CN1234159A

CN1234159A - Packet routing in telecommunications network

Info

Publication number: CN1234159A
Application number: CN 97199115
Authority: CN
Inventors: 奥缇·玛吉
Original assignee: Nokia Telecommunications Oy
Current assignee: Nokia Oyj
Priority date: 1996-09-24
Filing date: 1997-09-24
Publication date: 1999-11-03
Also published as: AU4387197A; CA2266390A1; AU728588B2; BR9711412A; EP0927478A1; FI104670B; FI964268A0; FI964268A; WO1998013975A1; NZ334771A

Abstract

The invention relates to a method for packet transfer in a telecommunications network. In a node(N1)at the boundary of the network, the data packets to be transmitted are segmented into lower protocol layer data units. The data units are transferred in the network both in default channels and in dedicated channels. For efficient routing of the data packets, the node(N1)at the boundary of the network monitors packets that pertain to the same flow, and upon detecting that a given flow meets predetermined criteria, a dedicated lower protocol layer connection identifier is assigned for its use. In a node(N2)within the network, the connection identifiers of incoming data packets are monitored, and upon detection of a new connection identifier that is not associated with a default channel or a dedicated channel, an output port and a new outgoing lower protocol layer connection identifier are defined for the data unit on the basis of the destination address of that packet, and the data unit is routed to the relevant output port. Upon detection of the same lower protocol layer connection identifier as said new connection identifier in subsequently arriving data units, routing is performed in the node(N2)merely on the basis of the lower protocol layer connection identifier. Dedicated connections are discarded from use in the nodes after no traffic in a flow having a dedicated connection identifier has been detected during a given period of time.

Description

Grouping Route Selection in the telecommunications network

The grouping Route Selection that relate generally to of the present invention carries out in the packet-switched telecommunications net relates in particular to through ATM net transmission IP grouping (IP, IP(Internet Protocol)).

Mainly due to popularizing of internet, IP is current most popular network layer (in the osi model the 3rd a layer) agreement.Along with the Exponential growth of the main frame that is connected with the internet, the throughput of IP network has become bottleneck, thereby needs a kind of ratio to transmit the method for IP operation at present quickly.

Fig. 1 illustrates the typical structure of IP network.Under working environment, personal computer PC or similar terminal equipment and the local area network (LAN) LAN1 that typically is Ethernet ... LAN3 connects, local area network (LAN) then and backbone network (WAN, wide area network) interconnected, comprising router (RT1 based on net as node ... RT6).All computers that are arranged in same local area network (LAN) have identical IP network address.When certain computer that is connected with local area network (LAN) sends packet, whether the IP layer check IP destination-address of the protocol stack of this sending computer is identical with the IP network address of himself.If the address is identical, do not need router, only this grouping is sent on the computer with destination-address along this regional network.If the IP network address of the IP network address of destination and sending computer is different, computer sends to certain router to grouping, and router transports the packet to other network again.

Transmissions links between the router can realize by PDH or SDH technology or with Packet Based Network technology (for example ATM, frame relay, X.25).

Router has two major functions: grouping transmits and the renewal routing table.Carry out by this way on the grouping transport process principle: router is at first read the network address of destination from entering the IP grouping that comes, and searches the output port relevant with this address then and by described port this grouping is sent to next router from its routing table.This grouping sends to another router from a router, and till certain router found that destination-address is identical with its network address, this router sent to destination host to this grouping in this case.

Along with the raising of transmission rate request, introduced new technology.As a kind of backbone network technology, the level of application of atm technology constantly increases.In this case, in the routing nodes of ATM net, making up interface; At first,, and again packet decomposition is become to be used for to net the ATM cell that transmits at ATM afterwards to a minute group selection route from rebuilding grouping from the cell of ATM net.Standard A TM adaptation layer (AAL) is carried out the decomposition and the reconstruction of IP grouping.To be described in more detail this below, so that provide background for explanation of the present invention.

