CN1514605A - Hierarchical exchange network system - Google Patents
Hierarchical exchange network system Download PDFInfo
- Publication number
- CN1514605A CN1514605A CNA03120998XA CN03120998A CN1514605A CN 1514605 A CN1514605 A CN 1514605A CN A03120998X A CNA03120998X A CN A03120998XA CN 03120998 A CN03120998 A CN 03120998A CN 1514605 A CN1514605 A CN 1514605A
- Authority
- CN
- China
- Prior art keywords
- network
- address
- hierarchical
- nodes domains
- port
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 claims abstract description 75
- 238000009826 distribution Methods 0.000 claims description 19
- 238000005538 encapsulation Methods 0.000 claims description 12
- 230000002093 peripheral effect Effects 0.000 claims description 10
- 238000004891 communication Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 230000009191 jumping Effects 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 2
- 230000008520 organization Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 23
- 238000013461 design Methods 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 230000002776 aggregation Effects 0.000 description 5
- 238000004220 aggregation Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 238000003780 insertion Methods 0.000 description 3
- 230000037431 insertion Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 239000012467 final product Substances 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 238000000275 quality assurance Methods 0.000 description 2
- 101710104937 Non-specific acid phosphatase Proteins 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The system includes a group of sets containing methods for organizing network and control methods including topological structure organization method, address assignment method, data packet exchange and control method etc. The system organizes topological structure of network and assigns address space of network based on hierarchy. Topological hierarchical structure and address hierarchical structure are matched strictly. Hierarchical switched network built based on the arrangement above can realize exchanging data packets rapidly according to hierarchical address, replacing IP route by IP exchange completely.
Description
Affiliated field:
The present invention relates to a kind of network system, adopt hierarchy type IP exchange (IP Switching) technology to replace IP route (IP Routing) technology especially.
Background technology:
Along with increasing rapidly of network size and number of users, Internet just progressively becomes following information basic platform, but the progress of new application demand and mechanics of communication order about the Internet architectural goals from opening, integrated to high-performance, can expand management orientation development.Original Internet architecture Design condition changes, and some principles wherein are no longer suitable under existing network technology condition.For example at present for guaranteeing network reliability and convenient interconnected, employing be the address structure of no center, structureless topology design and one dimension tiling, advantages such as this structure has flexibility is arranged, robustness.But along with network size expansion and user are exponential increase, this structure will cause network to be challenged at aspects such as performance, efficient, extensibility, service quality assurance, manageabilitys, be in particular in: the address structure and the structureless address allocation policy of one dimension tiling have caused huge routing table, the bottleneck that causes the core router bag to transmit; Structureless network topology and the topology design of separating with the address cause the overall routing iinformation to complexity to exchange with route calculating, thereby consumed the core router ample resources, and caused large-scale, overall, dynamic Routing Protocol to have that stability is not enough, convergence time waits problem slowly; Uncertain route causes that easily bursts of traffic and flow rate mode are difficult to prediction, thereby causes service quality assurance and flow control difficulty or the like.The structureless arbitrary topology of current Internet connects and address structure, and the network address characteristic of separating with topology, is to cause present Internet routing issue, service quality to guarantee and the insoluble basic reasons of problem such as performance expansion.
The current Internet route is finished by two levels, is divided into route in inter-domain routing and the territory.Border Gateway Protocol (BGP) 4 (Border Gateway Protocol 4) is the unique de facto standard of present inter-domain routing.This agreement is moved in about 100,000 routers, exchanges the reachability information between global Internet autonomous system, thereby makes Internet become one of maximum in the world distributed system.But the address structure of one dimension tiling, unordered address allocation policy, topology are separated with the address, overall situation feature dynamic and in large scale has caused BGP4 to have serious problems at aspects such as extensibility, convergence efficiency, stability, are difficult to bear more extensive Internet route task in future.The hierarchical routing method that F.Kamoun and L.Kleinrock proposed (" Hierarchical Routing forLarge Networks:Performance Evaluation and Optimization ", Computer Networks, 1 (3): 155--174, January 1977) and PNNI (Private Network-Network Interface) by the topology, converging of address, hide with abstract technology realization information, the route of constitutive logic hierarchical structure on the network of arbitrary topology, solve the routing table scalability problem, but cost is to have increased average path length, and information is abstract very complicated with route calculating.
For the forwarding speed that improves router with provide service quality to guarantee, people attempt to replace route with exchange as far as possible, IP/MPLS route/label switched structure (Rosen has been proposed, E., Viswanathan, A.and R.Callon, " A Proposed Architecture for MPLS ", RFC 3031, and January 2001), the advantage of comprehensive utilization network core switching technology and network edge IP route technology.Under the IP/MPLS structure, route is kept away not open forever, can only accomplish partly to reduce the routing table lookup number of times, this is a bit improving on existing bag retransmission technique basis, but do not tackle the problem at its root: the first, the huge routing table of complicated routing table generative process, storage, and to routing table time-consuming inquiry work all not have elimination; The second, do not have quick, strict method to the independent IP bag that arrives at random be belong to same stream also the same stream of right and wrong assert effectively; The 3rd, label switching has also been introduced more complicated problems, and for example generation of mark, distribution, management, mark make core net inside do more complicated and more complicated to the inquiry of port and the interoperability of mapping, label switched and other agreements etc.MPLS has increased the complexity of Internet core, has departed from the Internet structure.
Said method and technology are to the perfect of existing Internet structure and repair do not have important breakthrough.
Goal of the invention:
In order to solve huge, the overall Routing Information Exchange of routing table that current Internet causes because of reasons such as the topology that connects arbitrarily, one dimension tiling and unordered address assignment and address and topological no direct relations and the problem of calculation of complex, the invention provides a kind of hierarchical switched network (Hierarchically Switched Network) system, hereinafter to be referred as HSNET, be used to solve the problem of route scale and network performance extensibility.
The technical scheme of invention:
The technical solution adopted in the present invention is as follows:
Hierarchical switched network mainly by the rule of stipulating network topology, the network address again and the relation that must satisfy between the two, thoroughly replaces the IP route with the IP exchange.Its basic scheme comprises: by hierarchical structure tectonic network topological structure, by the address space of hierarchical structure distribution network, the level of topological structure and the strictness of the level of address structure are mated equally; In the network of setting up according to above method, realize the quick exchange of packet, thereby avoid routing operations according to the level address.
Hierarchical switched network is the set of a group network method for organizing and control method, mainly comprises topological structure method for organizing, address distribution method, packet switch control method etc.
1. the basic terms used of the present invention:
(1) hierarchy type exchange (Hierarchical Switching): hierarchical structure and strict coupling are adopted in the topology of network and address, and packet is transmitted the exchange that the corresponding level that relies on destination address is realized packet.
(2) (Hierarchically Switched Network, HSNET): refer to meet the network of hierarchy type exchange, the use hierarchy type exchanges hierarchical switched network in this network.
(3) nodes domains: also be logical node or dummy node, be one of HSNET basic logic element, refer to the autonomous territory of being formed by the server (as strategic server, bandwidth broker server, all kinds of management information banks etc.) of one group of interconnected switch and other functions.
(4) node: the switch of nodes domains inside and other function servers are referred to as node.
(5) rhizosphere: be in the nodes domains of hierarchical switched network top layer, rhizosphere can carry out interconnected with other HSNET networks or non-HSNET network.
(6) father field: the nodes domains that is in certain direct upper strata of nodes domains.
(7) subdomain: the nodes domains that is in certain direct lower floor of nodes domains.
(8) fraternal territory: all nodes domains of certain direct lower floor of nodes domains claim fraternal territory mutually.
(9) link: realize the media or the facility of inter-node communication in data link layer, as ppp link, Ethernet, X.25, frame relay, ATM net or high-rise tunnel etc.
