CN100407707C - Network configuration for loading high-speed data business and trans mitting method - Google Patents
Network configuration for loading high-speed data business and trans mitting method Download PDFInfo
- Publication number
- CN100407707C CN100407707C CN2006100587124A CN200610058712A CN100407707C CN 100407707 C CN100407707 C CN 100407707C CN 2006100587124 A CN2006100587124 A CN 2006100587124A CN 200610058712 A CN200610058712 A CN 200610058712A CN 100407707 C CN100407707 C CN 100407707C
- Authority
- CN
- China
- Prior art keywords
- network
- node
- packet
- time
- information
- 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
Links
Images
Abstract
The present invention relates to network architecture used for loading high-speed data services and a transmitting method for the network architecture. The present invention has the main design that network architecture which is logically independent, is centrally controlled and is connection-oriented, namely an optical core network, is established for the transmission of plate services, such as multimedia, large files, etc., and general services are still loaded on a traditional optical burst switching network. The network architecture is an all optical overlay network and comprises a service classifier, the optical core network and the optical burst switching network, wherein the optical core network and the optical burst switching network have identical physical topology. After entering network fringe nodes, services are first judged by the service classifier and are respectively sent to the optical core network and the optical burst network to be transmitted according to plate services and non-plate services. In the optical core network, a space-time tag protocol with the functions of both resource assignment and a router is adopted to centrally schedule resources of the whole net, which realizes the statistical multiplexing of network resources. Thus, rich bandwidth resources are fully utilized.
Description
Technical field
The present invention relates to a kind of network system and transfer approach thereof that is used to carry high-speed data service, particularly a kind ofly can support high-speed data service efficiently, the network system and the transfer approach thereof that on backbone optical network, transmit as big text data, interaction multimedia and the Grid Application Grid data etc. of Streaming Media, HD digital TV (HDTV), computer center.
Background technology
In recent years, the Network feature is from transferring to the packet switching service based on data service based on the circuit-switched service of voice, next generation network will be the multibusiness network that data service and multimedia service are combined together, for the user provides high-speed data and multimedia service, comprise the clear degree TV of large file, large scale database and multi-path high-definition of video work station, computer center etc.; On the other hand, the scientist of U.S.'s Bell Laboratory is through finding accurately measuring in real time with the research back of network institute loaded service such as LAN local area network (LAN), WAN wide area network for a long time, all embodies statistical self-similarity nature based on the actual volume of the network of packet switching service on the discrete time yardstick.This characteristic shows, shows on the scale similar suddenly at any time based on the traffic carrying capacity of IP grouped data, has a great difference with the Poisson process business model of traditional description speech business network.Therefore, the network system of studying under the new network traffic model has crucial meaning for support many (stream) media business of future generation.
Target based on the of future generation key information network (NGN) of data service is to realize All-Optical Communication Network.Because dense wave division multipurpose (DWDM) broad application, the light packet switching (OPS, Optical Packet Switching) that makes based on DWDM's and have features such as exchange fine size, bandwidth availability ratio height, flexibility are good becomes the effective switch mode of realization all-optical network.But, owing to still do not have ripe light logic and light caching technology, at present data service is carried on the wavelength routing technology that solution on the DWDM mainly adopts photosphere in order to realize full light packet switching.Yet the configuration of the bandwidth resources of this optical wavelength route switching (OCS) lacks flexibility, and efficient is not high.For this reason, ITU-T and IETF have provided the ASON model that can realize the optical network resource dynamic assignment---ASON (ASON) and broad sense multi protocol label exchange (GMPLS).This ASON model is that intelligentized chain of command is introduced photosphere, thereby makes optical-fiber network finish the dynamic assignment of light path, the processes such as automatic discovery of resource under the control of signaling.The automatic discovery of resource can be for signaling, route and administrative institute's usefulness, and carry out the renewal of full mesh topology and state information on this basis.Though realized from transmission, controlled and managed the automation that routing procedure was coordinated and handled in three aspects, but the route and the resource allocation that are based on wavelength, wavestrip or optical fiber exist with circuit switching mode, the exchange granularity is too thick, especially when network transmission speed from 2.5Gbps, when 10Gbps rises to 40Gbps, do not find the business of utilizing optical-fiber network layer large pipeline, bulky grain degree resource as yet, cause the optical network resource utilance not high, and the resource of for the optical-fiber network user, calling, manage and use the bulky grain degree flexibly is difficult to.In view of this, acted on the self-similarity nature of IP grouping, the exchange granularity between optical link exchange and light packet switching, can realize the OBS (OBS of Internet resources statistic multiplexing, Optical Burst Switching), become a kind of having a great attraction and the light switching technology of application prospect of key all optical network of realizing.