When certain work station in another local area network (LAN) of certain work station in the above-mentioned ethernet local area network sends data, the data division P1 that is formed by the application program of this work station at first is packaged into TCP grouping P2, (agreement that the supposition transport layer adopts is TCP, transmission control protocol) for example as shown in Figure 2.Then the TCP packet encapsulation is become IP grouping P3, and and then the IP packet encapsulation become Ether frame P4, Ether frame P4 is sent to a router that is connected with local area network (LAN) and have the interface of netting towards ATM along local area network (LAN).This router removes the ether section and the IP packet decomposition is become ATM cell in the ATM Adaptation Layer.Note that encapsulation can be included in the front of grouping and the insertion of back (so-called afterbody).

Fig. 3 illustrates the structure of an IP grouping 30 (being IP datagram).The minimum length of packet header is 20 bytes, and " word " that it is divided into five nybbles is present in the adjacent lines among the figure.The length of field is shown with figure place in bracket among this figure in each field behind the name.Title at first comprises 4 version field 31, represents employed IP version.Then be length field 32 (IHL, internet length for heading), the expression length for heading is the word of 4 bytes.Type field 33 indicating services types, field 34 is the total lengths that comprising the datagram of title.Identification field 35 is used for the reconstruction sign IP grouping together with grouping.Tag field 36 makes system infer whether certain fragment of the grouping after decomposing is last fragment of original packet.Fragment departs from the position of this fragment of content representation in original packet of field 37.This is grouped in the maximum duration that may exist in the network life time field 38 expressions.Push away from the value of this field and to unravel silk each router that this grouping is gone through.The upper-layer protocol (for example TCP) of the data that protocol fields 39 these datagrams of expression are carried.Field 40 comprise header checking and.

Field

41 and 42 is source address and destination-address, that is, and and the address that they are represented to send main frame and receive main frame with the form of 32 bit address.It after the address field option field that seldom uses.The data that transmit in this field are usually with network test or to search fault relevant; These data for example define certain given route that this datagram should be passed through.This field has the filling hyte in case of necessity, thereby byte quantity can divide exactly by four.

After above-mentioned title, be the real data in the IP grouping.The length of data field is variable, but its upper limit delimited by the length of field 34, and the maximum length that this means whole group is 2 ¹⁶Byte.

As previously described, in the ATM of foundation Fig. 1 net, transmit the IP grouping with the ATM cell form.Fig. 4 a represents the basic structure of a cell being transmitted in the ATM net.A plurality of cells that will transmit in the network comprise the pay(useful) load of one 48 byte and the title of one 5 byte, but the definite structure of title (content of title) depends on that part of using the ATM net under every kind of situation, depend on related interface (that is which part of network) because comprise the header syntax that adopts in careful interface of stipulating of some various standards and the ATM cell in the ATM network architecture.

Fig. 4 b illustrates the cell header syntax that the uni interface (user is to network interface) of ATM net is located, and this interface is the interface between atm end device and the ATM node.Fig. 4 C illustrates the cell header syntax that the NNI interface (network-network interface) of atm network is located, and this interface is the interface between two ATM nodes (or in a network or between two networks).

The Route Selection field of cell title comprises virtual route identifier (VPI) and virtual path identifier (VCI).In the mark header syntax that only is used for user terminal shown in Fig. 4 b, for Route Selection field (VPI/VCI) keeps 24 altogether.Header syntax at the other parts place that is used for ATM shown in Fig. 4 C is for Route Selection field (VPI/VCI) keeps 28.Mean the Route Selection field is served as the route cell in the ATM net basis as its name.Virtual route identifier VPI is mainly used in the interior section of network, and in fact it often determines that cell is with that physical connection of route.On the other hand, virtual path identifier VCI only is used for the Route Selection on the network boundary usually.Yet note that VPI and the VCI route of regulation cell clearly together.

Other field in the standard in the defined ATM cell title is:

-GFC (class current control) is used in the professional field of user terminal supervision, not accurately definition as yet,

-PTI (pay(useful) load type indicator) is mainly used in and distinguishes network management cell and user profile cell, might basis whether detect congested differentiation user profile cell on route in addition.

-CLP (cell loss priority) is used for determining the cell priority DE position of frame relay network (mostly corresponding to) when relating to the cell loss concealment possibility,

-HEC (header error control), header checking and.