(10) peripheral link: the link that this nodes domains links to each other with other nodes domains.
(11) inner link: the link that connects this nodes domains internal node.
(12) logical links: also claiming virtual link, is one of HSNET basic logic element.Article one, logical links refers to that this nodes domains connects the set of all peripheral links of other certain nodes domains.The logical links of a nodes domains comprises that one is pointed to the logical links in father node territory, logical links and several short circuit logical links in several sensing child node territories.
(13) port: port is that node is used for and link-attached parts.In nodes domains, port is identified at the label of this node by node ID and port.Port generally comprises input, output two parts.In nodes domains, be divided into outside port and internal port according to the link difference that is connected.
(14) outside port: the port that links to each other with peripheral link claims outside port.An outside port can be divided into input outside port and output outside port two parts.
(15) internal port: the port that links to each other with inner link claims internal port.
(16) logic port:, be of the outside port set of the peripheral link that comprises of logical links in this nodes domains with logical links port one to one.
(17) neighbours: the node that is connected to same link claims neighbours mutually.
(18) boundary node: the node that links to each other with other networks (HSNET network or non-HSNET network).
(19) boundary port: the port that links to each other with other networks (HSNET network or non-HSNET network).
2.HSNET topological structure method for organizing
The method that HSNET topological structure method for organizing has been adopted when having stipulated to set up the hierarchical switched network topological structure.
At first, can be divided into two parts to network: backbone network and Access Network according to the difference of management domain.Fig. 1 is the two-layer logical construction schematic diagram of network, and (1) is backbone network, and (2) are the backbone network interior nodes, (A)-(H) is Access Network.Backbone network is by operator's operation and management, and Access Network is managed by Subscriber Unit.Metropolitan area network is included in the backbone network, is responsible for by operator.The unit that has Access Network can be a school, a hospital, a company, a government department, a residential quarter etc.All user access networks are coupled together with a backbone network.HSNET is meant that backbone network follows HSNET and set up, and operation HSNET agreement.
Then, the topological structure of HSNET backbone network is formed tree by a plurality of exchange levels.Because the link of any two nodes is unique in the simple tree, therefore simplified route.But there is intrinsic shortcoming in it at aspects such as reliability, autgmentability, flexibilities: if for example there is Single Point of Faliure, the inefficacy of channel or node so, can make all subtrees that connect under this channel or the node lose connectivity, thereby cause the disconnection of a collection of subordinate's subtree; If the capacity of the bandwidth of single channel and single switch is limited, then there is scalability problem; Because the path between the tree arbitrary node is unique, so can not satisfy a plurality of inter-entity route need flexibly.Therefore when design HSNET topology of networks, the present invention expands the tree of strictness, to avoid above-mentioned shortcoming.Expansion comprises two kinds:
First kind is based on logical node territory and logical links expansion.This extended method is used to solve reliability and capacity scalability problem.This method substitutes the node of tree with logical node (or claiming nodes domains), substitute the branch of tree with logical links, being basic element with logical node and logical links is organized in logic hierarchical structure network by tree, keeps the hierarchical structure characteristics.From realizing angle, there are a plurality of nodes interconnected in each nodes domains, this has guaranteed the extensibility of switching node capacity and the flexibility of Path selection.Because every logical links has comprised many physical links, therefore expanded the bandwidth of logical links, strengthened the ability that the wall scroll physical channel lost efficacy of resisting.Therefore the logical extension method had both kept the characteristics of tree, had guaranteed reliability and extensibility again.
Fig. 2 is based on local hierarchy's formula switching network schematic diagram of logical node territory and logical links expansion, and (1) is nodes domains, and (2) are logical links, and (3) are server, and (4) are switch.Nodes domains (1) is a switching node on the tree, and its function such as same switch are finished the flow forwarding of both direction up and down.Connecting peripheral link set between two adjacent node territories abstract is logical links (2), and each logical links is made up of many peripheral links.For a certain nodes domains, a up logical links and one group of descending logical links are arranged: up logical links is connected the father node territory; Descending logical links has many, connects a plurality of child nodes territory.Each descending logical links is by this node layer territory management, and distributes a logical links number, and this logical links number is exactly the subdomain number in next straton territory in fact.
Second kind of expansion that is based on short circuit.This extended method is used to solve flexibility and efficiency.This method increases direct link between different subtrees, allow have the second active path between subtree, thereby make the flow between subtree not necessarily need their common ancestor node territories to transmit.In based on the short circuit expansion, the short circuit link is a partial autonomy, has only the existence of this short circuit of nodes domains perception of short circuit both link ends direct correlation.The method of short circuit expansion has kept topological level, has topological structure flexibly simultaneously.The realization of short circuit link can realize according to the make of logical links in first kind of extended method, promptly is made up of one group of physical link.
Expansion has five kinds of application scenarioss based on short circuit:
(1) first kind of application scenarios is to have two even more link service provider to adopt hierarchical switched network to provide the network insertion service to the user in areal (for example same city): these link service providers' (ISP) hierarchical network may cross one another overlapping, the a collection of hospital that for example is in same city receives on the hierarchical network that a link service provider provides, and some insurance companies in same city have received on the hierarchical network that another link service provider provides; The top layer of two trees could arrive the other side if communicating by letter between hospital and the insurance company all will be detoured, and the interconnection of this top layer may be beyond thousands of kilometers, and this is irrational, therefore needs to introduce the short circuit expansion.
(2) though second kind of application scenarios is to have only a link service provider that the hierarchical network access service is provided in the area (for example city), has only stalk tree, but some inserts the user is a group, at different subregions a plurality of branches are arranged, the bigger traffic is arranged between the branched structure, or higher communication quality requirement is arranged; Therefore require the link service provider between it has the bottom layer node territory that branch inserted of the big traffic, to set up channel, directly connection.
(3) the third application scenarios is that link service provider self passes through monitoring and the analysis to network traffics, and thinking should increase a direct channel between certain two sub regions (subtree), can optimize network performance, therefore carries out short circuit between two subareas.
(4) the 4th kinds of application scenarioss carry out short circuit when being HSNET and the interconnection of non-HSNET net, and making does not need the forwarding of process overlay network between HSNET network and non-HSNET net.
The sensible the other side's of the short circuit of (5) the 5th kinds of application scenarioss whole network, and be not the other side's stalk tree.
The application scenarios of above-mentioned preceding four kinds of short circuit demands all belongs to the locality short circuit, it is long-range that main purpose is that local flow does not detour, make the flow between the different branched network of the network that is in different I SP in this area or same ISP network more reasonable, play the local optimum effect; The 5th kind then is short circuit of overall importance.
In order to provide when networking maximum flexibility, become hierarchical switched network in the network switch that will have netted connection now, the short circuit mode allows to carry out arbitrarily short circuit.Short circuit comprises following method arbitrarily:
(1) allows that many arbitrarily short circuit links are arranged in the hierarchical network;
(2) allow a nodes domains that other any a plurality of nodes domains are had the short circuit link;
(3) allow to be between the nodes domains of any different levels of tree the short circuit link is arranged.
By any short circuit, do not add in network topology structure under the situation of any change, can become the network that exchanges based on level to the existing network rebuilding.Fig. 3 is the schematic diagram of any short circuit expansion example, wherein (D
1)-(D
10), (D
1')-(D
10') be nodes domains, (A)-(H) be Access Network, (1)-(8) are the short circuit links.
HSNET network and other network interconnections have three kinds of patterns: comprise interconnected with other HSNET, with the non-HSNET network interconnection and interconnected with Access Network.