As shown in Figure 1, 2, the notable feature of light burst-switched is that the control grouping is used the different channels transmission with the data burst bag.Generate the data burst bag that control is divided into groups (comprising routing iinformation) and is made of the mass data grouping at network edge node, and, offset time comes trigger data burst packets (being payload user data) by being set, control is grouped in core node and is data burst bag bandwidth resource in advance by offset time and signaling mechanism, therefore data burst wraps in the transmission that can realize full optical transparency in the optical burst switched network, and do not need opto-electronic conversion, this has just broken away from the restriction of optical device immature development, also solve a difficult problem that is difficult to make full use of bandwidth because of electronic bottleneck, promised to be the switching technology of full optical core network of future generation.Therefore, the distribution of research light burst exchange network bandwidth resources and dispatching method problem have great importance for the performance that improves network, the demand that satisfies high-speed data service.
Up to now, the more RSVP of research has bilateral agreements and unilateral arrangement in OBS.Bilateral reservation protocol is to connect by setting up end to end, and service quality is protected, but time delay is long, bandwidth availability ratio is low; Monolateral reservation protocol is to transmit along the selected link of source node by the control grouping, and bandwidth resource is in order to transmit the data burst bag, typical as " just measuring the time " agreement (JIT), " enough time " signaling protocol (JET) etc. in the transmission course.JIT and JET utilize " offset time " between control grouping and the data burst bag to finish RSVP, wherein the JET agreement proposes to take " just measuring the time " strategy, promptly " offset time (T) " is more than or equal to the processing delay δ (i) of all intermediate nodes between the control grouping is from the source node to the destination node
The time of the grouping of network node processing controls at present is approximately a few μ s, suppose that every kilometer transmission time of burst packets is 5 μ s, along with forwarding of control grouping, processing delay and the error between actual delay estimated by source node also are μ s magnitudes, and this error may cause the waste with link circuit resource of losing of data burst bag after accumulative total.Simultaneously, all there is a common problem in JIT and the JET agreement, promptly can't grasp the whole network resource dynamic information in real time, under the condition of the whole network resource operating position the unknown, carry out Route Selection voluntarily by each fringe node, will cause time window scheduling and spatial port to be distributed in both and independently carry out resource assignation under the situation.Under the condition of network at high flow capacity load, this mode of separate assignment at random will cause burst packets to distribute resource the competition conflict but to occur on another dimension on the one dimension therein, cause obstruction, under large file service transmission situation, block, losing of big data quantity will be caused, the paralysis in a certain zone of network might be caused.
The main body of high-speed data service will be many (stream) medium and large file service, and its principal character is a self-similarity nature, no longer follows Poisson distribution on Mathematical Modeling, but heavy-tailed distribution (Pareto).An important results of heavy-tailed distribution is: it not only makes packet have burst characteristic on each time scale in practical business, also will cause plate (CLUSTER) effect simultaneously.The plate effect is meant open-minded along with many (stream) media business such as large file of HDTV, video request program, video conference, computer center, and the big young pathbreaker of packet who enters the core network fringe node reaches up to a hundred million even the gigabit magnitude.Fig. 3 has shown under different network service self similarity characteristic parameter (the Hurst parameter in the heavy-tailed distribution), in a certain special time period, enter big packet and the shared ratio of small data packets in the self similarity business of network edge node, as can be seen, the business long greater than average packet accounts for more than 75% of total traffic, when Network self similarity characteristic parameter is 0.8, business with plate (CLUSTER) effect accounts for more than 80% of total traffic, if still adopt traditional blindness resource assignation mode at random, to cause the significantly reduction of network resource utilization, even the paralysis of network.
Summary of the invention
At above problem, the object of the present invention is to provide a kind of network system that is used to carry high-speed data service, it is mainly conceived: for the transmission of plate business such as many (stream) medium and big file set up one in logic independently, carry out central authorities' control, connection-oriented network system---optical core network (CLUSTERNET), and general service (NON-CLUSTER, the i.e. professional and low QoS service of small documents) then still is carried on traditional optical burst switched network (NON-CLUSTERNET).As shown in Figure 4, for light burst-switched overlapping network schematic diagram, in this network system, CLUSTERNET can have identical physical topology with NON-CLUSTERNET, but separate on logical construction and physical structure.Described network system is a kind of full light overlay network, and its structure comprises:
1. professional grader
Professional grader is the gateway of full light overlay network, is positioned at network edge node, connects CLUSTERNET network and NON-CLUSTERNET network respectively.Major function is according to type of service (as multimedia, big data service etc.) and demand (as service quality, time delay etc.) data that enter network edge node to be packaged, produce CLUSTER and NON-CLUSTER business (packet) according to group bag algorithm, respectively CLUSTER and NON-CLUSTER business are sent to corresponding C LUSTERNET network edge processor and NON-CLUSTERNET network edge processor.Wherein, be prior art about the group bag algorithm of large file service, another patent application " method for self similar cluster packet of large file service optical transmission " (application number 200510086709.9) that can the application reference people.
