router-level topology LSP8, during establishing LSP8 using distributed signaling mode, when LSP signalings
When reaching aggregated links endpoint, refering to what is shown in Fig. 2, LSP flows the step of implementing link aggregation control technology function during creating
Journey figure is as follows:
Step 1, LSP8 (3,30M, Link1) establishes signaling and reaches node A, according to the scheduling request information, determines that LSP is needed
By link Link1 (and LSP is in local outbound).
It inquires Link1 (10M), and it is 30M that LSP8, which needs CIR, without idling-resource, needs to seize and enter step 2.
Step 2, for Link1-1, a LSP of lowest priority (it is 5 to keep priority):LSP4 (5,10M,
Link1-1), first the LSP4 bandwidth occupied is temporarily set as " can be used ", then have 10M in Link1-1 Member Links can
With bandwidth, it is unsatisfactory for the requirement of LSP8;
For available free bandwidth 10M in Link1-2, it is unsatisfactory for the set up the condition of LSP8, lowest priority (keeps priority
For a LSP 6):LSP7 (6,10M, Link1-2), by LSP7 occupy bandwidth be temporarily set as " can be used ", at this time can be used
Bandwidth is:Idle bandwidth 10M does not still meet the condition of LSP8 plus the not reserved bandwidth total 20M of 10M of LSP7, remaining
The minimum holding priority of LSP is 3, establishes that priority is identical, and therefore, Member Link Link1-2 is simultaneously unsatisfactory for LSP8 with LSP8
Preemptive condition.
Step 3, continue a LSP of next lowest priority (it is 4 to keep priority) for Link1-1:By LSP2 (4,
20M, Link1-1) occupy LSP2 bandwidth be temporarily set as " can be used ", then in Link1-1 Member Links share 30M can be used
Bandwidth meets the condition of LSP8;In addition, also a kind of method seized, i.e., seize LSP3 (4,10M, Link1-1), LSP2 again
(4,20M, Link1-1), choose a kind of Least-cost i.e.-kind, enters step N.
Step N is seized and is dispatched successfully, the LSP (s) for returning to Member Link and being preempted:
Link1-1 (LSP1 (3,10M, Link1-1), LSP3 (4,10M, Link1-1), LSP8 (3,30M, Link1-1))
Pending list is added in the LSP2 (4,20M, --) that is preempted, LSP4 (5,10M, --), waits pending.
Hereafter, the LSP in pending list enters step 1 again.Specific processing is as follows:
Step 1, it is 4, CIR 20M to wait pending LSP2 to establish priority, Link1 (10M) is inquired, without enough
Idling-resource, need to seize and enter step 2.
Step 2, for there is available bandwidth 10M in Link1-2, the set up the condition of LSP2 is not met, lowest priority (is protected
It is a LSP 6) to hold priority:The LSP7 bandwidth occupied is temporarily set as " can be used " by LSP7 (6,10M, Link1-2), full
The condition of sufficient LSP2, so Member Link Link1-2 meets the condition of LSP8.
Step N is seized and is dispatched successfully, the LSP (s) for returning to Member Link and being preempted:
Link1-2 (LSP7 (2,20M, Link1-2), LSP6 (3,10M, Link1-2), LSP2 (4,20M, Link1-2));
Pending list is added in the LSP7 (6,20M, --) that is preempted, LSP4 (5,10M, --), waits pending.
It should be noted that being that the present invention reschedules model for the LSP for seizing minimum LSP numbers and being preempted above
The feasible specific example of description is enclosed, the example is not the protection domain for limiting patent of the present invention, and the technology of the present invention may be used on
Minimum cost (is included in and seizes minimal bandwidth under the premise of seizing minimum LSP numbers, preempting priority is weighed with LSP numbers are seized
Recanalization) and any particular technology range such as reschedule.
In addition, above-described embodiment is the processing seized for the aggregated links FF modes of control plane level, wherein
Trigger condition to the control of the control plane of link circuit resource scheduling can be scheduling request, network management configuration request etc..As for poly-
The contents such as link bandwidth link adjustment bottom layer realization process and SE (shared filtering) preempt-mode are closed, are no longer retouched in detail
It states.And foregoing description is to be described in detail to carrying out the minimum method of LSP numbers change on the basis of priority-based again,
But other costs calculating also should be within present invention protection.
From said program it can be seen that:The present invention can find one and be preempted according to the actual conditions of different members link
The minimum Member Link of LSP numbers, and attempt to reschedule resource to the LSP being preempted, to effectively realize link aggregation
High efficient and reliable, ensures the quality requirement of each LSP, and seeming in the packet network with traffic-engineering requirements extremely has
With.
In addition, additionally providing a kind of dispatching device of aggregated links in packet network, the device in the embodiment of the present invention
Include mainly:Resource query module and schedule process module, wherein
The resource query module is used for, and is inquired and is judged whether each Member Link meets the LSP's for waiting for preempting resources
Preemptive condition, and the one or more Member Links for meeting preemptive condition are returned into the schedule process module;
The schedule process module is used for, and preemption cost is selected from the one or more Member Links for meeting preemptive condition
It is minimum as the LSP groups being preempted.
Further, the schedule process module is additionally operable to, and after selecting the LSP groups being preempted, is robbed described
LSP in the LSP groups accounted for is added in the LSP queues for waiting for preempting resources;
Wherein, the LSP groups being preempted include one or more LSP.
Further, the resource query module is used for, when judging whether each Member Link meets preemptive condition, tool
Body includes:
If the LSP of the lowest priority of the Member Link is built at or above the current LSP's for waiting for preempting resources
Vertical priority, then judge that the Member Link is unsatisfactory for preemptive condition;
If the LSP of the lowest priority of the Member Link establishes priority less than the current LSP's for waiting for preempting resources,
It is then successively that the LSP of lowest priority in the Member Link to priority is excellent less than establishing for the LSP for currently waiting for preempting resources
The not reserved bandwidth of the LSP of first grade is set to available, the sky up to being set to available total not reserved bandwidth and the Member Link
The summation of not busy bandwidth is more than or equal to the guaranteed bandwidth of the current LSP for waiting for preempting resources, then judges that Member Link satisfaction is robbed
Account for condition, and to the schedule process module return reserved bandwidth in the Member Link, and/or the Member Link be set to it is available
LSP groups;If all idle bands for being set to available total not reserved bandwidth and the Member Link in the Member Link
Wide summation then judges that the Member Link is unsatisfactory for seizing item still less than the guaranteed bandwidth of the current LSP for waiting for preempting resources
Part.
Further, the LSP groups of the preemption cost minimum of the schedule process module selection being preempted, including:It is robbed
The number of the LSP accounted for is minimum;The priority for the LSP being preempted is low;Alternatively, it is most short to seize the time.
The preferred implementation case that these are only the present invention, is not intended to restrict the invention, the present invention can also have other
Various embodiments, without deviating from the spirit and substance of the present invention, those skilled in the art can be according to this hairs
It is bright to make various corresponding change and deformations, but these corresponding change and deformations should all belong to appended claims of the invention
Protection domain.
Obviously, those skilled in the art should be understood that each module of the above invention or each step can be with general
Computing device realize that they can be concentrated on a single computing device, or be distributed in multiple computing devices and formed
Network on, optionally, they can be realized with the program code that computing device can perform, it is thus possible to which they are stored
It is performed by computing device in the storage device, and in some cases, it can be with different from shown in sequence execution herein
The step of going out or describing, either they are fabricated to each integrated circuit modules or by them multiple modules or
Step is fabricated to single integrated circuit module to realize.In this way, the present invention is not limited to any specific hardware and softwares to combine.