[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN100486219C - A method to realize the end-to-end stream transmission - Google Patents

A method to realize the end-to-end stream transmission Download PDF

Info

Publication number
CN100486219C
CN100486219C CNB2005101090561A CN200510109056A CN100486219C CN 100486219 C CN100486219 C CN 100486219C CN B2005101090561 A CNB2005101090561 A CN B2005101090561A CN 200510109056 A CN200510109056 A CN 200510109056A CN 100486219 C CN100486219 C CN 100486219C
Authority
CN
China
Prior art keywords
resource
node
message
intermediate node
resource request
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.)
Expired - Fee Related
Application number
CNB2005101090561A
Other languages
Chinese (zh)
Other versions
CN1953415A (en
Inventor
陈洪飞
李斌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to CNB2005101090561A priority Critical patent/CN100486219C/en
Priority to PCT/CN2006/002730 priority patent/WO2007045163A1/en
Publication of CN1953415A publication Critical patent/CN1953415A/en
Application granted granted Critical
Publication of CN100486219C publication Critical patent/CN100486219C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/10Flow control; Congestion control
    • H04L47/12Avoiding congestion; Recovering from congestion
    • H04L47/125Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/10Flow control; Congestion control
    • H04L47/18End to end
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/70Admission control; Resource allocation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/70Admission control; Resource allocation
    • H04L47/72Admission control; Resource allocation using reservation actions during connection setup
    • H04L47/726Reserving resources in multiple paths to be used simultaneously

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention relates to a flow transmission method from terminal to terminal. Wherein, it is characterized in that: performing resource for said flow in several paths between sender and receiver; transmitting the flow from the sender to the receiver. The invention can confirm the consistence between preformed path and data flow path, to utilize the preformed resource of IPv6. The invention can utilize the load resource most, to improve the network utilization.