2. optical core network network
Optical core network network (CLUSTERNET) is to be the professional core net of the CLUSTER that the packet with plate effect provides fast, high-quality transmits.Its principal character is: 1) adopt the space-time tag protocol to the unified allocation of resources of the whole network resource; 2) logical topology has been introduced the powerful central management face that is used to allocate the whole network resource on data surface and the chain of command basis innovatively in original light burst exchange network, is used for unified allocation of resources route and signaling; 3) physical topology is made up of core (CLUSTERNET) network edge processor, core (CLUSTERNET) network central control node and core (CLUSTERNET) network core node.The essential characteristic of this core net is: central authorities control with towards being connected.
3. core network edge processor
As shown in Figure 5, the major function of CLUSTERNET network edge processor is: 1) receive the CLUSTER packet from professional grader; 2) according to CLUSTER business service grade (QOS), destination node produce that the routing information request, the CLUSTER packet bag that comprise source node and destination node are long, the grade of service is in the resource bid packets of information of interior CLUSTER packet all properties, described resource bid packets of information can also comprise other attributes of packets such as type of service; 3) the resource bid packets of information is sent to CLUSTERNET network central control node; 4) the transmission buffer unit that the CLUSTER packet is transferred in the network edge processor is stored, and waits for the resource feedback information bag that returns; 5) information of carrying according to the resource feedback information bag that returns produces the control grouping that comprises routing table accordingly and arrive the information such as time, the node resource time that need take and wavelength of each node in the routing table for the CLUSTER packet that successfully obtains route resource; 6) will control on the control channel of packet bearer in the CLUSTERNET network; 7), will send to CLUSTERNET network data channel with the corresponding CLUSTER packet of control grouping according to the offset time that calculates by central control node; 8) will successfully not obtain the CLUSTER data packet discarding of route resource.
4. core network central control node
The major function of core network central control node is: 1) receive the resource bid packets of information that CLUSTERNET network edge processor sends; 2) content of extraction resource bid packets of information adopts space-time tag protocol and space-time mark routing algorithm to be and optimal path of the corresponding CLUSTER allocation of packets of resource bid packets of information, generates corresponding space-time mark; 4) this CLUSTER packet is arrived the time of each node on the above-mentioned optimal path and the time span information of shared this node switching port and write the whole network resource state information database; 5) produce the space-time that includes with the corresponding optimal path of this CLUSTER packet and be marked at interior resource feedback information bag, send to corresponding C LUSTERNET network edge processor.
Wherein, the space-time tag protocol is made up of along time mark (temporal information) and free token (port information) that selected routing channel arrives each node the CLUSTER packet, form with labeled slots is contained in the resource bid packets of information simultaneously, it is a kind ofly to be generated by central control node, is the two-dimentional RSVP of CLUSTER packet reserving network resources according to the whole network resource information and the space-time solicited message of the whole network resource state information database.
1) time mark
The CLUSTER packet is defined as " time mark " (T along the accurate moment that route arrives each node in the link
Ikn), represent with following formula:
T
Ikn＝L
Ikn/v+T
IIn 1)
T
L＝L/P 2)
Wherein, T
Ikn(n=1,2 ... m; M represents to have the m paths) expression CLUSTER packet arrive from network edge node I along alternative path n intermediate node K (k=1,2 ... N; N represents that N node arranged on the path) the moment; L
IknDistance between the intermediate node K on expression source node I and the alternative routing channel n; V represents the light velocity in the optical fiber, v=c/n
1, c is the light velocity in the vacuum, n
1Be optical fibre refractivity; T
IInThe expression burst packets is from the moment of fringe node I transmission; T
LExpression packet time span, L represents the long bit number of packet bag, P represents network data transmission speed; T
Ikn+ T
LThe expression burst packets is left the moment of intermediate node K, as shown in Figure 6.