With other HSNET are interconnected two kinds of methods are arranged: a kind of is that short circuit is interconnected, is promptly undertaken interconnected by the short circuit link between two HSNET net top mode territories.Its advantage is interconnected simple, only need configuration top mode territory to get final product, and shortcoming is can only be interconnected between two HSNET, can not realize the interconnected entirely of a plurality of (more than three) HSNET, as shown in Figure 3.Another kind is the top layer extended method, promptly sets up a super dummy node territory as root node, with the child node territory as this super node territory, the top mode territory of each HSNET network, constructs new wider HSNET network.Fig. 4 is a top layer extended method schematic diagram, (D
1)-(D
10), (D
1')-(D
10') be nodes domains, (A)-(H) be Access Network, (1) is HSNET
1, (2) are HSNET
2, (D
0) be the super node territory.
With the method for the non-HSNET network interconnection be: HSNET top mode territory is directly interconnected with non-HSNET server backbone network, perhaps sets up a peer-to-peer with non-HSNET network.The foundation of peer-to-peer need move bgp protocol, perhaps with the HSNET network as lower floor's network of non-HSNET backbone network and need not to move bgp protocol.Fig. 5 is HSNET network and non-HSNET network interoperability schematic diagram, (1) right and wrong HSNET server layer, and the distribution layer of (2) right and wrong HSNET network, (3) right and wrong HSNET network, (4) are the HSNET networks, (A)-(Q) are Access Networks.
With Access Network is interconnected three kinds of methods are arranged: comprise the interconnected and terminal extended method of HSNET with directly interconnected, the many place of current Access Network.Directly interconnected method is fundamental method, is about to user network and directly inserts HSNET leaf node territory (also claiming the HSNET edge port) by router, thereby realize the access of user network; Many places interconnected (Multi-Homing) method refers to that a user network links different ISP or the diverse access point of same ISP, so that increase reliability, equally loaded; The networking of the terminal extended method finger of HSNET itself is also constructed according to hierarchical switched network, thereby has cancelled router when inserting, and realizes the seamless interconnected of HSNET network and Access Network.
Fig. 5 has reflected the scene of HSNET network and other network interoperabilities, and wherein the HSNET network couples together dissimilar Access Networks at bottom, can interconnect at top layer with other HSNET networks or non-HSNET network simultaneously.Fig. 6 provides the schematic diagram of two user access networks (C), many places of (E) employing interconnected method access HSNET network, (D
1)-(D
10), (D
1')-(D
10') be nodes domains, (A)-(Q) be Access Network, (1) is HSNET
1, (2) are HSNET
2User access networks (C) are linked two different subtrees of same ISP net, and user access networks (E) are linked the HSNET network of two different I SP.
3.HSNET address distribution method
The HSNET address distribution method is followed the principle with IPv6 address structure compatibility, observes the IPv6 address allocation plan, sets up consistent address structure at unified address space.This distribution method is structure level address structure in the address space of authorizing, and this address not only has NAPT function, and has addressing function, and promptly the structured message of network can be represented in the network address.HSNET is a backbone network, and its address distribution method relates generally to design, distribution and the management of backbone network layer address, does not relate to the address assignment and the management of user access network and internal node thereof.HSNET address distribution method of the present invention is mainly considered the global unicast address distribution, an IP address is divided into the network address number (Network ID) and interface number (Interface ID) two parts: in the HSNET network, any logic subtree is a level subnet, and its network address is number by the subnet address prefix designates; The last layer network distributes a logical port number for next layer network simultaneously, and next straton net address is the combination of the logical port number of last layer subnet address prefix and its distribution.Pointed out the logical path of packet exchange by hierarchical network address number explicitly.
The address style of IPv6 has three classes: clean culture (Unicast) address, appoint and to broadcast (Anycast) address and multicast (Multicast) address.IPv6 address purposes is specified by the bit preamble pattern of address, claims Format Prefix FP (Format Prefix) again.FP is that 11111111 these addresses of indication are multicast address, other broadcast the address with appointing for clean culture.Divide address space with FP, its purposes, Format Prefix and address space are as shown in table 1.
Purposes Format Prefix (FP) address space (binary system) |
Reserved??????????????????????????????0000?0000????1/256 |
Unassigned????????????????????????????0000?0001????1/256 |
Reserved?for?NSAP?Allocation??????????0000?001?????1/128 |
Reserved?for?IPX?Allocation???????????0000?010?????1/128 |
Unassigned????????????????????????????0000?011?????1/128 |
Unassigned????????????????????????????0000?1???????1/32 |
Unassigned????????????????????????????0001?????????1/16 |
Aggregatable?Global?Unicast?Addresses?001??????????1/8 |
Unassigned????????????????????????????010??????????1/8 |
Unassigned????????????????????????????011??????????1/8 |
Unassigned????????????????????????????100??????????1/8 |
Unassigned????????????????????????????101??????????1/8 |
Unassigned????????????????????????????110??????????1/8 |
Unassigned????????????????????????????1110?????????1/16 |
Unassigned????????????????????????????1111?0???????1/32 |
Unassigned????????????????????????????1111?10??????1/64 |
Unassigned????????????????????????????1111?110?????1/128 |
Unassigned????????????????????????????1111?1110?0??1/512 |
Link-Local?Unicast?Addresses??????????1111?1110?10?1/1024 |
Site-Local?UnicastAddresses???????????1111?1110?11?1/1024 |
Multicast?Addresses???????????????????1111?1111????1/256 |
Table 1:IPv6 address
Fig. 7 is that IPv6 can converge global unicast address structure (Aggregatable Global Unicast Addresses) form: (A) be Format Prefix FP, value is 001,3 bit; (B) be top level aggregation territory TLA, 13 bits; (C) be reserved field RES, be preserved for expanding in the future TLA or NLA territory, 8 bits; (D) be that one deck converges territory NLA, 24 bits down; (E) be site level territory SLA, 16 bits; (F) be interface identifier territory Interface ID, 64 bits; (G) be the public address scope; (H) be website topological address scope.TLA, NLA, SLA are divided into the top layer of three big level: TLA at the route level with the address, distribute to the big tissue that flow transfer service is provided by registration body; NLA is used to obtain the own inner address level of tissue foundation of a TLA address and identify website; SLA is used for user group and sets up the home address hierarchical structure of oneself and identify subnet.
The address assignment of HSNET is to carry out address assignment for global unicast address under IPv6 address assignment framework, two kinds of methods are arranged: a kind of method is to repartition the global unicast address that converges of IPv6, redistribute by HSNET, set up the hierarchical structure address space of oneself, as shown in Figure 8.Fig. 8 is that the HSNET of compatible IPv6 can converge global unicast address architecture: (A) be Format Prefix FP, value is 001,3 bit; (B) be top level aggregation territory TLA, 13 bits; (C) be reserved field (RES), be preserved for expanding in the future TLA or HA territory, 8 bits; (D) be level address field HA, 32 bits; (E) be site level territory SLA, 8 bits; (F) be interface identifier territory Interface ID, 64 bits; (G) be the public address scope; (H) be the hierarchical topology address realm; (I) be website topological address scope.The address space that another kind method is to use IPv6 not assign is 010,011,100,101,110 etc. as the FP bit pattern, sets up the address space of level switching network, as Fig. 9.Fig. 9 is to use the not HSNET global unicast address architecture in assigned address space of IPv6: (A) be Format Prefix FP, 3 bits; (B) be level address field HA, 61 bits; (C) be site level territory SLA, 16 bits; (D) be interface identifier territory Interface ID, 64 bits; (E) be the hierarchical topology address realm; (F) be website topological address scope.The present invention advises adopting first kind of address scheme, so as with the direct intercommunication of existing IPv6 network.