Description

A kind of realization is the flow transmission method end to end
Technical field
The present invention relates to the network communications technology field, be specifically related to a kind of realization flow transmission method end to end.
Background technology
Stream is meant one group of sequence of message between the specific source and destination, and these messages arrive identical destination from identical source, and have identical traffic identifier.The source requires intermediate node that these messages are done particular procedure by signaling or alternate manner.
Consider the needs of real time business, IPv6 (Internet protocol the 6th edition) is doing further improvement aspect the support of convection current and the resource reservation, defined stream label (Flow Label) territory of a 20bit in the ipv6 header, when the source sends message, if desired message is placed in the stream and transmits, need in the stream label territory, insert corresponding stream numbering.The stream label thresholding is that 0 message does not belong to any stream, is regarded as general message and handles.The tlv triple that the IPv6 node is formed according to the source address in the message header, destination address and stream label is discerned stream, and handles message in this stream according to the stream mode of having set up.
In the IPv6 network, because IPv6 extension header length and position is unfixing, and TCP/UDP (transmission control protocol/User Datagram Protoco (UDP)) port numbers is arranged in IPv6 load and owing to reasons such as burst or encryption are difficult to obtain, therefore, IPv6 uses tlv triple and can more effectively accelerate message classification than the five-tuple of being made up of source address, destination address, protocol type, TCP/UDP source port, TCP/UDP destination interface that uses among the IPv4, promptly accelerates the speed that stream is discerned; In addition, because IPv6 makes the message classification device only depend on the information of IP (Internet protocol) head, make to be easy to introduce new upper-layer protocol on the IPv6; Thereby, make IPv6 have remarkable advantages than IPv4.
IETF (Internet Engineering Task Force, internet task working group) IPv6 group has defined the purposes in stream label territory in RFC (draft) 3697, stipulated the minimum requirement that source node marked flows, forward node forwarding flow message and stream mode are set up; Wherein, about the method that the IPv6 stream mode is set up, RFC3697 has only done following 2 requirements:
1), need in all or part of IPv6 node of source to the destination path, set up the IPv6 stream mode in order convection current to carry out particular procedure.IPv6 stream mode method for building up and stream transaction module are stipulated in independent RFC.
2) for various IPv6 stream mode method for building up can and be deposited, these IPv6 stream mode method for building up must satisfy two basic demands: first, must provide stream mode to remove means, source node can be specified than default 120 seconds longer stream mode life cycles by signaling; The second, the stream mode method for building up must recover from the situation that the requirement stream mode can not be supported.
Though stipulated 2 basic demands among the RFC3697, do not stipulate concrete method and transaction module, and up to the present, the draft or the individual motion of this respect also do not occur the stream mode method for building up.
RSVP that IETF works out in RFC2205 (RSVP) and QoS (service quality) IntServ (integrated service model) that works out in RFC2210 can be used to IPv4/IPv6 stream to apply for resource on intermediate node, set up priority and safeguard soft state, and intermediate node adopts specific scheduling strategy to carry out particular procedure according to the soft state of stream to stream packet.That is to say, be that the stream packet of specific identifier produces a special passage in network, guarantees that it promptly transmits.Wherein, described specific identifier is meant the tlv triple of five-tuple or the IPv6 of IPv4.
The RSVP message generally is encapsulated in the Raw IP message, also can be encapsulated in the UDP message to send.RSVP and ICMP agreements such as (internet control messages agreements) is control protocol, and it does not deliver any application data and only transmits the reservation signaling parameter, and supports clean culture and multicast.RSVP intercepts the RSVP message in order to notify intermediate router, and does further processing as a new protocol type, has introduced the ip router alert options according to RFC2113 in the IP head.The RSVP message does not allow burst and reorganization, and the default refresh cycle of soft state is 30 seconds.
The design original intention of RSVP is the end-to-end QoS signaling protocol that designs for Intserv, goes up the real-time application data of transmission to be supported in Internet.Owing to think that when design RSVP the multicast application in real time of multiple spot is that the key that RSVP must support is used, and the multicast receiving terminal is distributed in different places, network condition each other may fall far short, so, the basic design philosophy of RSVP is: periodically send message and refresh protocol status on router and the main frame, be i.e. soft state, the exchange of two-way ground signaling message, receiving terminal is initiated resource reservation request, and qos signaling is independent of Routing Protocol.
Application program at first triggers RSVP, satisfies the resource of certain QoS for data flow request to be set up.In concrete resource request process, the transmission course of message is:
Setting transmitting terminal Src is connected with receiving terminal Dst by intermediate node R1, R2, R3.Transmitting terminal Src is forwarded to receiving terminal Dst by intermediate node R1, R2, R3 step by step with PATH (path) message, and receiving terminal Dst oppositely is sent to transmitting terminal Src with RESV (reservation) message of feedback step by step by original route.The concrete processing procedure of above-mentioned each node may further comprise the steps:
Step 1, transmitting terminal Src periodically sends PATH message along data flow path.
Step 2 after intermediate node such as R1, R2, R3 receive PATH message, obtains the reverse path information of data flow by the parameter in the analysis PATH message.
Step 3 after receiving terminal Dst receives PATH message, sends RESV message request resource reservation.
Step 4, each intermediate node is analyzed RESV message, and carry out policy control and admit control, after policy control and admittance control are all passed through, be data flow reserved resource and setting processing parameter, and according to the reverse path information that from PATH message, obtains, along the path forwarding RESV message reverse with data flow.
If in the default time period, intermediate node does not receive periodic PATH and RESV update, and then the processing parameter for data flow reserved resource and setting will be cancelled in the intermediate node.Transmitting terminal and receiving terminal also can be by sending dismounting message, and initiatively cancelling in the intermediate node is the processing parameter of data flow reserved resource and setting.
Because RSVP has favorable expansibility, therefore, reserving parameter and specific operation can be with object mode transparent enclosure in RSVP message, and different transaction modules can be expanded the reservation parameter and the specific operation of own needs.
After RFC2205, IETF for various need do ten to RSVP surplus kind expansion, the analysis and summary of in draft-ietf-nsis-signalling-analysis-03.txt, these expansions being done according to NSIS working group, the RSVP expansion comprises: RSVP passes through IP tunnel, support IPSec, the support policy control framework, reduce the soft-state refreshes expense, the RSVP polymerization, 802 local area network (LAN)s are passed through in support, the ATM net is passed through in support, DiffServ (differentiated service) is passed through in support, support empty COS, support MPLS TE (multi-protocol label switching traffic engineering), support GMPLS (exchange of general multiple protocols label), support the RSVP agency, localization RSVP promptly only reserves at Access Network, support mobile IP etc.
In these expansions, it is successful having only the RSVP-TE agreement, and set up explicit route LSP (label switched path) in the MPLS network to be supported in by a large amount of operator's widespread deployments, and other expansion and RSVP and intserv model are not disposed widely itself.
RSVP-TE is used for the MPLS network, for metastable polymerization traffic is set up explicit route LSP, and bandwidth reserved.Because the life cycle of LSP is long, flow and path can not change continually, the quantity of LSP in network becomes the quantity of how much grades of ground less than the IP miniflow, therefore, the refresh message expense of RSVP-TE and initiate request and there is not scaling concern in request merging mechanism for the receiving terminal of supporting the multicast design.
Yet, the design original intention of RSVP is used to tiny IP application data stream reserved resource and particular procedure in network, and the resource reservation of support IP multicast data flow, so, RSVP at data flow be the miniflow of application layer, life cycle is short, change frequent, if be each miniflow reserved resource on intermediate router, carry out particular procedure, then RSVP refresh message expense and receiving terminal are initiated the mechanism of request and request merging, problem aspect complexity and autgmentability is just very outstanding, and, be difficult to the support of realization to mobile IP, therefore, IntServ or RSVP only are applicable in the mininet at present.
In big-and-middle-sized network, the DiffServ that does not rely on signaling, coarseness that extensibility is good is in occupation of dominant position, still, DiffServ can only provide relative QoS, can not guarantee predictable, QoS end to end, therefore, DiffServ still can not satisfy the requirement of real time business fully.
From top description as can be seen, though the IPv6 agreement is by adopting fixed header, 8 bit traffic grade such as TC, Traffic Class territory etc., the Flow Label of 20 bits, methods such as variable length extension header make mobile IP and IPsec problem be simplified largely and solve, make the IPv6 network at outstanding moving in the IPv4 network, problem such as safety and QoS is improved, but aspect QoS, only stipulated that in RFC3697 3 tuples are used for traffic classification and to 2 basic demands of stream mode method for building up, do not stipulate concrete stream mode method for building up and transaction module.In addition, the end-to-end measurable QoS of the scalability problem of IntServ/RSVP, DiffServ guarantees also keeping intact in conjunction with problem of problem, IntServ and DiffServ in present IPv6, do not solved at all.Facts have proved for many years, complicated RSVP is successful in the RSVP-TE expansion, but fail aspect setting up in the application layer data flow state, therefore, there is not a gratifying stream mode to set up scheme at present as yet, can satisfy the requirement of real time business, solve the real time business problem of transmission as the QoS demand of application layer real time data stream.
At present, the existing stream mode method for building up of IPv6 as shown in Figure 1.
Among Fig. 1, transmitting terminal Src sends the message that includes the data flow resource request by the forwarding one by one of intermediate node R1, R2, R3 to receiving terminal Dst; On the path of this message process, start stream mode and set up the intermediate node of function, judge whether this node satisfies the demand condition of resource request in the current message, if satisfy, then carry out resource reservation according to this resource request, and transmit the message include the data flow resource request to next node, transmit the message that includes the data flow resource request to R2 as R1; Otherwise, respond to transmitting terminal feedback resources request refusal, and change this resource request into the resource acquisition request, transmit the message that this comprises the resource acquisition request to next node, send the response of resource request refusal as R2 to transmitting terminal Src, and send the resource acquisition request to R3.If include the data flow resource request in the message that receiving terminal receives, then carry out resource reservation, and confirm response to the request of transmitting terminal feedback resources according to this resource request; If include the resource acquisition request in the message that receiving terminal receives, then collect response to the transmitting terminal feedback resources, when receiving the resource acquisition request that the R3 transmission comes, collect response as receiving terminal Dst to transmitting terminal Src feedback resources.Transmitting terminal is by periodically sending the reservation that the message that includes the data flow resource request keeps resource.