A most important characteristics of this agreement is: in case the CLUSTER packet sends at source node, it (comprises each node that passes through) and will make the transparent transmission (also not needing optical buffer at middle each node) of no additional delay with the light velocity in whole network, as long as therefore determined CLUSTER packet sending the time at source node, just can predict its " time mark " exactly, thereby avoid the waste with resource of losing of estimation control lag that the rough estimate offset time causes and the accumulation of working control delay error and packet that time delay caused by way of each node.Central control node is sought an optimal channel all available on time and space according to time mark inquiry the whole network resource status data for the data burst bag, has realized the statistic multiplexing of Internet resources, and has greatly reduced network blocking probability.
2) free token
As shown in Figure 7, free token is similar to the labeling method in the Multiprotocol Label Switching Protocol, to wait to transmit on the alternative path of CLUSTER packet each address of node and certain port attribute, in resource feedback information bag and control grouping, confirm that with node symbol (Node or Router ID) and port confirm symbol (Port ID) expression respectively as spatial information.Each node is given in the network input port and output port are called local port, and the port of other adjacent node belongs to far port.Port label can be the port of any granularity in the light territory, comprises three necessary informations: the bandwidth of port switch type, maximum (but and bandwidth reserved), digital coding type.The purpose of introducing free token is to utilize the routing iinformation realization to have the reasonable distribution of the port resource of space attribute.The free token of each node port has uniqueness and differentiability in the network in network-wide basis.
5. core network core node
The major function of CLUSTERNET network core node comprises: 1) receive and opto-electronic conversion is carried out in the control grouping, read time tag information and the free token information and the wavelength information of control grouping; 2) according to the indicated time point of the time mark that is positioned at the labeled slots top optical switch on a certain wavelength channel is switched to the port that the corresponding space mark is indicated; 3) eject labeled slots top information; 4) rewrite the control grouping information, and send to next jumping core node; 5) before the CLUSTER packet arrives core node, the control grouping is for it configures light path, and the CLUSTER packet can not have any obstruction ground with full optically transparent manner and directly passes through core node.
CLUSTERNET server technology is to adopt the space-time tag protocol that the whole network resource (comprising bandwidth, optical switch etc.) is carried out unified centralized management.The main execution mode of management is, by the situation that takies of the whole network resource status data-base recording Internet resources, network edge node sends service request to central control node, and the central control node Query Database is with Resources allocation and refresh data library information.Compare with the distributed control that is widely used in optical burst switched network at present, the present invention can realize the statistic multiplexing of the whole network resource, improve switching performance, greatly reduced network blocking probability, more be applicable to the transmission of plate business (CLUSTER) on all optical network.
6.NON-CLUSTERNET network
The NON-CLUSTERNET network is to be non-plate (NON-CLUSTER) business, and for example traditional telecommunication service, small documents business, low service quality (QOS) business etc. provide the network of low delay, low blocking rate.Its principal character is: the logical construction and physical structure and signaling protocol such as the JET that adopt traditional light burst exchange network.The present invention does not do change to the system and the transfer approach thereof of NON-CLUSTERNET network in the prior art, therefore is not described in detail.
Another object of the present invention is to provide and the corresponding transfer approach of above-mentioned network system, comprise step:
1) business that will enter network edge node is divided into plate and non-plate business (packet), and is sent to the network edge processor of optical core network and optical burst switched network respectively;
2) in optical core network, the network edge processor extracts professional raw information, generates the resource bid packets of information that comprises the packet attribute, is sent to central control node, simultaneously, the plate packet is transferred to the network edge processor sends the buffer unit storage;
3) central control node generates the resource feedback information bag that comprises time mark and free token information or reserve resource failure information according to space-time mark routing algorithm and returns to the core network edge processor;
Wherein, space-time mark routing algorithm is according to network topology and the whole network resource state information database, is the be responsible for assigning routing algorithm of an optimal path of data burst by what central control node was implemented.Algorithm mainly divided for four steps: a) the core network central control node is according to network topological information, adopt shortest path first to calculate the temporary space routing table that comprises n bar alternative path, and according to the ordering of the length in path, shortest path is positioned at the route heading, and the like; B) generate the time mark that arrives each node on above-mentioned every alternative path, generate corresponding time routing table; C) in regular turn time mark in the pairing time routing table of alternative path in the temporary space routing table and the temporal information in the whole network resource state information database are compared, if all node time corresponding sections are all available on a certain alternative path, confirm that then this alternative path is an optimal path, detailed process is: packet compared through the occupied moment of this node in moment of each node on the shortest path of determining by above-mentioned routing algorithm and the whole network resource state information database, if none is overlapping, then this path is available optimal path; Overlapping as if having, then compare, till finding not have constantly overlapping path with next bar second shortest path.If than right n (n bar alternative path) inferior still have overlapping, then this packet reserve resource failure; Otherwise packet is gone up the moment of each node and the space coordinates of each node and is time tag information and free token information by way of the path of reserve resource success (optimal path), be collectively referred to as the space-time label information; D) time mark and the free token information with optimal path writes the whole network resource state information database.Although the optimal path that space-time mark routing algorithm calculates is shortest path not necessarily spatially, be one and be the conflict free transparent transmission channel of the special assignment of this CLUSTER packet.