The HSNET address distribution method adopts elongated exchange field, according to the address structure of the needs planned network of real network topological structure.The length of HSNET address layer time number and each layer exchange field is to be closely related with topology of networks, is independently determined oneself to exchange the length of field and is exchanged the logic number of the value of field as next straton net by each nodes domains.The length of exchange field has determined the maximum subdomain number of this nodes domains fan-out.Though the topological level of real network without limits, generally within 3 to 5 layers of scope, the exchange field number of address is also just in 3 to 5 scope.During exchange data packets, each nodes domains need be known the position and the length of respective quadrature transpose section.Concrete, HSNET level address space can be divided into a plurality of exchange field SF
i, each exchange field SF
iTwo parameter S are arranged
iAnd L
i, S wherein
iBe the length of subnet prefix, expression SF
iOriginal position in the IP address, L
iBe exchange field SF
iLength.There is special purposes the address of bit pattern complete 1, is used to indicate this section point domain addresses; Carrying out flexibly when the address of bit pattern complete 0 is used for lacking layer, level inserts; Whether adopt the exchange field of equal in length to need the complexity and the forward efficiency of taking into account system.
The HSNET address distribution method identifies nodes domains and internal node address thereof, comprises nodes domains and logic port address thereof, nodes domains internal node and port address thereof, nodes domains internal server address.Be that the design of nodes domains home address is netted certain i of branch node layer territory A with HSNET and the internal node address is an example, its form is FP:TLA:RES:SF1:SF2: ...: SF
i: ...: SLA:EHA:TYPE:X:Y: SF wherein
iBit pattern is complete 1; EHA is extendible level address, accounts for 4 bytes; Type TYPE field is the type of special address, accounts for 1 byte; X field accounts for 1 byte; 2 bytes of Y fields account.Whether address resolution method is " 0 " and different according to the bit pattern of EHA field entirely: if the bit pattern of EHA field is not complete " 0 ", then address resolution is a station server address of this nodes domains inside, and low 64 bits of this address are exactly the interface identifier number of this server EUI-64 form; The bit pattern of EHA field is complete " 0 " else if, and then resolve according to following mode the address:
(1) if TYPE=0, X, Y are 0, this address designation nodes domains A self address then;
(2) if TYPE=1, the Y span is 0 to 65535.When the Y value is 0 to 1023, then logical port number is the logic port of Y in this this nodes domains of address designation A; When the Y value is 1024 to 2047, then the short circuit logical port number is the port of Y in this this nodes domains of address designation A; When the Y value is 65535, the up logic port in this this nodes domains of address designation A then.
(3) if TYPE=2, then the exchange plane No. is the switch of X in this address designation nodes domains A.
(4) if TYPE=3, then the exchange plane No. is that X, physical port number are the physical port of Y in this address designation nodes domains A.
(5) if TYPE=4, X=0, Y value are 0 to 255, then the type in this address designation nodes domains A is the server of Y.
The address assignment of HSNET can be supported the insertion of level when the network layer vacancy.The level number of branching networks may be different in the HSNET network, and promptly nodes domains or user access network can be skipped some hierarchical direct (HD) and be connected with the ancestor node territory of high level.The level of skipping is the vacancy level, and the exchange field of vacancy level is made as 0.When this part address space will allow more user access network use, just can carry out the insertion of level, insert one or more level.In scarce layer branch, insert level, promptly increase nodes domains, set up its hierarchical network then.In the branch that lacks layer, can insert many levels, but all to guarantee to lack each original in floor branch nodes domains thereon the territory in node layer territory number be 0.Level inserts does not influence in the former scarce layer branch addresses distributed space.
4. packet level formula switching control method
The topological sum address of above-mentioned hierarchical switched network all is a hierarchical structure, and address structure and topological structure match each other, so the present invention is according to these features of hierarchical network, uses the hierarchy type exchange in that hierarchical switched network is inner.Because the level characteristics of HSNET network topology, so the logic switching path between any two nodes domains determines do not have the route of the overall situation to calculate; Because the strict coupling of the level characteristics of HSNET address and address and topology, the certain topology information of location, old place reflection, address itself has seeks the footpath function, exchange field accordingly according to the IP destination address, determine next jumping of logical path, and then, determine actual forward-path in conjunction with nodes domains configuration information and control procedure, finish the exchange of packet in nodes domains.In nodes domains, the exchange of packet level formula has three processes: determine logic port, select target outside port and select next to jump output port.The bag that arrives nodes domains from outside port first need be determined logic port, and the select target outside port is selected next to jump port then and is sent to next node; And the bag of port arrival internally only need carry out the operation of final stage, promptly selects next jumping according to the selected target external port in front, and hop-by-hop arrives the target external port.For last process, because node carries out need knowing target external output port information (being called inner exchanging mark ISL) when next jumps selection, therefore when packet enters this nodes domains first and finish the selection of target external port, need carry out the ISL encapsulation to bag.
Figure 13 is an ISL encapsulation IP packet format: wherein (A) is the FLAG field, 4 bits; (B) be reserved field RES, 4 bits; (C) be the EPID field, 24 bits; (D) be the IP bag; (E) be ISL.ISL is made up of three fields, is respectively FLAG, RES and EPID.The FLAG bit pattern is 1111 o'clock sign ISL encapsulation, and RES keeps, and the EPID field is inserted the target external output slogan EPID of bag, and EPID is made up of exchange plane No. SID (8) and port numbers PID (16).
Beneficial effect:
Hierarchical switched network of the present invention can be eliminated huge routing table, does not have route complicated, the overall situation to calculate, network control and management partial autonomy, thus improved network performance, simplified the complexity of core network.Hierarchical switched network is fit to the structure high-performance, can expands, can manage and reliable Internet of future generation, be particularly suitable for the IPv6 network, it follows IPv6 and related protocol (as the IPv6 address system structure) thereof, transmits scheme for the IPv6 network provides a kind of new bag.
Beneficial effect of the present invention comprises:
(1) key character of hierarchical switched network is that network topology is that address according to hierarchical structure structure, network also is to distribute and the address structure of network and the structure of network topology are mated according to hierarchical structure.The forwarding that this specific character makes Internet go up packet does not rely on the routing table routing, thereby does not need the Routing Information Exchange overall situation, complicated and route to calculate.Level exchange control procedure has partial autonomy's property, exchange of control information is confined to switching node inside, need not be in whole Internet communication control information, need not consider interoperability problem with other nodes domains, not only simplified the design of switch, and simplified Network Control Protocol, greatly reduce the complexity of procotol.The reduction of complexity is the key of structure high performance network.
(2) the hierarchy type exchange adopts the IP exchange to replace the IP route, each node layer exchanges the simple switched that field is carried out local control accordingly according to purpose IP address, do not rely on the huge routing table of inquiry, no longer have the implication that route is selected on the traditional sense, router has been degenerated to switch.Compare with label switched such as MPLS, the foundation of exchange is that the destination address of IP bag exchanges field accordingly, does not therefore need very long mark is distributed, manages, retrieves, and does not need to generate label switched path (LSP) for each stream.In the hierarchy type exchange, each nodes domains is only used the value of a very short field in the address, determine the whereabouts of next jumping, greatly simplified the complexity of switch, eliminate complicated route and calculated and searched the task of huge routing table, can finish main function of exchange with hardware, improve the performance of switch, alleviate the bottleneck problem of router greatly.