In the network of reality, owing to reasons such as load balancings, the path that intermediate node arrives receiving terminal may all can have many, be to have a reachable path between transmitting terminal and the receiving terminal at least, in accompanying drawing 2, three reachable paths are arranged between transmitting terminal Src and the receiving terminal Dst, be respectively: Src-R1-R2-R3-Dst, Src-R1-R4-R3-Dst and Src-R1-R2-R5-Dst.According to the existing stream mode method for building up of IPv6, though transmitting terminal has been finished the resource reservation path by resource request, but, actual data traffic may be not path by this reservation send, but be sent to receiving terminal by other path, in accompanying drawing 2, transmitting terminal Src has finished Src-by resource request〉R1-〉R2-〉R3-〉resource reservation in Dst path, but in fact the data that send of transmitting terminal Src may be passed through Src-〉R1-〉R4-〉R3-〉Dst sends to receiving terminal Dst.Thereby the bandwidth that causes reserving does not have the data utilization, has wasted bandwidth, and the data flow that really needs bandwidth to guarantee does not obtain bandwidth and guarantees.
Summary of the invention
The objective of the invention is to, a kind of realization flow transmission method end to end is provided, improved resource utilization, made data flow can access the bandwidth assurance.
For achieving the above object, a kind of realization provided by the invention is the flow transmission method end to end, comprising:
A, stream transmitting terminal and the mulitpath between the receiving terminal in be described stream reserved resource;
B, described stream transfer to receiving terminal according to the reserved resource in each path from transmitting terminal.
Described step a specifically comprises:
A1, transmitting terminal send the resource request message that carries the corresponding reserved resource parameter of stream to receiving terminal;
A2, receive resource request message intermediate node when the resource of this node satisfies reserved resource parameter in the described resource request message, carry out resource reservation according to described reserved resource parameter;
A3, described intermediate node determine the reserved resource parameter of each load balancing item according to the reserved resource allocation strategy, and transfer to downstream node by resource request message.
Described step a2 specifically comprises:
A21, when the intermediate node that receives resource request message is not the resource that arrives last intermediate node of receiving terminal and this node when satisfying the reserved resource parameter of its reception, carry out resource reservation according to the reserved resource parameter of its reception;
A22, when the intermediate node that receives resource request message be that the resource that arrives last intermediate node of receiving terminal and this node is when satisfying the reserved resource parameter of its reception, reserved resource parameter according to its reception is carried out resource reservation, and sends the resource request response message of resource request success to its each upstream node;
A23, receive the intermediate node of the resource request response message of resource request success, continue to send described resource request response message, transfer to transmitting terminal up to the resource request response message to its upstream node.
Described step a23 also comprises:
When intermediate node receives the resource request response message of resource request success, begin, and when in life span, receiving transmitting terminal once more, again life span is picked up counting at the resource request message of described stream to the life span timing;
Intermediate node keeps described reserved resource in life span.
Described resource request message, resource request response message send with the form that includes the prolate-headed IPv6 message of resource reservation.
When intermediate node had a plurality of upstream node, the step that the resource of this node satisfies the reserved resource parameter of its reception among described step a21, the a22 comprised:
The resource of this node satisfies each reserved resource parameter of its reception simultaneously;
And described step a23 comprises:
When having the intermediate node of a plurality of upstream nodes to receive the resource request response message of resource request success, continue to send described resource request response message, transfer to transmitting terminal up to the resource request response message of resource request success to its each upstream node.
Carry the maximum reserved resource scope that this node can provide for its upstream node in the resource request response message that described intermediate node sends.
Described method also comprises after step a22:
Intermediate node with load balancing item is adjusted the reserved resource allocation strategy according to the maximum reserved resource scope in the resource request response message of each downstream node of its reception.
Reserved resource allocation strategy among the described step a3 is: each load balancing item has the reserved resource allocation strategy of equal value of identical weights or the non-equivalence reserved resource allocation strategy that each load balancing item has different weights.
Described step a also comprises:
A4, when the intermediate node that receives resource request message is not the resource that arrives last intermediate node of receiving terminal and this node can not satisfy the reserved resource parameter of its reception the time, send resource inquiry information respectively to its each downstream node;
A5, when the intermediate node that receives resource request message is the resource that arrives last intermediate node of receiving terminal and this node can not satisfy the reserved resource parameter of its reception the time, send the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node respectively to its each upstream node;
A6, receive the intermediate node of resource inquiry information, send the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node to its upstream node respectively;
A7, when the intermediate node that can not satisfy the reserved resource parameter receives the resource query response message, to the resource request message of its upstream node transmission resource request failure, until transferring to transmitting terminal.
Also comprise after the described step a5:
Intermediate node with load balancing item is adjusted the reserved resource allocation strategy according to the maximum reserved resource scope in the resource query response message of each downstream node of its reception.
When carrying the maximum reserved resource scope that this node can provide for its upstream node in the resource request response message that intermediate node sends, described step a1 comprises:
Transmitting terminal is adjusted the corresponding reserved resource parameter of described stream according to the maximum reserved resource scope in the resource request response message of its reception, and resends the resource request message that carries adjusted reserved resource parameter to receiving terminal.
Described method also comprises:
Transmitting terminal sends resource inquiry information to receiving terminal;
Each intermediate node returns the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node step by step to transmitting terminal.
Described step a1 comprises:
When the maximum reserved resource scope of transmitting terminal in the resource query response message of determining its reception can satisfy the resource reservation request of described stream, send resource request message to receiving terminal again.
Described resource inquiry information, resource query response message send with the form that includes the prolate-headed IPv6 message of resource reservation.
Described step c comprises:
Each intermediate node transmits described stream according to hash algorithm or weighting algorithm by reserved resource.
Description by technique scheme as can be known, by carrying out resource reservation for stream in each path between transmitting terminal and receiving terminal, the path of reserved resource and the inconsistent phenomenon in path of the actual forwarding of data flow have been avoided, make the reserved resource among the IPv6 can obtain practical application, and, farthest utilized the load balancing resource, Internet resources are fully used; The present invention provides a kind of perfect stream mode end to end to create a mechanism by utilizing resource request message, resource request response message, resource inquiry information, resource query response message; By above-mentioned each message is sent with the form that includes the prolate-headed IPv6 message of resource reservation, avoided the use of other specialized protocol messages, reduced message overhead; By in the resource request response message, carry this node in the resource query response message and can reserve scope for the maximum resource that its upstream node provides, make each intermediate node can know resource situation in each load balancing item, make transmitting terminal can know the resource situation in the path between itself and the receiving terminal, like this, intermediate node can carry out rational reserved resource and distribute, transmitting terminal can suitably be adjusted the corresponding reserved resource parameter of stream, or the resource situation in the path satisfies when flowing corresponding reserved resource parameter, rebulid stream mode, stream mode can successfully be set up; Thereby realized improving network resource utilization, improving the purpose of stream mode practicality end to end by technical scheme provided by the invention.
Description of drawings
Fig. 1 is a realization of the prior art flow transmission method schematic diagram end to end;
Fig. 2 is the problem schematic diagram of realization of the prior art flow transmission method existence end to end;
Fig. 3 is a realization of the present invention flow transmission method schematic diagram one end to end;
Fig. 4 is a realization of the present invention flow transmission method schematic diagram two end to end;
Fig. 5 is a realization of the present invention flow transmission method schematic diagram three end to end;
Fig. 6 is a realization of the present invention flow transmission method schematic diagram four end to end;
Fig. 7 is a realization of the present invention flow transmission method schematic diagram five end to end;
Fig. 8 is a realization of the present invention flow transmission method schematic diagram six end to end.
Embodiment
Core of the present invention is: be described stream reserved resource in transmitting terminal that flows and the mulitpath between the receiving terminal, described stream transfers to receiving terminal according to the reserved resource in each path from transmitting terminal.
Based on core concept of the present invention technical scheme provided by the invention is further described below.
The present invention can be by the transmission of resource request message at transmitting terminal and each intermediate node between the receiving terminal of IPv6, be that each intermediate node duplicates forwarding with resource request message on many load sharing links, make the resource request message of transmitting terminal can travel through all load sharing links of intermediate node.
When intermediate node of the present invention can satisfy the reserved resource parameter of stream correspondence in the resource of this node, the duplicate resource request message.Intermediate node is when the duplicate resource request message, should make amendment to the reserved resource parameter in the resource request message of its reception according to the quantity of load balancing item, during the resource request message of the application 1M bandwidth that receives as intermediate node, if this intermediate node has 2 load balancing items, then this intermediate node should equal 1M with the summation that guarantees 2 load balancing item bandwidth reserveds according to predetermined reserved resource allocation strategy for each load balancing distributes bandwidth reserved.By that analogy, all intermediate nodes between transmitting terminal and the receiving terminal all carry out above-mentioned processing procedure.