Space-time tag protocol and the space-time mark routing algorithm situation that all-optical network signaling is for a long time separated with route that fundamentally is through with, and for the CLUSTER packet provides the connection-oriented service that can guarantee QOS, that has realized Internet resources to the full extent has made full use of and has guaranteed professional transmission quality.
4) the core network edge processor abandons the raw data packets of reserve resource failure according to the information of resource feedback information bag, otherwise, generation comprises the control grouping of the wavelength information of the offset time information of time period information, control grouping that optimal channel port information, CLUSTER packet arrive the time of each node and take node resource along optimal channel and packet and transfer data packets in interior control information, and sends the control grouping and packet arrives control corresponding channel and data channel at preset time point.
5) in the process of control packet distribution, each core node is distributed the space-time mark along optimal channel after receiving the space-time mark, is labeled as packet according to space-time and sets up the space-time label switched path, resets the space-time labeled slots subsequently, sends out to next redirect.
More than from step 2) begin to step 5) be the process that packet transmits optical core network, i.e. the core of method provided by the present invention.After packet arrived destination node, the network edge Egress node was destroyed the control grouping, and added up the data burst bag number that successfully receives, and calculated packet loss.
Enter the NON-CLUSTER business of NON-CLUSTERNET network edge processor from professional grader, after edge processor is finished route and is generated the task of control grouping, packet is assigned with a wavelength at random, after the control grouping transmits an offset time, begin to transmit with the JET signaling protocol.Similarly, after packet arrived destination node, the network edge Egress node of optical burst switched network was destroyed the control grouping, and added up the data burst bag number that successfully receives, and calculated packet loss.The translator unit of above-mentioned packet in the NON-CLUSTER network belongs to prior art, and the present invention no longer does detailed description.
Technique effect of the present invention is: compare with the present all optical network networking plan that proposes, the present invention is according to core optical network loaded service type of future generation and feature, proposed a kind of based on good, with low cost on technical feasible, the performance, eclipsed form light burst-switched core network network-building method efficiently.Be characterized in that two kinds of overlapped dissimilar business of business network carrying have separate logical construction and physical structure, and are continuous by the professional grader of fringe node; Set up the CLUSTER service transmission channels of at a high speed, extremely hanging down blocking rate by this grader for many (stream) medium with sudden and plate (CLUSTER) and the large file service that account for more than the Network total amount 75% (even up to 90%); Adopting the space-time tag protocol that not only has the resource assignation function but also have a routing function is the unified scheduling that the professional optical core network of CLUSTER (CLUSTERNET) carries out the whole network resource, realized the statistic multiplexing of Internet resources, thereby abundant bandwidth resources are fully utilized.On the other hand, the arithmetic speed of high-speed digital signal process chip DSP and the switch speed of fast optical switch based have reached 100 psecs and 10 nanosecond orders respectively, make that focusing on, calculate and allocate route can finish in the microsecond magnitude, and data packet traffic time spans such as many (stream) medium mostly are a millisecond magnitude, make that utilizing central centralized way to assign route for these long packets becomes possibility, only is the one thousandth of this data packet length or lower because spend in the time of calculating with assigning route! Therefore ripe hardware technology provides solid hardware foundation for the application of high-speed data service.
Description of drawings
Fig. 1 is a light burst-switched pattern diagram;
Fig. 2 is the light burst exchange network schematic diagram;
Fig. 3 represents that the CLUSTER business accounts for the percentage of total traffic with self similarity characteristic parameter (Hurst parameter);
Fig. 4 is a light burst-switched overlapping network schematic diagram;
Fig. 5 is a CLUSTERNET network edge node functional schematic;
Fig. 6 express time mark schematic diagram;
Fig. 7 representation space mark schematic diagram;
Fig. 8 represents the NSFNET network topology;
Fig. 9 represents the professional transfer process schematic diagram of overlapping network;
Figure 10 represents CLUSTERNET cellular logic structural representation;
Figure 11 represents relatively schematic diagram of NSFNET network JET agreement and space-time tag protocol (TSL-SP) blocking performance.