(3) hierarchical structure structure large scale network system has performance, scale, managerial extensibility.Along with Internet progressively becomes following global basis information transmission platform, number of users, network size increase by geometric progression.New application need provides the big data quantity real-time multimedia transmission service that QoS guarantees.But at present on structureless Internet, article one, channel is used by unknowable user group's piggybacking, add the sudden of communication, so flow rate mode of unpredictable network, can't reasonably be that every channel and corresponding ports thereof are distributed adequate resources quantity, the traffic engineering of network or QoS mechanism be difficult to carry out.This defective is more and more outstanding to the influence of network size expansion, network performance expansion, becomes the key factor of restriction network development.Hierarchical switched network is introduced the structuring notion in existing Internet, can be the flow restriction of the local communication in the subregion in subregion inside, do not take subregion Internet resources in addition, avoided at present the often phenomenon of the confusion of appearance, poor efficiency, waste Internet resources.The hierarchy type network has strict topological structure and relatively-stationary communication path, can simplify the model of describing the network dynamic change, reduces the complexity of network resource management and network traffics forecast of distribution, for the application of QoS mechanism provides basic assurance.Because the hierarchy type structure is with good expansibility and manageability, be fit to the network of effective structure and managing large scale simultaneously, can become the structure that following Internet adopts.When scale expansion, number of users or the user network of network increased, expansion only needed to increase the intermediate level and gets final product, and its influence is confined in the subrange, can not exert an influence to whole network.
(4) expansion of HSNET network logic-based nodes domains and logical links has increased the reliability of network and the extensibility of capacity.Expansion based on short circuit makes multi-level topology more flexible, has both logically kept the characteristics of hierarchical structure, has structure flexibly again on network is realized.In principle, the network configuration of any complexity of existing any connection under to its topological structure and channel connection state situation without any modification, can become the level switching network of band short circuit channel.
(5) compatibility of the design of HSNET and existing IPv6 and IPv4 network.The network of IPv6 can directly insert, so HSNET will help advancing IPv6.The design of HSNET makes existing network can carry out the transition to full HSNET structure in order, progressively.
Description of drawings:
The two-layer logical construction schematic diagram of Fig. 1 network
Local hierarchy's formula switching network schematic diagram of Fig. 2 logic-based nodes domains and logical links expansion
The example schematic diagram of any short circuit expansion of Fig. 3
Fig. 4 top layer extended method schematic diagram
Fig. 5 HSNET network and non-HSNET network interoperability schematic diagram
Fig. 6 Access Network inserts HSNET network topology example schematic based on many Su Fangfa
Fig. 7 IPv6 can converge global unicast address architecture
The HSNET of the compatible IPv6 of Fig. 8 can converge global unicast address architecture
Fig. 9 uses the not HSNET global unicast address architecture in assigned address space of IPv6
Figure 10 HSNET network and level address structure example schematic
Figure 11 HSNET network branches lacks layer address and represents example schematic
The example schematic that Figure 12 level inserts
Figure 13 ISL encapsulation IP packet format
Figure 14 level switching network embodiment schematic diagram
Figure 15 ISL encapsulation IP bag P form schematic diagram
Embodiment:
The present invention is further described below in conjunction with drawings and Examples.
Figure 10 is a HSNET network and level address structure example schematic.Wherein (1) is the HSNET network, (D
1)-(D
10) be nodes domains, (V)-(Z) be Access Network, (A) be Format Prefix FP, value is 001; (B) be top level aggregation territory TLA, 13 bits; (C) be reserved field RES, 8 bits; (SF
1)-(SF
10) be respectively nodes domains (D
1)-(D
10) the exchange field; (E) be site level territory SLA, 8 bits; (F) be interface identifier territory Interface ID, 64 bits; (G) be the public address scope; (H) be the hierarchical topology address realm; (I) be website topological address scope.Figure 10 (a) is a HSNET network diagram.Figure 10 (b) is (D of branch
1)-(D
2)-(D
4)-(D
7) level address structure schematic diagram, 32 level address fields (HA) are divided into four not isometric exchange fields, are respectively SF
1(24/8), SF
2(32/10), SF
4(42/6), SF
7(48/8).An Access Network that is attached to this HSNET network branches, host address is expressed as FP:TLA:RES:SF in the net
1: SF
2: SF
4: SF
7: SLA:EUI-64, wherein SF
1: SF
2: SF
4: SF
7For the level address of HSNET, manage by ISP; SLA is the address of user access networks, i.e. the website network No.; EUI-64 is 64 a host number.Figure 10 (c) is (D of branch
1)-(D
3)-(D
6)-(D
10) the level address structure, wherein 32 level address fields (HA) are divided into four isometric exchange fields, are respectively SF
1(24/8), SF
3(32/8), SF
6(40/8), SF
10(48/8).
Figure 11 is that the scarce layer address of HSNET network branches is represented example schematic.(D wherein
1)-(D
7) be nodes domains, (W)-(Z) be Access Network, (A) be Format Prefix FP, value is 001,3 bit; (B) be top level aggregation territory TLA, 13 bits; (C) be reserved field RES, 8 bits; (SF
1)-(SF
7) be respectively nodes domains (D
1)-(D
7) the exchange field; (M) be to lack layer field; (E) be site level territory SLA, 8 bits; (F) be interface identifier territory Interface ID, 64 bits; (G) be the public address scope; (H) be the hierarchical topology address realm; (I) be website topological address scope.Figure 11 (a) is a HSNET network diagram, and Backbone Server or Access Network are directly linked this HSNET net (D of branch
1)-(D
2)-(D
4)-(D
7) second layer nodes domains (D
2), lower floor's vacancy.Figure 11 (b) is (D of branch
1)-(D
2The level of)-(Y) address structure schematic diagram, wherein 32 level address fields (HA) are divided into two not isometric exchange fields, are respectively SF
1(24/8), SF
2(32/10), SF
2Value is 0; Lack layer field M (42/14), 14 bits are filled Binary Zero.
Figure 12 is the example schematic that a level inserts.(D wherein
1)-(D
9) be nodes domains, (V)-(Z) be Access Network, (A) be Format Prefix FP, value is 001,3 bit; (B) be top level aggregation territory TLA, 13 bits; (C) be reserved field RES, 8 bits; (SF
1)-(SF
9) be respectively nodes domains (D
1)-(D
9) the exchange field; (M) be to lack layer field; (E) be site level territory SLA, 8 bits; (F) be interface identifier territory Interface ID, 64 bits; (G) be the public address scope; (H) be the hierarchical topology address realm; (I) be website topological address scope.Figure 12 (a) is a HSNET network diagram, inserts nodes domains D in the scarce layer of branch of Figure 11 (a)
8(42/8), increased a level.Figure 12 (b) is (D of branch
1)-(D
2)-(D
8The level of)-(Y) address structure schematic diagram, wherein 32 level address fields (HA) are divided into three not isometric exchange fields, are respectively SF
1(24/8), SF
2(32/10), SF
8(42/8), SF
2And SF
8Value all is 0; Lack layer field M (50/6), 6 bits are filled Binary Zero; Wherein keeping the constant key in Access Network (Y) address, is the new nodes domains (D that inserts
8) subdomain number 0 distributed to the branch of (Y).Figure 12 (c) is (D of branch
1)-(D
2)-(D
8)-(D
9) level address structure schematic diagram, wherein 32 level address fields (HA) are divided into four not isometric exchange fields, are respectively SF
1(24/8), SF
2(32/10), SF
2(42/8), SF
9(50/6).
Figure 14 is the comprehensive embodiment schematic diagram of of level switching network.(D wherein
1), (D
2) ..., (D
10) be nodes domains; (A), (B), (C), (D) are switches; (E), (F), (G), (H), (I), (J) are user access networks; (1)-(6) is the port of switch (A), in switch (A), is numbered (1)-(6); (7)-(12) be the port of switch (B), in switch (B), be numbered (1)-(6); (13)-(16) be the port of switch (C), in switch (C), be numbered (1)-(4); (17)-(21) be the port of switch (D), in switch (D), be numbered (1)-(5); (22)-(37) be logical links; (38) be the short circuit link.