Intermediate node of the present invention is when the reserved resource parameter that can not satisfy in the resource request message, no longer simply resource request is converted to the resource acquisition request, but determine the maximum reserved resource scope that this node can provide for its upstream node, then, with this maximum reserved resource scope announcement upstream node, upstream node is by preserving the resource situation of downstream node, make it can be when receiving the reserved resource request next time, be each load balancing item reasonable distribution reserved resource, be that intermediate node is on the signaling aspect, can dynamic assignment reserved resource parameter, on the data forwarding aspect, can give different load balancing items by the dynamic assignment flow.
The reserved resource of each intermediate node among the present invention all is provided with lifetime (life span), when the lifetime of certain reserved resource finishes, reduce to 0 as lifetime, this resource is just no longer reserved, transmitting terminal needs its reserved resource to receiving terminal of periodic request, to guarantee that in stream transmission procedure the lifetime of each intermediate node reserved resource does not finish.
For making things convenient for the transmission of resource obligate information, redetermination of the present invention following 4 kinds of message:
1, resource request message mainly comprises: message identifier, traffic identifier, reserved resource parameter and life span etc.It is resource request message that message identifier is mainly used in this message of sign; Traffic identifier is mainly used in the sign transmitting terminal needs the stream of reserved resource to receiving terminal unique identification; The reserved resource parameter is mainly used in parameters such as CIR (committed rate), PIR (peak rate), delay requirement, the shake requirement of expression convection current, minimum message length, maximum message segment length; Life span is mainly used in the life span of expression resource reservation.
2, resource request response message mainly comprises: message identifier, traffic identifier, response message, maximum resource reservation scope etc.It is the resource request response message that message identifier is mainly used in this message of expression; Traffic identifier in the resource request response message should be identical with the traffic identifier in the resource request message; Response message is mainly used in represents whether resource request is successful; The maximum resource that maximum resource reservation scope is mainly used in this node of expression can be provided is reserved scope.
3, resource inquiry information mainly comprises: message identifier, inquiry sign etc.It is resource inquiry information that message identifier is mainly used in this message of sign; The inquiry sign is mainly used in the unique identification of sign transmitting terminal to the resource inquiry information of receiving terminal, i.e. the inquiry of each resource inquiry information sign is unique.
4, resource query response message mainly comprises: message identifier, inquiry sign, maximum resource reservation scope etc.It is resource inquiry information that message identifier is mainly used in this message of sign; Inquiry in resource query response message sign should with resource inquiry information in identical.The maximum resource that maximum resource reservation scope is mainly used in this node of expression can be provided is reserved scope.
The present invention can be a kind of IPv6 extension header of above-mentioned each message redetermination, be the resource reservation extension header, above-mentioned resource request message, resource request response message, resource inquiry information, the resource query response message of real-time application data stream is carried on transmission among the prolate-headed Option of resource reservation (option).The present invention also can be in present defined Hop-by-hop Options Header (hop-by-hop options header) several Option of redetermination, above-mentioned each message of real-time application data stream is placed on the middle transmission of Option (option) of redetermination, when carrying above-mentioned message of the present invention in the present defined hop-by-hop options header, this hop-by-hop options header can be called the resource reservation extension header.
Transmitting terminal is by in the resource reservation extension header that resource request is carried at the IPv6 message, and needn't use the message transmission of specialized protocol, thereby reduced message overhead.
Below in conjunction with accompanying drawing 3 to 8 pairs of realizations of the present invention of accompanying drawing end to end the flow transmission method describe.
When Fig. 3 all can satisfy transmitting terminal to the resource reservation request of receiving terminal for each intermediate node, realization of the present invention is the flow transmission method flow diagram end to end.
Among Fig. 3, at first transmitting terminal Src sends resource request message to receiving terminal Dst, for its stream that need be sent to receiving terminal Dst carries out resource reservation.Resource request message can send by the form that includes the prolate-headed IPv6 message of resource reservation.Include message identifier, traffic identifier, reserved resource parameter such as CIR, PIR etc. and life span in the resource request message that transmitting terminal Src sends.Be that example is set up process to stream mode end to end of the present invention and described only below with CIR in the reserved resource parameter and PIR.
CIR in the resource request message that setting transmitting terminal Src sends is 1M, and PIR is 2M.
Intermediate node R1 receives the IPv6 message, and judges in this IPv6 message whether include the resource reservation extension header, if do not comprise, intermediate node R1 transmits this IPv6 message according to normal forwarding process; If include the resource reservation extension header, and when determining to include resource request message in this IPv6 message according to the message identifier in the resource reservation extension header, intermediate node R1 obtains traffic identifier, life span and reserved resource parameter from the resource reservation extension header, as CIR, PIR, delay requirement, shake requirement etc.
Intermediate node R1 judges according to traffic identifier whether the resource reservation of this stream exists, be in the resource request message traffic identifier whether with other resource reservation request of this node record in traffic identifier equate, if equate, the resource request message of then representing this stream has passed through load balancing and in the polymerization of this node before arriving this node; If unequal, represent that then the resource request message of this stream arrives this node for the first time.
Owing to do not have the load balancing item between transmitting terminal Src and the intermediate node R1, so, do not record in the intermediate node with this resource request message in identical traffic identifier, it is new stream request resource reserved resource request message that this resource request message is one.At this moment, it is exactly that this receives reserved resource parameter in the resource request message that intermediate node R1 flows reserved resource parameter to be processed at this, and promptly CIR is 1M, and PIR is 2M.
Intermediate node R1 judges that this node is to whether there being many load balancing paths between the receiving terminal Dst, because intermediate node R1 can arrive receiving terminal Dst by R2 and R4, so intermediate node R1 needs to send resource request message to the downstream node R2 and the R4 of load balancing item respectively.The summation of the resource reservation value of each load balancing item correspondence should equal the resource reservation value that upstream node requires, promptly intermediate node R1 respectively to downstream node the CIR sum in the resource request message that sends of R2 and R4 should be 1M, the PIR sum should be 2M.
Intermediate node R1 to downstream node CIR, the PIR in the resource request message that sends of R2 can with intermediate node R1 to downstream node CIR, the PIR in the resource request message that sends of R4 equate, the CIR, the PIR that are about to the upstream node requirement are respectively divided by load balancing item 2, at this moment, intermediate node R1 is divided into 0.5M with CIR, and PIR is divided into 1M.At this moment, the resource request message that sends respectively to the path of two load balancings of intermediate node R1 is a kind of resource request message of equivalence.
Intermediate node R1 CIR, the PIR in the resource request message that sends of R2 to downstream node also can CIR, the PIR in the resource request message that sends of R4 be unequal to downstream node with intermediate node R1, promptly intermediate node R1 according to predetermined policy as wait according to the weight that disposes on the load balancing item to determine in its resource request message that is sent to downstream node R2 CIR, PIR with and be sent to CIR, PIR in the resource request message of downstream node R4.At this moment, the resource request message that sends respectively to the path of two load balancings of intermediate node R1 is a kind of non-equivalent resource request message.
Intermediate node R1 judges whether the resource summation of two load balancing items can satisfy upstream node requested resource reserved value, the resource summation that is set in two load balancing items in this flow process can satisfy upstream node requested resource reserved value, at this moment, intermediate node R1 is to the path down distributing resource request message of each load balancing, and the resource request message here is still for including the prolate-headed IPv6 message of resource reservation.Set intermediate node R1 to the CIR in the path down distributing resource request message of each load balancing be 0.5M, PIR is 1M.
Because intermediate node R1 is not last intermediate node that arrives receiving terminal Dst,, but wait for the resource request response message of its downstream node so intermediate node R1 does not send the resource request response message to its upstream node.
Intermediate node R2 receives that upstream node R1 sends includes the prolate-headed IPv6 message of resource reservation, intermediate node R2 is identical with above-mentioned description to intermediate node R1 to the processing procedure of this IPv6 message, to be intermediate node R2 to the path of two load balancings issue respectively includes the prolate-headed IPv6 message of resource reservation, promptly respectively to its downstream node R3, R5 issues and includes the prolate-headed IPv6 message of resource reservation, and the cir value in two prolate-headed resource request messages of resource reservation can be identical, also can be different, the value of PIR can be identical, also can be different.CIR among Fig. 3 in two resource request messages is 0.25M, PIR is 0.5M.Equally, intermediate node R2 also needs to wait for the resource request response message of its downstream node.
Intermediate node R4 receives that upstream node R1 sends includes the prolate-headed IPv6 message of resource reservation, intermediate node R4 is basic identical to the processing procedure of this IPv6 message and above-mentioned description to intermediate node R1, just because R4 can only arrive receiving terminal Dst by its downstream node R3, there is not the load balancing item, so, intermediate node R4 only will include the prolate-headed IPv6 message of resource reservation and be sent to downstream node R3, and the CIR in the prolate-headed resource request message of resource reservation is that 0.5M, PIR are 1M.Equally, intermediate node R4 also needs to wait for the resource request response message of its downstream node.
Intermediate node R3 can receive two and include the prolate-headed IPv6 message of resource reservation, because the traffic identifier in the prolate-headed resource request message of resource reservation of these two IPv6 messages is identical, so, intermediate node R3 determines that the resource request message of this stream is before arriving this node, passed through load balancing and in the polymerization of this node, at this moment, intermediate node R3 flows reserved resource parameter to be processed at this and is the reserved resource parameter in two resource request messages of its reception.