Embodiment
In order more clearly to describe content of the present invention, below further describe the present invention, but be not construed as limiting the invention by concrete enforcement example.
Present embodiment adopts the international NSFNET network topology with 21 links of 14 nodes that the network system of the present invention's proposition and the network performance of transfer approach thereof are tested by Computer Simulation, the network topology of NSFNET as shown in Figure 8, the service source that is adopted is that packet is obeyed self similarity the time of advent at interval and distributed, the long negative exponent of obeying of bag distributes, and average is the self similarity service source of 200Mbit.
With reference to Fig. 4, the network system of present embodiment comprises professional grader, optical core network, optical burst switched network.Wherein, optical core network has identical physical topology with optical burst switched network, includes central control node, core node and network edge node.The function of each part in the present embodiment network system will obtain embodying in following specific descriptions to transport process.
With reference to Fig. 9, present embodiment utilizes above-mentioned network system to realize that the concrete steps of high-speed data service are as follows:
1) the self similarity business professional grader that enters network edge node packages and type of service is differentiated.Adopt the method for the time thresholding judgement CLUSTER business that proposes in the prior art " method for self similar cluster packet of large file service optical transmission " (number of patent application 200510086709.9) that the self similarity business that enters professional grader is adjudicated, and CLUSTER and the NON-CLUSTER business that will organize bag judgement back generation are sent to CLUSTERNET network and NON-CLUSTERNET network respectively.
2) in conjunction with Figure 10, at first, behind the professional arrival of the CLUSTER CLUSTERNET network edge processor, CLUSTERNET network edge processor extracts professional raw information, comprises destination address, bag long letter breath and the grade of service; Secondly, CLUSTERNET network edge processor produce comprise that routing information request, CLUSTER packet bag are long, the grade of service is in the resource bid packets of information of all properties of interior CLUSTER packet; Edge processor sends to CLUSTERNET network central control node with the resource bid packets of information then; At last, the CLUSTER packet is transferred to the core network edge processor and sends the buffer unit storage, waits for the resource feedback information bag that returns.
3) after central control node is received the resource bid packets of information that fringe node sends,, adopt shortest path first to set up routing table (LSPA1, LSPA2, LSPA3 etc. as shown in Figure 10) at first according to network topological information.Optimal path is positioned at the routing table top, and the like.By formula 1), 2) calculate the initial moment T that every paths in the routing table takies the node bandwidth resource
IknWith termination moment T
Ikn+ T
L, i.e. rise time mark; Then the temporal information of having been preengage that writes down in the time mark of optimal path and the resource state information storehouse is compared, if not conflict, then this path is identified, otherwise continues relatively sub-optimal path, be identified until passage, the passage that is identified is an optimal channel; Then, central control node produces the resource feedback information bag of the time tag information, free token information and the wavelength information that comprise optimal channel, and sends it back the core network edge processor; At last, the time of optimal channel and free token information are write the whole network resource state information database.
4) after the core network edge processor is received the resource feedback information bag that returns from central control node, abandon the raw data packets of reserve resource failure according to feedback information, otherwise, generation comprises the optimal channel port information, the CLUSTER packet arrives the time and the time period information that takies node resource of each node along optimal channel, the offset time information of control grouping and packet and transmit of the control grouping of the wavelength information of data burst bag in interior control information, and send the control grouping and and send the data burst bag subsequently and arrive control corresponding channel and data channel at preset time point.
5) in the process of control packet distribution, each core node is distributed the space-time mark along optimal channel after receiving the space-time mark, according to time mark T
AB1N(n=1 among Figure 10,2 ... 4, B) reserve the node n be labeled port Pn (n=1 among the figure, 2 ... 4, bandwidth time slot (T b)
AB1N+ T
L) set up space-time label switched path (TS-LSP) for the data burst bag.Reset the space-time labeled slots subsequently, send out to next redirect.
6) after the data burst bag arrived destination node, the network edge Egress node was destroyed the control grouping, and added up the data burst bag number that successfully receives, and calculated packet loss.
7) enter the NON-CLUSTER business of NON-CLUSTERNET network edge processor from professional grader, after fringe node is finished route and is generated the task of control grouping, packet is assigned with a wavelength at random, after the control grouping transmits an offset time, begin to transmit with the JET signaling protocol.
8) after the data burst bag arrived destination node, the network edge Egress node was destroyed the control grouping, and added up the data burst bag number that successfully receives, and calculated packet loss.