Figure 15 is the schematic diagram of ISL encapsulation IP bag P form.Wherein (A) is the FLAG field, and 4 bits are worth and are binary one 111; (B) be reserved field RES, 4 bits, value is 1111; (C) be the EPID field, 24 bits, preceding 8 bit patterns are 00000010, expression exchange plane No. 2, back 16 bit patterns are 0000000000000110, expression port numbers 6; (D) be IP bag P; (E) be ISL.
The embodiment of HSNET is described as a comprehensive embodiment with Figure 14 below.
In Figure 14, press the hierarchical structure tissue between nodes domains.As (D
1) be (D
2) and (D
3) father node, (D
2) be (D
4) and (D
5) father node.Interconnect with logical links between nodes domains, be and (D as (22), (23)
1) two continuous descending logical links, at (D
1) in numbering be respectively (13) and (22).Logical links number is specified by the father node territory: as (D
2) the up logical links that links to each other is numbered 65535 (because up logical links number is fixed, not marking it among the figure), (24), (25) they are two descending logical links, numbering is respectively (7) and (43); (26), (27) are (D
3) two descending logical links, numbering is respectively (0) and (19); (28) be (D
4) a descending logical links, be numbered (17); (29), (30) are (D
5) two descending logical links, numbering is respectively (10) and (26); (31) be (D
6) a descending logical links, be numbered (53; (32), (33) are (D
7) two descending logical links, numbering is respectively (11) and (23); (34), (35) are (D
8) two descending logical links, numbering is respectively (7) and (15); (36) be (D
9) a descending logical links, be numbered (14); (37) be (D
10) a descending logical links, be numbered (26).(38) be (D
5) and (D
6) between the short circuit logical links, this short circuit logical links is at nodes domains (D
5) in be numbered 1088, at (D
6) in be numbered 2018.Access Network (E), (G), (H), (J) directly are connected with the bottom layer node territory; Wherein Access Network (F) adopts many hosts access way, access node territory (D
7), (D
8); Direct and the HSNET upper level node territory (D of Access Network (I)
3) connect.
Provided nodes domains (D among Figure 14 in detail
5) internal configurations embodiment.(D
5) form by four switches (A), (B), (C), (D), numbering is respectively (1), (2), (3), (4), adopts complete interconnected topological structure.What be connected with switch (1) has 6 links, and its port is numbered in switch (1) and is respectively (1)-(6).Wherein the link that links to each other with port (1), (2), (6) is a peripheral link; The link that links to each other with port (1) is (D
5) in be numbered a physical link of 26 logical links; The link that links to each other with port (2) is (D
5) in be numbered a physical link of 10 logical links; The link that links to each other with port (6) is (D
5) in be numbered a physical link of 1088 short circuit link.The link that links to each other with port (3), (4), (5) is an inner link, the switch of inside, connected node territory.
In this embodiment, the address prefix of supposing the HSNET top mode is 00111111:11111111:11111111/24, then the following hierarchical switched network address structure of structure oneself in space, location at this point.According to the hierarchical switched network address distribution method, lower level node domain addresses prefix is the logical links number that upper layer node domain addresses prefix splicing upper layer node territory is assigned to the lower level node territory, and prefix length is that the prefix length in upper layer node territory adds upper layer node territory exchange field length sum.Wherein the address prefix table of nodes domains is as shown in table 2:
Address prefix | The exchange field | |
?D 1 | ?00111111:11111111:11111111/24 | ?24/8 |
?D 2 | ?00111111:11111111:11111111:00001101/32 | ?32/10 |
?D 3 | ?00111111:11111111:11111111:00010110/32 | ?32/8 |
?D 4 | ?00111111:11111111:11111111:00001101:00000001:11/42 | ?42/6 |
?D 5 | ?00111111:11111111:11111111:00001101:00001010:11/42 | ?42/8 |
?D 6 | ?00111111:11111111:11111111:00010110:00010011/40 | ?40/8 |
?D 7 | ?00111111:11111111:11111111:00001101:00000001:11010001/48 | ?48/8 |
?D 8 | ?00111111:11111111:11111111:00001101:00001010:11000010:10/50 | ?50/6 |
?D 9 | ?00111111:11111111:11111111:00001101:00001010:11000110:10/50 | ?50/6 |
?D 10 | ?00111111:11111111:11111111:00010110:00010011:00110101/48 | ?48/8 |
Table 2: the address prefix table of nodes domains
The Access Network address prefix is that its access node domain addresses prefix is spliced logical links that this nodes domains is assigned to Access Network number, and prefix length is that the prefix length of nodes domains adds nodes domains exchange field length sum.Access Network address prefix table is as shown in table 3:
Access Network | Address prefix |
E | ?00111111:11111111:11111111:00001101:00000001:11010001:00001011/56 |
F (host D 7) | ?00111111:11111111:11111111:00001101:00000001:11010001:00010111/56 |
F (host D 8) | ?00111111:11111111:11111111:00001101:00001010:11000010:10000111/56 |
G | ?00111111:11111111:11111111:00001101:00001010:11000010:10001111/56 |
H | ?00111111:11111111:11111111:00001101:00001010:11000110:10001110/56 |
I (lacking layer) | ?00111111:11111111:11111111:00010110:00000000:00000000:00000000/56 |
J | ?00111111:11111111:11111111:00010110:00010011:00110101:00011010/56 |
Table 3: the address prefix table of Access Network
The appointment of Access Network address of internal host is responsible for by user of access network, generally is made up of three parts: Access Network address prefix, SLA address (8) and interface number (64 are the EUI-64 form).
In comprehensive embodiment shown in Figure 14, the embodiment of a nodes domains basic layer formula exchanges data is: the main frame (M) in the Access Network (E) is communicated by letter with a main frame (N) of Access Network (H), mainly analyzes packet here at (D
5) exchanges data.
Suppose that at (E) site level domain addresses (M) is 7, interface number is EUI-64-M, then its address ip
MBe expressed as 00111111:11111111:11111111:00001101:00000001:11010001:00 001011:00000111:EUI-64-M; At (H), site level domain addresses address (N) is 6, and interface number is EUI-64-N, then its address ip
NBe expressed as 00111111:11111111:11111111:00001101:00001010:11000110:10 001110:00000110:EUI-64-N.Then to send to the process of main frame (N) as follows for a packet (P) of main frame (M):
(1) source address of P is filled to IP
M, destination address is filled to IP
N
(2) (E) interior router forwards it to (D
7), (D
7) the relatively address prefix of self and the destination address IP of P
NCorresponding part, the unequal up link of then selecting of result is sent to (D
4).
(3) at (D
4) through similar procedure, P arrives (D
2).
(4) (D
2) the relatively address prefix of self and the destination address IP of P
NCorresponding part, the result equates; According to (D
2) exchange field attribute 32/10 choose IP
NCounterpart, its value are 43, and then selecting down link number is 43 logical links, and P is sent to (D
5).
(5) supposition P arrives switch (3), at first determines its logical links; Promptly compare (D
5) address prefix of self and the destination address IP of P
NCorresponding part, the result equates; According to (D
5) exchange field attribute 42/8 choose IP
NCounterpart, its value are 26, and then selecting down link number is 26 logical links.
(6) search (D according to logical links numbers 26
5) the logical links table, be 1/1,2/6 to obtain two peripheral links according to outer end slogan (SID/PID); Based on the distributed-load balancing method that guarantees bag order in the stream, use the Hash function of appointment, the source and destination IP address computation hash value of input packet (P), search the outside port Hash table of logical links number 26 correspondences according to this hash value, obtain the target external port, the port numbers of supposing the peripheral link of selection is 2/6.Fig. 5 is the schematic diagram of ISL encapsulation IP bag P form.Wherein (A) is the FLAG field, and 4 bits are worth and are binary one 111; (B) be reserved field RES, 4 bits, value is 1111; (C) be the EPID field, 24 bits, preceding 8 bit patterns are 00000010, expression exchange plane No. 2, back 16 bit patterns are 0000000000000110, expression port numbers 6; (D) be IP bag P; (E) be ISL.