Intermediate node R3 merges the CIR in two resource request messages of its reception, PIR is merged, and among Fig. 3, the numerical value after CIR merges is 0.75M, and the numerical value after PIR merges is 1.5M.Then, intermediate node R3 determines its resource reservation value after whether can satisfying merging.Be set in the flow process of Fig. 3, intermediate node R3 can satisfy the upstream node requested resource reserved value after the merging.
Because intermediate node R3 can satisfy resource reservation value after the merging for last intermediate node of arriving receiving terminal Dst and intermediate node R3, so, intermediate node R3 sends the resource request response message of resource request success respectively to two upstream node R4 and R5, and the resource request response message can send by the form that includes the prolate-headed IPv6 message of resource reservation.Not only include message identifier, the traffic identifier of resource request response message, the information of resource request success in the resource request response message here, also include intermediate node R3 and can reserve scope for the maximum resource that its upstream node provides, in Fig. 3, CIR in the resource request response message that intermediate node R3 upstream nodes R4 sends is 2M, PIR is 4M, CIR in the resource request response message that intermediate node R3 upstream nodes R2 sends is 10M, and PIR is 20M.Intermediate node R3 is when upstream nodes R2, R4 send the resource request message of resource request success, begin to be the life span timing in the resource request message of its reception, when clocking value reaches life span, and when intermediate node R3 does not receive the resource request message that transmitting terminal Src sends at this stream once more, be released to this stream reserved resource.
Intermediate node R5 only receives one and includes the prolate-headed IPv6 message of resource reservation, so intermediate node R5 directly determines according to the reserved resource parameter in the prolate-headed resource request message of this resource reservation whether it can satisfy upstream node R2 requested resource reserved value.In the flow process of Fig. 3, set intermediate node R5 and can satisfy upstream node R2 requested resource reserved value.
Because intermediate node R5 can satisfy upstream node R2 requested resource reserved value for last intermediate node and the intermediate node R5 that arrives receiving terminal Dst, so intermediate node R5 sends the resource request response message of resource request success to its upstream node R2.Equally, the resource request response message can send by the form that includes the prolate-headed IPv6 message of resource reservation, and the maximum resource that includes intermediate node R5 in the resource request response message here equally and can provide for its upstream node is reserved scope, in Fig. 3, CIR in the resource request response message that intermediate node R5 upstream nodes R2 sends is 10M, and PIR is 20M.In like manner, intermediate node R5 also begins the life span timing for its reception.
What intermediate node R2 received its downstream node R3 and R5 includes the prolate-headed IPv6 message of resource reservation, when determining that according to the message identifier in two resource reservation extension header message in these two resource reservation extension header is the resource request response message, intermediate node R2 needs and can feed back to its upstream node R1 for the maximum resource reservation scope that its upstream node provides, can be as intermediate node R2 with after the addition of two maximum resources reservation scopes, send the resource request response message of resource request success to its upstream node R1, maximum resource reservation scope in this resource request response message is the result after the above-mentioned addition, in Fig. 3, CIR in the resource request response message that intermediate node R2 upstream nodes R1 sends is 20M, and PIR is 40M.
What intermediate node R4 received its downstream node R3 includes the prolate-headed IPv6 message of resource reservation, when determining that according to the message identifier in the resource reservation extension header message in the resource reservation extension header is the resource request response message, send the resource request response message of resource request success to its upstream node R1, maximum resource reservation scope in this resource request response message is that the maximum resource in the resource request response message of its reception is reserved value range, in Fig. 3, CIR in the resource request response message that intermediate node R4 upstream nodes R1 sends is 0.5M, and PIR is 1M.
When the message of intermediate node R1 in the resource reservation extension header that the downstream node R2 and the R4 of definite its reception transmit the IPv6 message that comes respectively was the resource request response message of resource request success, intermediate node R1 was to the processing procedure of the resource request response message of downstream node R2 and R4 such as the description of above-mentioned centering intermediate node R2.In Fig. 3, the CIR of intermediate node R1 in the resource request response message of the resource request success that transmitting terminal Src sends is 22M, and PIR is 44M.
When transmitting terminal Src received the resource request response message of resource request success of downstream node R1, the stream mode between transmitting terminal Src and the receiving terminal Dst was successfully set up.At this moment, transmitting terminal Src can be according to above-mentioned reserved resource with the flow transmission of above-mentioned traffic identifier correspondence to receiving terminal Dst.Like this, intermediate node can guarantee that stream is by reserved resource transmission, the phenomenon of having avoided reserved resource not to be fully utilized when the transmission of the overload of will circulating is shared a path.
If each intermediate node between transmitting terminal Src and the receiving terminal Dst does not receive the resource request message of transmitting terminal Src once more in its life span separately, then delete it and be this stream reserved resource.
From to the description of Fig. 3 as can be seen, can reserve scope for the maximum resource that its upstream node provides by in the resource request response message, carrying intermediate node, make each intermediate node that receives the resource request response message, the resource situation that transmitting terminal all can be known its downstream node, thereby make intermediate node in the resource reservation process of carrying out next time, can adjust it according to the real resource situation of its downstream node and be sent to reserved resource parameter in the resource request message that different loads shares.
Part intermediate node between transmitting terminal and receiving terminal can not satisfy the reserved resource request of its upstream node, but the reserved resource request that can satisfy upstream node by redistributing of load balancing, thereby realize that method flow that stream mode end to end sets up is shown in accompanying drawing 4, accompanying drawing 5.
Among Fig. 4, transmitting terminal Src, intermediate node R1, R2, R5 sends resource request message, send among the process of resource request response message and above-mentioned Fig. 3 transmitting terminal Src, intermediate node R1, R2, the description of R5 is basic identical, its difference only is that intermediate node R1 obtains the maximum resource that downstream node R4 can provide according to the resource query response message of its downstream node R4 to reserve scope, and, because intermediate node R4 can not satisfy the resource reservation request of its upstream node R1, so intermediate node R1 should send the resource request response message of resource request failure to transmitting terminal Src.
Because the intermediate node R4 among Fig. 4 can not satisfy the resource reservation request of upstream node R1, so intermediate node R4 need be revised as resource inquiry information with resource request message, then, resource inquiry information is sent to downstream node R3.If intermediate node R4 has the load balancing item, then intermediate node R4 need send resource inquiry information to the downstream node of each load balancing item.
Resource inquiry information can send by the form that includes the prolate-headed IPv6 message of resource reservation.Intermediate node R4 waits for the resource query response message of downstream node feedback after sending resource inquiry information.
What intermediate node R3 received upstream node R4 includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in the resource reservation extension header message in the resource reservation extension header is resource inquiry information, this node can be reserved scope for the maximum resource that upstream node R4 provides transfer to upstream node R4 by the information query response message.
The resource query response message can send by the form that includes the prolate-headed IPv6 message of resource reservation.The maximum resource that mainly comprising message identifier, inquiry sign, traffic identifier and this node in the resource query response message can provide for the upstream node that sends resource inquiry information is reserved scope.Among Fig. 4, the CIR of intermediate node R3 in the resource query response message that its upstream node R4 sends is that 2M, PIR are 3M.
What intermediate node R4 received downstream node R3 includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in the resource reservation extension header message in the resource reservation extension header is the resource query response message, intermediate node R4 can be carried on for the maximum optimum reserved value that upstream node R1 provides in the resource request response message, and with the resource request response message by including the prolate-headed IPv6 message transmissions of resource reservation to its upstream node R1, simultaneously, intermediate node R4 also needs to write down the maximum resource reserved value that downstream node R 3 can provide.If intermediate node R4 is to a plurality of load balancing items are arranged between the receiving terminal Dst, then intermediate node R4 need gather the maximum resource reservation scope in the resource query response message of the downstream node of each load balancing item feedback, then, according to gathering the back result this node can be transferred to upstream node R1 by the resource request response message for the maximum resource reservation scope that upstream node R1 provides, simultaneously, intermediate node R4 also needs to write down the maximum resource reserved value that each load balancing item can provide.The maximum resource reserved value that intermediate node R4 can provide according to each load balancing item of its record is provided with the weight of each load balancing item.
If the resource reservation request that last intermediate node before the arrival receiving terminal Dst such as intermediate node R3, R5 among Fig. 4 can not satisfy upstream node, then do not need to send to downstream node resource inquiry information, directly this node can be reserved scope for the maximum resource that its upstream node provides and transfer to its upstream node by the resource query response message.
What intermediate node R1 received downstream node R4 includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in the resource reservation extension header message in the resource reservation extension header is the resource query response message, record downstream node R 4 can be reserved scope for its maximum resource that provides, the downstream node R2 that gets access to the resource request response message of the resource request success that intermediate node R1 can also feed back from the downstream node R2 of its reception in the load balancing item can be that its maximum resource that provides is reserved scope, intermediate node R1 gathers the maximum resource reservation scope of downstream node R2 and R4, and the resource request response message that the reservation of the maximum resource after will gathering scope is failed by resource request transfers to transmitting terminal Src, in Fig. 4, CIR in the resource request response message of intermediate node R1 resource request failure is set to 1.6M, PIR is set to 3.2M, and is sent to transmitting terminal Src.