In these other fields, mainly be that pti field is relevant with the present invention, might utilize the boundary between this field monitor packets.Last position (Fig. 4 a in the pti field ... item 2 among the 4c) when indication begins a new high layer packets (IP grouping).When this position is become 1, relate to last cell of certain IP grouping, and in this case, next grouping begins with the busy cell of the next one.

Usually, the signal of various forms arrives ATM Adaptation Layer, the task of ATM Adaptation Layer is that these signals were formed the required reference format of ATM net on the one hand before by the ATM net, and on the other hand signal relay is being rebuild the cell signal of netting from ATM before user interface or control interface.To different service class (A ... D) standardization dissimilar adaptation layer (AAL1 ... AAL5).For example, there is not the temporal correlation between the source and destination in AAL type 3,4,5 therein for application program provides the transmission service.

Fig. 5 illustrates the IP packet fragmentation being become ATM cell and rebuilding grouping from ATM cell of AAL5 execution, so that the operation of the ATM Adaptation Layer in the example network shown in the exemplary view 1.ATM Adaptation Layer generally is divided into two sublayers, and they are abbreviated as SAR (cut apart and ressemble sublayer) and CS (polymerization sublayer).Sublayer, back is carried out the encapsulation of user data part (for example IP grouping) and control data and is gone encapsulation.Carry out the frame that encapsulates and produce by the CS sublayer and be called CS-PDU (polymerization sublayer protocol Data Unit).The AAL5 encapsulation is carried out by this way, promptly increases an afterbody to user data cell (for example, the IP grouping), for example comprises an error checking part (CRC).The length of this afterbody is 8 bytes.The multiple correspondence of the length of whole C S-PDU and 48 bytes, this is that suitably to add length between the pay(useful) load of afterbody and grouping when needing be 0 ... the filling field PAD of 47 bytes obtains.

The SAR sublayer is divided into each CS-PDU of sending direction the section that calls SAR-PDU (cut apart and ressemble protocol Data Unit) of 48 byte longs.On receive direction, SAR-PDU forms CS-PDU by assembling.

ATM layer below the ATM Adaptation Layer is responsible for the header field CH of five bytes (Fig. 4 b and 4c) is added among the SAR-PDU for transmission, thereby generates the ATM cell 50 that sends to the ATM net.The ATM net is only handled the cell title, the pay(useful) load of not handling and even not reading 48 bytes in the ATM net.On receive direction, the ATM layer removes title from cell, and the pay(useful) load of 48 bytes is offered the SAR sublayer so that assembling.

In the network of Fig. 1, router sends the IP grouping, in a manner described packet fragmentation is become cell and grouping is sent to the ATM transmission link.The router at link other end place is rebuild grouping with cell in a manner described, makes route in a usual manner according to the IP address and judges, and once more packet fragmentation is become cell for next atm link.

Route judgement in the grouping is normally finished by software.When Route Selection that realizes with software and above-mentioned packet fragmentation with when rebuilding combination, obviously slow down based on the operation of the router network of ATM.In addition, realize that this conventional route selection method is expensive.

For alleviating these shortcomings, developed a kind of method of the IP of being called exchange.IP exchanges based on stream concept: stream (usually) is a string IP grouping to same destination from same source.Thereby a stream (usually) comprises source address and one group of identical IP grouping of destination-address.For example, it is a stream that TCP connects: when opening this TCP connection, a string grouping sends to the destination from the source.In IP exchange, router node is determined each stream, and the router that requires the network boundary place provides a unique flow identifier, for example unique VPI/VCI identifier for the grouping of every stream.When unique VPI/VCI identifier being provided for each relevant grouping of constant current with certain, the router in the network can utilize conventional ATM exchange to realize packet relay on the cell level.Thereby Route Selection does not need to enter into the 3rd layer (IP layer), but can carry out in the second layer (ATM layer).

But the shortcoming of the prior art solution of just having mentioned is, its prerequisite is an independently control protocol, by the negotiation of the node on node in this protocol network and the professional Incoming direction so that to the VPI/VCI identifier of given flow distribution special use.This solution makes network be still complicated and causes the supplementary service that network is loaded.

An object of the present invention is to remove above-mentioned each shortcoming, and provide any new stream protocol of a kind of unnecessary application to accelerate the solution of grouping Route Selection down.

This purpose is to reach by the solution that defines in claims.