(7) switch (3) carries out the ISL encapsulation to packet, and encapsulation format as shown in Figure 5; The nodes domains that wraps in after the encapsulation is transmitted according to ISL label lookup inside, because target D_SID is 2, checking in the result is that port one is sent bag internally, arrives switch 2; After switch 2 is received bag, know and should send from the outside port 6 of controlling oneself, slough the ISL encapsulation, P is sent to (D
9);
(8) (D
9) will wrap P and send to (H), arrive main frame (N), finish the transmission of packet P from (M) to (N).
Claims (10)
1. hierarchical switched network system, by stipulating network topology, the network address and the relation that satisfies between the two again, replace the IP method for routing with the IP switching method, it is characterized in that: topology of networks is by the hierarchical structure structure, the address space of network also distributes by hierarchical structure, the strict coupling of the level of topological structure and the level of address structure; System comprises topological structure method for organizing, address distribution method and packet exchange control method; The wherein expansion of topological structure method for organizing logic-based nodes domains and logical links, substitute the node of tree with logical node, substitute the branch of tree with logical links, being basic element with logical node and logical links is organized in logic tree network by tree, makes hierarchical structure keep the tree characteristics; Address distribution method is set up the address space of level switching network, elongated exchange field is adopted in the address, each nodes domains freely determines to exchange accordingly separately field length, and the length of exchange field determines the maximum subdomain number of this nodes domains fan-out, the topology information of reflection level characteristics; Packet exchange control comprises definite logic port, select target outside port and selects next to jump three processes of output port, exchanges field accordingly according to destination address, determines next jumping of logical path, finishes the exchange of packet in nodes domains.
2. hierarchical switched network according to claim 1 system is characterized in that: by the expansion based on short circuit, increase direct link between different subtrees on the logic tree, make to have the second active path between subtree.
3. hierarchical switched network according to claim 2 system is characterized in that: the sight based on the short circuit expansion comprises locality short circuit and short circuit of overall importance; The locality short circuit directly connects the big flow nodes in communication of local needs territory, makes the flow between the different branched network that are in different link service provider's networks or same link service provider's network in this area more reasonable, plays the local optimum effect; Short circuit of overall importance directly connects two networks, directly sensible the other side's whole network.
4. hierarchical switched network according to claim 3 system is characterized in that: interconnected between hierarchy type network and other internet, and perhaps by the short circuit interconnect scheme, interconnected by the short circuit link between two network top nodes domains; Perhaps set up a super dummy node territory as root node,, construct new wider hierarchy type network child node territory, the top mode territory of formula network at all levels as this super node territory.
5. hierarchical switched network according to claim 1 system, it is characterized in that: hierarchical switched network and other non-hierarchical switched networks are interconnected, and perhaps running boundary gateway protocol BGP sets up peer-to-peer between two networks; Perhaps with the lower floor network of hierarchical switched network, need not running boundary gateway protocol BGP and use the route of acquiescence as non-hierarchy type backbone network.
6. hierarchical switched network according to claim 1 system, it is characterized in that: the interconnected pattern of hierarchical switched network and Access Network has three classes, comprises and the interconnected and terminal extended method in directly interconnected, the many place of current Access Network; Wherein directly interconnected method is by the direct access level formula of router switching network edge port with user network; Many places interconnected method is that a user network is connected on different link service providers or same link service provider's the diverse access point; Terminal extended method be with Access Network itself also according to the hierarchical switched network structure, when inserting and hierarchical switched network carry out seamless interconnected.
7. hierarchical switched network according to claim 1 system, it is characterized in that: the hierarchical switched network address distribution method is that the global unicast address space of converging of IPv6 is repartitioned, redistribute by hierarchical switched network, set up the hierarchical structure address space of oneself; Wherein the corresponding exchange of the length level number of each layer exchange field adopts elongated exchange field, by the length of each nodes domains decision oneself exchange field.
8. hierarchical switched network according to claim 7 system, it is characterized in that: in the hierarchical switched network address assignment, nodes domains and internal node address thereof are identified, comprise nodes domains and logic port address thereof, nodes domains internal node and port address thereof, nodes domains internal server address; Address resolution resolves to corresponding server address, switch address, logic port or physical port according to the content of its extendible level address and type field.
9. hierarchical switched network according to claim 8 system is characterized in that: when the vacancy level occurring when nodes domains is skipped some hierarchical direct (HD) and high-rise ancestor node territory is connected, the exchange field of vacancy level is made as 0; Can insert level lacking in the layer branch, promptly increase nodes domains, but guarantee to lack each original in floor branch nodes domains thereon the territory in node layer territory number be 0.
10. hierarchical switched network according to claim 1 system, it is characterized in that: the hierarchical switched network exchange data packets comprises three steps, determines logic port, select target outside port process and packet is encapsulated and inner routing; Wherein each nodes domains determines that logic port passes through the position and the length of packet exchange field, determines that this packet arrives the logical links that the next node territory need be passed through; Select target outside port process is selected an outside output link according to load balance mechanism from the selected peripheral link that logical links comprised; When packet enters this nodes domains first and finish the selection of target external port, need carry out the encapsulation of inner exchanging mark to bag; When selecting next to jump output port, then directly output to the known target external output port later on.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB03120998XA CN100444573C (en) | 2003-03-27 | 2003-03-27 | Hierarchical exchange network system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB03120998XA CN100444573C (en) | 2003-03-27 | 2003-03-27 | Hierarchical exchange network system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1514605A true CN1514605A (en) | 2004-07-21 |
CN100444573C CN100444573C (en) | 2008-12-17 |
Family
ID=34239519
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB03120998XA Expired - Fee Related CN100444573C (en) | 2003-03-27 | 2003-03-27 | Hierarchical exchange network system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100444573C (en) |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1333564C (en) * | 2004-12-15 | 2007-08-22 | 中国科学院计算机网络信息中心 | Short circuit communication method of layered type switching network |
WO2007147328A1 (en) * | 2006-06-16 | 2007-12-27 | Beijing Jiaotong University | A method of vector network address coding |
CN100421407C (en) * | 2005-11-22 | 2008-09-24 | 中国科学院计算机网络信息中心 | Separating and merging IPv6 address space of switching network in hierarchy mode |