Under identical services source and network environment, the result and the networking plan that adopts the JET agreement of the foregoing description have been carried out network performance emulation contrast, its result is as shown in figure 11.Figure 11 demonstrates the light burst exchange network of the traditional JET agreement of under self similarity service source employing and the packet loss of the burst packets of the CLUSTERNET network that adopts the space-time tag protocol.As we can see from the figure, the switching performance of CLUSTERNET network is far superior to JET, particularly under the condition of high flow capacity, the space-time tag protocol will be improved 3 orders of magnitude than JET performance, so the space-time tag protocol can overcome the shortcoming of traditional light burst-switched agreement poor performance under high flow condition.This shows that new system of high-speed data service network and transmission method thereof that the present invention proposes can be the transmission service that next generation network service provides high-quality, the contour performance of high efficiency, and can significantly reduce network cost.
More than be embodiments of the present invention, according to content disclosed by the invention, some identical, replacement schemes that those of ordinary skill in the art can expect apparently all should fall into the scope of protection of the invention.
Claims (9)
1. transfer approach that is used to carry high-speed data service comprises step:
1) packet that will enter network edge node is divided into plate packet and non-plate packet, is sent to the network edge processor of optical core network and optical burst switched network respectively;
2) core network edge processor information extraction from the packet that receives, the resource bid packets of information that generation comprises the packet attribute is sent to the core network central control node, simultaneously the plate packet is transferred to the core network edge processor and sends buffer unit;
3) the core network central control node generates according to space-time mark routing algorithm and comprises the space-time label information of optimal path or the resource feedback information bag of reserve resource failure information returns to the core network edge processor;
4) the core network edge processor abandons the packet of reserve resource failure according to the information of resource feedback information bag, the packet that is the reserve resource success generates the control grouping, and sends the control grouping and packet arrives control corresponding channel and data channel at preset time point;
5) in the process of control packet distribution, the core network core node is distributed the space-time mark along optimal channel after receiving the space-time label information, be labeled as packet according to space-time and set up the space-time label switched path, reset the space-time labeled slots subsequently, send out to next redirect.
2. the method for claim 1, it is characterized in that, also comprise step: the non-plate packet that is sent to light burst exchange network is after the network edge processor is finished route and generated the task of control grouping, packet is assigned with a wavelength at random, after the control grouping transmits an offset time, begin to transmit with enough time signaling protocols.
3. method as claimed in claim 1 or 2 is characterized in that, also comprises step: after packet arrived destination node, the network edge Egress node was destroyed the control grouping.
4. the method for claim 1 is characterized in that, also comprises step in the described step 3): the time mark and the free token information of optimal path are write the whole network resource state information database.
5. the method for claim 1 is characterized in that, described space-time mark routing algorithm comprises step:
A) the core network central control node adopts shortest path first to calculate the temporary space routing table that comprises n bar alternative path according to network topological information, and sorts according to the length in path, and shortest path is positioned at the route heading;
B) generate the time mark that arrives each node on above-mentioned every alternative path, generate corresponding time routing table;
C) in regular turn time mark in the pairing time routing table of alternative path in the temporary space routing table and the temporal information in the whole network resource state information database are compared, if all node time corresponding sections are all available on a certain alternative path, confirm that then this alternative path is an optimal path, the moment of packet each node on optimal path and the space coordinates of each node are the space-time label information, if can not find then reserve resource failure of optimal path.
6. the method for claim 1 is characterized in that, described control grouping comprises optimal channel port information, packet and arrives the time of each node and take the time period information of node resource along optimal channel.
7. network system that is used to carry high-speed data service comprises:
Professional grader is positioned at network edge node, connects optical core network and optical burst switched network respectively;
Optical core network is the core net that transmits the packet with plate character, comprising: core network edge processor, core network central control node and core network core node;
Optical burst switched network is the network that transmits non-plate packet.
8. network system as claimed in claim 7 is characterized in that, described optical core network has identical physical topology with optical burst switched network.