Thought of the present invention is the stream (the reception grouping of higher protocol layer) that is formed by grouping on the monitoring network boundary node.When detecting certain, to the low protocol layer connection identifier (CID of a special use of this flow distribution to the satisfied special use connection of constant current criterion.When the next node that makes progress when downstream side receives and has the data cell of " newly " connection identifier (CID, its clear immediately stream of specifying dedicated channel that connected.This node is carried out conventional Route Selection according to the IP destination-address in first grouping, but the follow-up data cell with identical connection identifier (CID is only forwarded according to low protocol layer connection identifier (CID.In addition, node has out-of-date control, and after preset time did not receive the business relevant with related streams in the cycle, this control stopped the use that this special use connects.

In most of preferred embodiments of the present invention, according to above-mentioned, low layer data unit is that ATM cell and high layer packets are the IP groupings, but this method also can realize together with other agreement or transmission method.

Owing to according to solution of the present invention, can increase the throughput of node, no longer need carry out packet fragmentation and reconstruction at the node place of network.In addition, it is not needing the mode of extra stream protocol to be issued in network, thereby it remains more simply router node and does not produce supplementary service in the network.

By according to the example in the accompanying drawing, describe the present invention and its preferred embodiment in detail below with reference to Fig. 6-11g, accompanying drawing is:

Fig. 1 illustrates environment for use of the present invention,

Fig. 2 forms grouping before illustrating in sending to network,

Fig. 3 illustrates the structure of IP grouping,

Fig. 4 a ... the general structure of 4c explanation ATM cell,

What Fig. 5 illustrated that ATM Adaptation Layer 5 carries out becomes ATM cell to packet fragmentation, and rebuilds grouping from ATM cell,

The operation of Fig. 6 explanation method of the present invention on three successive nodes of network,

Fig. 7 is a flow chart, is illustrated in the operation of the inventive method on the first node of Fig. 6,

Fig. 8 is a flow chart, is illustrated in the operation of the inventive method on second, third node of Fig. 6,

Fig. 9 illustrates the structure of gateway node,

Figure 10 illustrates the structure of network node, and

Figure 11 a ... the table that the node of 11g explanation Figure 10 uses.

In solution according to the present invention, locate to carry out cutting apart of IP grouping by above-mentioned node (RT1 among Fig. 1 or RT4), and ATM cell is sent to the ATM transmission link at network boundary.

Fig. 6 illustrates ATM net node N1 on three routes that one after the other are positioned at grouping and that realize the inventive method ... N3.In the figure, the default channel between the node indicates with extra heavy pipe, and the dedicated channel between the node indicates with tubule.The different step circles mark of this method.

System includes only default channel at first, and system is as required to connecting assigned with dedicated channel when beginning is professional.Default channel is extended from node along all directions, the grouping with different source addresses and/or destination-address can advance along a default channel (even they have identical VPI/VCI identifier).Here, default channel is to be ATM usefulness the channel of transmission path only, and realizes Route Selection at high-rise (IP layer) in a usual manner therein.For this reason, the business in the default channel be system attempt to eliminate do not wish business.But, must use default channel, because the major part in all business continues very short (may only send several groupings in the TCP connection) in the reality, thereby do not need they are distributed special-purpose VPI/VCI identifier.Only be worthwhile for and continue to set up the particular virtual connection than long business.

Locate only to arrive at the IP grouping at first node (N1), packet fragmentation is become cell, and cell relay is arrived node N2 in mode recited above.Node N1 determines the header field (source address and destination-address and may also have other field) of the required usefulness of Incoming IP grouping in known mode.When node N1 detects (circle 1) certain given IP when connecting (having the identical source address and/or the grouping of destination-address) and satisfying some and connect (osi layer 2) pre-determined criteria for the special-purpose second layer, node N1 decision specifies the second layer of a special use to connect (promptly to this stream, connection identifier (CID), this is a new VPI/VCI identifier under the situation of ATM net.

For example can there be conventional or heavy business (classified counting of node to arriving) in node or makes above-mentioned decision according to detecting a certain amount of business of IP connection needs according to detecting in this IP connection.Node for example may find to relate to the FTP transmission that needs fastext.Above-mentioned criterion can be large-scale all kinds.