Intermediate node R1 can share the resource situation of item according to the different loads of its record, weight is set for each load balancing item, like this, when resource reservation request message arrival next time, intermediate node R1 can be according to the resource situation of many load balancing items of its acquisition, be the weight of each load balancing item, adopt the non-equivalence method for obligating resource to carry out resource reservation and distribute.
When transmitting terminal Src received the resource request response message of resource request failure of downstream node R1, transmitting terminal Src determined that it is set up with the stream mode between the receiving terminal Dst and fails.At this moment, transmitting terminal Src can according to the reserved resource parameters such as CIR, PIR in the resource request response message of its reception determine needs once more to receiving terminal Dst to sending resource request message.
When transmitting terminal Src once more when receiving terminal Dst sends resource request message, the method that the present invention sets up the stream mode end to end between transmitting terminal Src and the receiving terminal Dst is as shown in Figure 5.
Among Fig. 5, when intermediate node R1 receives the resource request message of transmitting terminal Src and once more when its downstream node R2, R4 send resource request message, the CIR of intermediate node R1 in can its resource request message that downstream node R2 is issued is set to 0.75M, PIR is set to 1.5M, and the CIR of intermediate node R1 in can its resource request message that downstream node R4 is issued is set to 0.25M, PIR is set to 0.5M.Thereby, the stream mode end to end between transmitting terminal Src and the receiving terminal Dst can successfully be set up.
Part intermediate node between transmitting terminal and receiving terminal can not satisfy the reserved resource request of upstream node, even during the reserved resource request that can not satisfy upstream node by redistributing of load balancing, the method flow that stream mode end to end of the present invention is set up is shown in accompanying drawing 6, accompanying drawing 7, accompanying drawing 8.
When transmitting terminal Src sends resource request message for the first time for the stream that it need be sent to receiving terminal Dst, as shown in Figure 6 with the process of setting up stream mode.
Among Fig. 6, transmitting terminal Src, intermediate node R1 send among resource request message, the process that sends the resource request response message and above-mentioned Fig. 4 basic identical to the description of transmitting terminal Src, intermediate node R1; The process that intermediate node R4 sends resource inquiry information, resource query response message also with above-mentioned Fig. 4 in basic identical to the description of intermediate node R4.
Because the intermediate node R2 among Fig. 6 can not satisfy the resource reservation request of upstream node R1, so intermediate node R2 is modified as resource inquiry information with resource request message, then, resource inquiry information is sent to downstream node R3 and R5.
What intermediate node R3 and R5 received upstream node R2 respectively includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in the resource reservation extension header this message is resource inquiry information, the maximum resource that all this node can be provided is reserved scope and is transferred to upstream node R2 by the information query response message.Among Fig. 6, the CIR of intermediate node R3 in the resource query response message that its upstream node R2 sends is that 10M, PIR are 20M; The CIR of intermediate node R5 in the resource query response message that its upstream node R2 sends is that 10M, PIR are 20M.
What intermediate node R2 received downstream node R3, R5 includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in the resource reservation extension header this message is the resource query response message, maximum resource reservation scope in two resource query response messages is gathered, as two CIR are carried out addition, two PIR are carried out addition, then, according to the result after gathering it can be reserved scope for the maximum resource that upstream node R1 provides and transfer to its upstream node R1 by the resource request response message.Simultaneously, intermediate node R2 also needs to write down the maximum resource reserved value that downstream node R 3, R5 can provide.
What intermediate node R1 received its downstream node R4, R2 respectively includes the prolate-headed IPv6 message of resource reservation, and when determining that according to the message identifier in two resource reservation extension header these two message are the resource query response message respectively, the maximum resource that record downstream node R 2, R4 can provide is reserved scope.Intermediate node R1 gathers the maximum resource reservation scope of R2 and R4, and according to the result after gathering its maximum resource that can provide for transmitting terminal Src is reserved the resource request response message of scope by the resource request failure and transfer to transmitting terminal Src, in Fig. 6, CIR in the resource request response message of intermediate node R1 resource request failure is set to 0.8M, PIR is set to 1.6M, and is sent to transmitting terminal Src.Equally, intermediate node R1 can share the resource situation of item according to the different loads of its record, weight is set for each load balancing item.
After transmitting terminal Src receives the resource request response message of resource request failure, whether this stream that judgement need be sent to receiving terminal Dst must satisfy resource reservation and just can be sent out, if must satisfying resource reservation, this stream just can be sent out, then transmitting terminal Src no longer sends the resource reservation request of this stream to receiving terminal Dst, but, periodically send resource inquiry information, resource situation on the detective path, in Fig. 7, when intermediate node R1 receives the resource inquiry information of transmitting terminal Src, continuation is to its downstream node R2, R4 sends resource inquiry information respectively, so analogize, when resource inquiry information transfers to last the intermediate node R3 that arrives receiving terminal Dst and R5 respectively, the maximum resource reserved value that R3 and R5 can provide it respectively feeds back to its upstream node by the resource query response message, the intermediate node that receives the resource query response message can feed back to upstream node by the resource query response message according to the maximum resource reserved value in the resource query response message for the maximum resource reserved value that upstream node provides with this node, so analogize, until transmitting terminal Src.Among Fig. 7, the CIR in the resource query response message of the intermediate node R1 that transmitting terminal Src receives is set to 0.8M, PIR is set to 1.6M.
When transmitting terminal Src determines that reserved resource parameter in the resource query response message of its reception can satisfy the resource reservation request of stream, send resource request message to receiving terminal Dst once more, for this stream is set up stream mode end to end.
Receive the resource request response message of resource request failure as transmitting terminal Src, and when definite this stream that need be sent to receiving terminal Dst can reduce the numerical value of reserved resource parameter, transmitting terminal Src adjusts the reserved resource parameter value in the resource request message, send new resource reservation request to receiving terminal Dst, in accompanying drawing 8, transmitting terminal Src is adjusted into 0.8M with the CIR in the resource request message, PIR is adjusted into 1.6M, then, send resource request message, intermediate node R1 and R2 can adjust the reserved resource parameter value that it is sent to downstream node according to the weight of its each load balancing item correspondence that writes down respectively, set up process according to the stream mode end to end among Fig. 8, the present invention can successfully set up stream mode end to end for above-mentioned stream.
Above-mentioned to Fig. 3 in the description of Fig. 8, the reserved resource parameter that each message is carried, be not limited to CIR and PIR, should comprise: committed rate, peak rate, delay requirement, the shake requirement, minimum message length, reserved resource parameters such as maximum message segment length, wherein, delay requirement, the shake requirement, minimum message length, reserved resource parameters such as maximum message segment length are at the resource request message of each load balancing item of intermediate node, the resource request response message, all be the same in the resource query response message, need not carry out five equilibrium according to the quantity of load balancing item, five equilibrium not, processing such as gather, requiring time delay as transmitting terminal is 1 millisecond, and then the time delay that is required of each load balancing all should be 1 millisecond.
Behind the stream reserved resource of successfully having set up by above-mentioned 4 kinds of messages at stream, when transmitting terminal Src begins to send data flow, in order to guarantee that each intermediate node can the invention provides weighting algorithm and HASH algorithm according to the data of certain ratio forwarding flow on many load balancing paths.
At first introduce the method for carrying out data flow transmission according to the HASH algorithm below.
Because stream mode is set up at same stream end to end, the tlv triple of the IPv6 message of this stream is the same, and promptly source IP address, purpose IP address and Flow Lable are the same, so, can not use the input of tlv triple as HASH.
The result of HASH algorithm will guarantee certain length, as the shortest be the numerical value of 1-1000.For instance, the path that 3 load balancings are arranged as an intermediate node, detecting the flow rate ratio that the path of 3 load balancings can bear according to resource request message is 17:20:63, three ratios add up to 100, if the HASH arithmetic result is shorter, as 2bits, like this, can only be divided into 4 piece of data to stream, and can't carry out load balancing according to the requirement of 100 piece of data.Example hereto, if the result of HASH algorithm is the numerical value of 1-1000, the expression node can be divided into 1000 parts with levelling by the HASH algorithm, HASH result<=170 walk article one path, 170<HASH result<=370 walk the second path, 370<HASH result<=1000 walk the 3rd paths.
Based on above-mentioned consideration, the HASH algorithm is the clock information of collection network equipment directly, and as gathering the millisecond numerical value of clock information, per second is 1000 milliseconds, so numerical value is 0-999, can reach the requirement that complex load is shared.
The method of carrying out data flow transmission according to weighting algorithm is as follows:
The application of weighting algorithm in flow transmission is simpler than HASH algorithm, weighting algorithm needs intermediate node to generate a counter, the data message of a stream of the every processing of intermediate node, counter adds one, the counter maximum is the summation of the weight of each load balancing item, surpass this value counter clear 0, the numerical value of counter drops on the scope of which weighted value, which just transmit from load balancing item, for instance, as an intermediate node path of 3 load balancings is arranged, the flow rate ratio that the path of 3 load balancings that detect according to resource request message can be born is 17:20:63, and three ratio additions equal 100, the count range of counter is 0-99, like this, when the count value of 0<=counter<17, data flow is transmitted by first load balancing item, when the count value of 17<=counter<37, data flow is transmitted by second load balancing item, and when the count value of 37<=counter<99, data flow is transmitted by the 3rd load balancing item.
Though described the present invention by embodiment, those of ordinary skills know, the present invention has many distortion and variation and do not break away from spirit of the present invention, and the claim of application documents of the present invention comprises these distortion and variation.