9. network system as claimed in claim 7 is characterized in that, described optical core network has the central management face that is used to allocate the whole network resource, unified allocation of resources route and signaling on the logical topology structure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006100587124A CN100407707C (en) | 2006-03-03 | 2006-03-03 | Network configuration for loading high-speed data business and trans mitting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006100587124A CN100407707C (en) | 2006-03-03 | 2006-03-03 | Network configuration for loading high-speed data business and trans mitting method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1816018A CN1816018A (en) | 2006-08-09 |
| CN100407707C true CN100407707C (en) | 2008-07-30 |
Family
ID=36907981
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2006100587124A Expired - Fee Related CN100407707C (en) | 2006-03-03 | 2006-03-03 | Network configuration for loading high-speed data business and trans mitting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100407707C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101621714B (en) | 2008-06-30 | 2013-06-12 | 华为技术有限公司 | Node and data processing system and data processing method |
| CN104427377B (en) * | 2013-09-09 | 2018-05-18 | 杭州海康威视数字技术股份有限公司 | The point-to-point optical transmitter and receiver of polymorphic type business |
| CN106063200B (en) * | 2013-12-09 | 2019-11-01 | 华为技术有限公司 | A kind of device for content caching, method |
| CN105337658A (en) * | 2015-10-29 | 2016-02-17 | 电子科技大学 | Method for designing incremental monitoring traces in all-optical network based on service routes |
| CN105610494B (en) * | 2015-10-29 | 2018-05-04 | 电子科技大学 | A kind of monitoring mark design method based on node pre-selection in all-optical network |
| CN110730470B (en) * | 2019-10-24 | 2020-10-27 | 北京大学 | Mobile communication equipment integrating multiple access technologies |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1089498A2 (en) * | 1999-09-30 | 2001-04-04 | Alcatel | Control architecture in optical burst-switched networks |
| CN1377154A (en) * | 2002-04-30 | 2002-10-30 | 上海交通大学 | Multi-wavelength optical albel transceiver based on wavelength-time encode method |
| WO2005076625A1 (en) * | 2004-01-30 | 2005-08-18 | Wiltel Communications Group, Llc, | Method for the transmission and distribution of digital television signals |
| CN1725756A (en) * | 2005-07-21 | 2006-01-25 | 上海交通大学 | Passive optical network system based on generalized multiprotocol label switching (GMPLS) protocol |
-
2006
- 2006-03-03 CN CN2006100587124A patent/CN100407707C/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1089498A2 (en) * | 1999-09-30 | 2001-04-04 | Alcatel | Control architecture in optical burst-switched networks |
| CN1377154A (en) * | 2002-04-30 | 2002-10-30 | 上海交通大学 | Multi-wavelength optical albel transceiver based on wavelength-time encode method |
| WO2005076625A1 (en) * | 2004-01-30 | 2005-08-18 | Wiltel Communications Group, Llc, | Method for the transmission and distribution of digital television signals |
| CN1725756A (en) * | 2005-07-21 | 2006-01-25 | 上海交通大学 | Passive optical network system based on generalized multiprotocol label switching (GMPLS) protocol |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1816018A (en) | 2006-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200196034A1 (en) | Reduced-complexity integrated guaranteed-rate optical packet switch | |
| Rosberg et al. | Performance analyses of optical burst-switching networks | |
| CN100407707C (en) | Network configuration for loading high-speed data business and trans mitting method | |
| KR20080088351A (en) | Dynamic bandwidth allocation and service differentiation for broadband passive optical networks | |
| US10305785B2 (en) | Adaptive traffic routing in communication networks | |
| Bjornstad et al. | A packet-switched hybrid optical network with service guarantees | |
| Bouabdallah et al. | Resolving the fairness issues in bus-based optical access networks | |
| González-Ortega et al. | LOBS-H: an enhanced OBS with wavelength sharable home circuits | |
| Chua et al. | Quality of service in optical burst switched networks | |
| CN1794880B (en) | Parallel dispatching processing method in optical burst switched network | |
| CN101753450A (en) | Method for joint optimization of resources in three-layered network | |
| Xu et al. | Dual-fiber-link OBS for metropolitan area networks: modelling, analysis and performance evaluation | |
| Li et al. | A distributed fast light-path set-up protocol for grid over obs networks | |
| Mohan et al. | Efficient techniques for improved QoS performance in WDM optical burst switched networks | |
| Cao et al. | Dynamic optical packet switching network with advanced SDN/OpenFlow control | |
| Jin et al. | APOSN: operation, modeling and performance evaluation | |
| Nandi et al. | A new contention avoidance scheme in optical burst switch network | |
| Garg et al. | Performance analysis of optical burst switching high-speed network architecture | |
| Singh et al. | Dynamic bandwidth allocation in GMPLS optical networks using minimum execution time technique | |
| Pallavi et al. | A simple node architecture for optical burst switching | |
| Naji et al. | Impact Of Using Fiber Delay Scheme On Burst Loss Ratio And Delay Using Offset Time Algorithm For Optical Burst Switching Networks | |
| Priya et al. | Wavelength division multiplexing for optical burst switched networks | |
| Liamcharoen | Performance analaysis of a proposed hybrid optical network | |
| Radivojević et al. | The way forward truly QoS aware EPON | |
| Okorafor | Integrated hybrid optical node using dynamic reconfiguration and routing algorithms for bimodal traffic |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080730 Termination date: 20130303 |