CN100428736C (en) * | 2005-02-02 | 2008-10-22 | 华为技术有限公司 | Topology method of one-time route computing to realize hierarchical route |
CN100449997C (en) * | 2006-06-02 | 2009-01-07 | 华为技术有限公司 | Netting communication system and its node address distribution method |
CN101841573A (en) * | 2010-01-20 | 2010-09-22 | 中国科学院计算机网络信息中心 | Method and device for processing address information of Internet and Internet system |
CN101521613B (en) * | 2008-12-31 | 2011-12-14 | 张建国 | Method for designing optical fiber communication network structure, and extensible network structure |
CN101635645B (en) * | 2009-08-28 | 2012-01-11 | 中兴通讯股份有限公司 | Method and device for managing network topology |
CN101626343B (en) * | 2009-08-05 | 2012-04-04 | 华为技术有限公司 | Method and apparatus for exchanging data packet, and communication device |
CN103124237A (en) * | 2011-11-18 | 2013-05-29 | 中国科学院沈阳计算技术研究所有限公司 | Addressing method used for MAC (Media Access Control) layer of switching device of star-shaped industrial Ethernet |
CN103269387A (en) * | 2013-05-21 | 2013-08-28 | 中国科学院计算技术研究所 | Automatic configuration method for data center network addresses |
CN103391254A (en) * | 2012-05-09 | 2013-11-13 | 百度在线网络技术(北京)有限公司 | Method and device for managing distributed CDN flows |
CN104079491A (en) * | 2014-07-07 | 2014-10-01 | 中国科学院计算技术研究所 | Router for high-dimensionality network and routing method |
CN102811167B (en) * | 2011-06-02 | 2017-11-21 | 三星电子株式会社 | Method and apparatus for the network based on hierarchical name structure |
CN108063686A (en) * | 2017-12-12 | 2018-05-22 | 长春理工大学 | A kind of data center network topology structure and construction method based on cube structure |
CN109729107A (en) * | 2017-10-27 | 2019-05-07 | 北京乾唐视联网络科技有限公司 | The method of network entry and relevant device of autonomous cloud in a kind of autonomous networks |
CN112615851A (en) * | 2020-12-16 | 2021-04-06 | 盐城工学院 | Boundary router combining multiple safety inspection mechanisms under CoLoR architecture |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3833466B2 (en) * | 2000-11-08 | 2006-10-11 | 三菱電機株式会社 | Network system and network control method |
JP2002271374A (en) * | 2001-12-28 | 2002-09-20 | Sony Corp | Information processing device and method therefor |
-
2003
- 2003-03-27 CN CNB03120998XA patent/CN100444573C/en not_active Expired - Fee Related
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1333564C (en) * | 2004-12-15 | 2007-08-22 | 中国科学院计算机网络信息中心 | Short circuit communication method of layered type switching network |
CN100428736C (en) * | 2005-02-02 | 2008-10-22 | 华为技术有限公司 | Topology method of one-time route computing to realize hierarchical route |
CN100421407C (en) * | 2005-11-22 | 2008-09-24 | 中国科学院计算机网络信息中心 | Separating and merging IPv6 address space of switching network in hierarchy mode |
CN100449997C (en) * | 2006-06-02 | 2009-01-07 | 华为技术有限公司 | Netting communication system and its node address distribution method |
WO2007147328A1 (en) * | 2006-06-16 | 2007-12-27 | Beijing Jiaotong University | A method of vector network address coding |
US7958265B2 (en) | 2006-06-16 | 2011-06-07 | Beijing Jiaotong University | Coding method of vector network address |
CN101521613B (en) * | 2008-12-31 | 2011-12-14 | 张建国 | Method for designing optical fiber communication network structure, and extensible network structure |
CN101626343B (en) * | 2009-08-05 | 2012-04-04 | 华为技术有限公司 | Method and apparatus for exchanging data packet, and communication device |
CN101635645B (en) * | 2009-08-28 | 2012-01-11 | 中兴通讯股份有限公司 | Method and device for managing network topology |
CN101841573B (en) * | 2010-01-20 | 2013-08-07 | 中国科学院计算机网络信息中心 | Method and device for processing address information of Internet and Internet system |
CN101841573A (en) * | 2010-01-20 | 2010-09-22 | 中国科学院计算机网络信息中心 | Method and device for processing address information of Internet and Internet system |
WO2011088657A1 (en) * | 2010-01-20 | 2011-07-28 | 中国科学院计算机网络信息中心 | Method, device and internet system for processing internet address information |
US8984112B2 (en) | 2010-01-20 | 2015-03-17 | Computer Network Information Centre, Chinese Academy Of Sciences | Internet address information processing method, apparatus, and internet system |
CN102811167B (en) * | 2011-06-02 | 2017-11-21 | 三星电子株式会社 | Method and apparatus for the network based on hierarchical name structure |
CN103124237A (en) * | 2011-11-18 | 2013-05-29 | 中国科学院沈阳计算技术研究所有限公司 | Addressing method used for MAC (Media Access Control) layer of switching device of star-shaped industrial Ethernet |
CN103391254B (en) * | 2012-05-09 | 2016-07-27 | 百度在线网络技术(北京)有限公司 | Flow managing method and device for Distributed C DN |
CN103391254A (en) * | 2012-05-09 | 2013-11-13 | 百度在线网络技术(北京)有限公司 | Method and device for managing distributed CDN flows |
CN103269387A (en) * | 2013-05-21 | 2013-08-28 | 中国科学院计算技术研究所 | Automatic configuration method for data center network addresses |
CN104079491A (en) * | 2014-07-07 | 2014-10-01 | 中国科学院计算技术研究所 | Router for high-dimensionality network and routing method |
CN104079491B (en) * | 2014-07-07 | 2018-04-27 | 中国科学院计算技术研究所 | A kind of router and method for routing towards high-dimensional network |
CN109729107A (en) * | 2017-10-27 | 2019-05-07 | 北京乾唐视联网络科技有限公司 | The method of network entry and relevant device of autonomous cloud in a kind of autonomous networks |
CN109729107B (en) * | 2017-10-27 | 2021-10-19 | 北京乾唐视联网络科技有限公司 | Network access method of autonomous cloud in autonomous network and related equipment |
CN108063686A (en) * | 2017-12-12 | 2018-05-22 | 长春理工大学 | A kind of data center network topology structure and construction method based on cube structure |
CN108063686B (en) * | 2017-12-12 | 2023-05-16 | 长春理工大学 | Data center network topology structure based on cube structure and construction method |
CN112615851A (en) * | 2020-12-16 | 2021-04-06 | 盐城工学院 | Boundary router combining multiple safety inspection mechanisms under CoLoR architecture |
Also Published As
Publication number | Publication date |
---|---|
CN100444573C (en) | 2008-12-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1216472C (en) | Control method for layered switching network node region | |
CN1514605A (en) | Hierarchical exchange network system | |
CN102150148B (en) | Differentiated services for unicast multicast frames in layer 2 topologies | |
CN101459698B (en) | Intra-domain and inter-domain network interconnecting method and system thereof | |
US9148300B2 (en) | Method and system for telecommunications including self-organizing scalable Ethernet using IS-IS hierarchy | |
CN102037685B (en) | IP forwarding across a link state protocol controlled Ethernet network | |
CN104270475B (en) | The system and method for IPv4 networks and IPv6 network interworkings is realized based on NAT64 | |
Jain et al. | Viro: A scalable, robust and namespace independent virtual id routing for future networks | |
CN101155115B (en) | Method and apparatus for implementing automatic tunnel | |
CN105850102A (en) | Control of a chain of services | |
CN102075438B (en) | unicast data frame transmission method and device | |
CN101960785A (en) | Implementation of VPN over a link state protocol controlled Ethernet network | |
CN101114971A (en) | Method for implementing virtual private network based on IPv6 address structure | |
CN1513253A (en) | Tunneling through access network | |
CN102421162B (en) | Hybrid router supporting IPv6 (Internet Protocol version 6) wireless sensor network | |
CN114095305A (en) | BIER message forwarding method, equipment and system | |
CN101047636A (en) | Method and system for end-to-end pseudo-line simulation virtual leased line access virtual special network | |
CN100452734C (en) | Global Internet topology knowledge-based P2P application construction method | |
Rosenberg | A primer of multicast routing | |
CN105681438A (en) | Centralized caching decision strategy in content-centric networking | |
CN100428681C (en) | Method for identifying multicast by using unicast address in IP network | |
US20050053014A1 (en) | System and method of implementing contacts of small worlds in packet communication networks | |
CN102098353B (en) | System and method for interworking between IPv4 (internet protocol version 4) and IPv6 (internet protocol version 6) based on DHT (distributed hash table) | |
CN107241457B (en) | Method for realizing network end-to-end communication | |
Francis | Addressing in internetwork protocols |
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: 20081217 Termination date: 20160327 |