Claims (15)

1, a kind of realization flow transmission method end to end is characterized in that, comprising:
A, stream transmitting terminal and the mulitpath between the receiving terminal in be described stream reserved resource;
B, described stream transfer to receiving terminal according to the reserved resource in each path from transmitting terminal;
Described step a comprises:
A1, transmitting terminal send the resource request message that carries the corresponding reserved resource parameter of stream to receiving terminal;
A2, receive resource request message intermediate node when the resource of this node satisfies reserved resource parameter in the described resource request message, carry out resource reservation according to described reserved resource parameter;
A3, described intermediate node determine the reserved resource parameter of each load balancing item according to the reserved resource allocation strategy, and transfer to downstream node by resource request message.
2, a kind of realization as claimed in claim 1 flow transmission method end to end is characterized in that described step a2 specifically comprises:
A21, when the intermediate node that receives resource request message is not the resource that arrives last intermediate node of receiving terminal and this node when satisfying the reserved resource parameter of its reception, carry out resource reservation according to the reserved resource parameter of its reception;
A22, when the intermediate node that receives resource request message be that the resource that arrives last intermediate node of receiving terminal and this node is when satisfying the reserved resource parameter of its reception, reserved resource parameter according to its reception is carried out resource reservation, and sends the resource request response message of resource request success to its each upstream node;
A23, receive the intermediate node of the resource request response message of resource request success, continue to send described resource request response message, transfer to transmitting terminal up to the resource request response message to its upstream node.
3, a kind of realization as claimed in claim 2 flow transmission method end to end is characterized in that described step a23 also comprises:
When intermediate node receives the resource request response message of resource request success, begin, and when in life span, receiving transmitting terminal once more, again life span is picked up counting at the resource request message of described stream to the life span timing;
Intermediate node keeps described reserved resource in life span.
4, a kind of realization as claimed in claim 2 flow transmission method end to end is characterized in that, described resource request message, resource request response message send with the form that includes the prolate-headed IPv6 message of resource reservation.
5, a kind of realization as claimed in claim 2 flow transmission method end to end is characterized in that:
When intermediate node had a plurality of upstream node, the step that the resource of this node satisfies the reserved resource parameter of its reception among described step a21, the a22 comprised:
The resource of this node satisfies each reserved resource parameter of its reception simultaneously;
And described step a23 comprises:
When having the intermediate node of a plurality of upstream nodes to receive the resource request response message of resource request success, continue to send described resource request response message, transfer to transmitting terminal up to the resource request response message of resource request success to its each upstream node.
6, a kind of realization as claimed in claim 2 flow transmission method end to end is characterized in that, carries the maximum reserved resource scope that this node can provide for its upstream node in the resource request response message that described intermediate node sends.
7, a kind of realization as claimed in claim 6 flow transmission method end to end is characterized in that described method also comprises after step a22:
Intermediate node with load balancing item is adjusted the reserved resource allocation strategy according to the maximum reserved resource scope in the resource request response message of each downstream node of its reception.
8, as claim 1,2,3,4,5,6 or 7 described a kind of realizations flow transmission method end to end, it is characterized in that the reserved resource allocation strategy among the described step a3 is: each load balancing item has the reserved resource allocation strategy of equal value of identical weights or the non-equivalence reserved resource allocation strategy that each load balancing item has different weights.
9, as claim 1,2,3,4,5,6 or 7 described a kind of realizations flow transmission method end to end, it is characterized in that described step a also comprises:
A4, when the intermediate node that receives resource request message is not the resource that arrives last intermediate node of receiving terminal and this node can not satisfy the reserved resource parameter of its reception the time, send resource inquiry information respectively to its each downstream node;
A5, when the intermediate node that receives resource request message is the resource that arrives last intermediate node of receiving terminal and this node can not satisfy the reserved resource parameter of its reception the time, send the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node respectively to its each upstream node;
A6, receive the intermediate node of resource inquiry information, send the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node to its upstream node respectively;
A7, when the intermediate node that can not satisfy the reserved resource parameter receives the resource query response message, to the resource request message of its upstream node transmission resource request failure, until transferring to transmitting terminal.
10, a kind of realization as claimed in claim 9 flow transmission method end to end is characterized in that, also comprises after the described step a5:
Intermediate node with load balancing item is adjusted the reserved resource allocation strategy according to the maximum reserved resource scope in the resource query response message of each downstream node of its reception.
11, a kind of realization as claimed in claim 9 flow transmission method end to end is characterized in that:
When carrying the maximum reserved resource scope that this node can provide for its upstream node in the resource request response message that intermediate node sends, described step a1 comprises:
Transmitting terminal is adjusted the corresponding reserved resource parameter of described stream according to the maximum reserved resource scope in the resource request response message of its reception, and resends the resource request message that carries adjusted reserved resource parameter to receiving terminal.
12, a kind of realization as claimed in claim 9 flow transmission method end to end is characterized in that described method also comprises:
Transmitting terminal sends resource inquiry information to receiving terminal;
Each intermediate node returns the resource query response message that carries the maximum reserved resource scope that this node can provide for its upstream node step by step to transmitting terminal.
13, a kind of realization as claimed in claim 12 flow transmission method end to end is characterized in that described step a1 comprises:
When the maximum reserved resource scope of transmitting terminal in the resource query response message of determining its reception can satisfy the resource reservation request of described stream, send resource request message to receiving terminal again.
14, a kind of realization as claimed in claim 9 flow transmission method end to end is characterized in that, described resource inquiry information, resource query response message send with the form that includes the prolate-headed IPv6 message of resource reservation.
15, a kind of realization as claimed in claim 1 flow transmission method end to end is characterized in that described step b comprises:
Each intermediate node transmits described stream according to hash algorithm or weighting algorithm by reserved resource.
CNB2005101090561A 2005-10-17 2005-10-17 A method to realize the end-to-end stream transmission Expired - Fee Related CN100486219C (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CNB2005101090561A CN100486219C (en) 2005-10-17 2005-10-17 A method to realize the end-to-end stream transmission
PCT/CN2006/002730 WO2007045163A1 (en) 2005-10-17 2006-10-17 Method for realizing end-to-end stream transmission and the node equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2005101090561A CN100486219C (en) 2005-10-17 2005-10-17 A method to realize the end-to-end stream transmission

Publications (2)

Publication Number Publication Date
CN1953415A CN1953415A (en) 2007-04-25
CN100486219C true CN100486219C (en) 2009-05-06

Family

ID=37962195

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2005101090561A Expired - Fee Related CN100486219C (en) 2005-10-17 2005-10-17 A method to realize the end-to-end stream transmission

Country Status (2)

Country Link
CN (1) CN100486219C (en)
WO (1) WO2007045163A1 (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5737723B2 (en) * 2009-02-20 2015-06-17 トムソン ライセンシングThomson Licensing Scheduling method, end node and core network switch in packet switched communication network
FR2958490B1 (en) * 2010-03-31 2012-03-09 Alcatel Lucent METHOD FOR CONTROLLING THE ESTABLISHMENT OF A CONNECTION IN A TRANSPORT NETWORK
CN103281252B (en) * 2013-05-14 2017-04-26 华为技术有限公司 Message flow control method and device based on multi-path transmission
CN113395735B (en) * 2020-03-11 2023-07-07 华为技术有限公司 Message transmission method, device and network equipment

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3016023B1 (en) * 1999-01-22 2000-03-06 株式会社超高速ネットワーク・コンピュータ技術研究所 How to reserve network resources
US6999436B2 (en) * 2001-04-16 2006-02-14 Nokia Corporation Method and apparatus for efficient routing of mobile node packets
CN100387023C (en) * 2004-04-26 2008-05-07 华为技术有限公司 Method of flow state establishment

Also Published As

Publication number Publication date
WO2007045163A1 (en) 2007-04-26
CN1953415A (en) 2007-04-25

Similar Documents

Publication Publication Date Title
US6058113A (en) Method for enhancing resource reservation communication
Gambiroza et al. Design, analysis, and implementation of DVSR: a fair high-performance protocol for packet rings
US7636781B2 (en) System and method for realizing the resource distribution in the communication network
CN101854697B (en) Multi-constraint quality-of-service control routing method and system for wireless mesh network
CN106533960A (en) Data center network routing method based on Fat-Tree structure
US7092359B2 (en) Method for distributing the data-traffic load on a communication network and a communication network for implementing this method
CN101841487A (en) Configuration method for aggregating link service flow and packet switching device
Fotohi et al. A Hybrid routing method for mobile ad-hoc networks
CN100387023C (en) Method of flow state establishment
CN100486219C (en) A method to realize the end-to-end stream transmission
Kato et al. A congestion control method for named data networking with hop-by-hop window-based approach
CN1567891B (en) A method for implementing data service transmission routing
Striegel et al. A scalable protocol for member join/leave in DiffServ multicast
CN1953409A (en) A networking method for semi-network configuration of network and its system
Psaras et al. Revisiting resource pooling: The case for in-network resource sharing
CN105072046A (en) Delay tolerant network congestion prevention method based on data concurrence and forwarding by token control node
Hui et al. Analysis and research on OLSR protocol for multi-channel assignment of wireless mesh network
CN100514912C (en) Method of realizing Pipe model based on distinction service
Hua et al. A DTN congestion mechanism based on Distributed Storage
Bharadwaj et al. A bandwidth and energy aware QoS routing protocol for enhanced performance in ad-hoc networks
Parameswari Opportunistic Routing Protocol for Resource Optimization in Vehicular Delay-Tolerant Networks (VDTN)
Rong et al. LSMR: A label switching multipath routing protocol for ad hoc networks
Vutukury et al. SMART: a scalable multipath architecture for intra-domain QoS provisioning
Vutukury et al. A scalable architecture for providing deterministic guarantees
Lowu et al. On efficient distribution of data in multicast networks: QoS in scalable networks

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
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: 20090506

Termination date: 20191017