When above-mentioned decision posterior nodal point N1 detected first grouping, it selected a spendable second layer connection identifier (CID (VPI/VCI identifier) according to the IP title of this grouping.

For example, connect when node detects the special-purpose second layer of this connection request at the tenth fen group place, node N1 uses the cell of its decomposition from described grouping beginning and belongs to the identifier that the special-purpose second layer connects.Preceding nine groupings have transmitted (with the form of cell) in default channel.In default channel with each nonoverlapping mode of dividing into groups send continuously each the grouping in cell.

After this, the next node (N2) in downstream has received and has carried the cell of this node for " new " (being used by any other stream) VPI/VCI identifier, thereby the node that clear (circle 3) this stream relates to a updrift side has distributed the stream of special-purpose second layer connection identifier (CID to it.This node is read the IP destination-address from this cell, and as can be as seen from Figure 3, the IP destination-address always makes on the fixed position of its payload bytes 13-17 that is in cell when cutting apart.According to the destination-address of reading, node is searched the identifier of the output port relevant with this address from routing table, and cell is sent to this output port.Then, all cells with identical VPI/VCI identifier all send to identical output port.Node only need be checked first IP address, and carries out conventional Route Selection (at the IP layer) according to this address.After this, it can directly carry out exchange on the ATM layer according to the VPI/VCI identifier, and utilizes new VPI/VCI identifier that cell is then sent to next node (circle 4).

Fig. 7 illustrates the flow chart of the operation of explanation network boundary node (gateway node N1).In different steps, illustrate among this figure in the bracket with reference to T _n(n=1 ... 5); Each table that on behalf of this node, they use, these tables are described in Fig. 9.

At first, node receives an IP grouping, reads source address and destination-address (step 71) at least from its title, and is that this address is to (or only being destination-address) update packet counting (step 72, table T1).

Then, node takes a decision as to whether this address to setting special-purpose connect (step 73).This is by checking in this table whether (table T1) sets with this address certain relevant dedicated channel is finished, in other words, whether to this flow distribution special-purpose second layer connection identifier (CID.If this situation is searched and the Route Selection mark of this join dependency from routing table (table T2); This mark is the internal identifier of atm switching matrix, relies on the exchange (step 74) of its atm switching fabric (fabric) execution to correct output port.

If still unallocated special-purpose the connection, next step is to judge whether this stream satisfies the special-purpose criterion (step 75) that connects, in other words, for example, measure the phase traffic carrying capacity at certain and whether surpass certain set-point, perhaps whether professional quality requirement is certain given type (that is, needing the high-throughput possibility) in this stream.Node for example can use the connection type field 33 (Fig. 3) by indication service quality such as relaying, throughput and reliabilities.If do not satisfy these criterions, search the Route Selection mark (step 74, table T2) of the default channel of relevant IP destination-address, and carry out Route Selection accordingly.

If in step 75, find to satisfy the special-purpose criterion that connects, then search with relevant address to or the relevant output port (table T2) of destination-address, and search and of this output port correspondence can use Route Selection mark (table T3), i.e. still unappropriated and and the Route Selection mark of this IP address correspondence.Then, set certain dedicated channel (table T1), and in table T2, upgrade the Route Selection mark of this connection corresponding to this IP address.At the output port place, be spendable VPI/VCI of this new Route Selection label lookup (table T5) and renewal Route Selection mark conversion table (table T4).

After this, this method enters step 78, therein the IP packet fragmentation is become cell, and above retrieve the Route Selection mark and be attached on these cells.Then these cells are sent to the output port (step 79) relative with this Route Selection mark.Before cell is sent to next link, substitute Route Selection mark (table T4) with corresponding VPI/VCI identifier.

This method enters and receives next IP grouping then.

In addition, node has an independently aging logical circuit, and preset time did not receive the business relevant with this specific stream in the cycle after, this circuit did not re-use special-purpose the connection and relevant connection data.Be preferably in all node places and adopt similarly mechanism, and start the deletion that special use is connected from start node.Still have business if node has been discarded special-purpose connection in this stream, the explanation of above then being similar to " first " being divided into groups is carried out route and is judged.

Fig. 8 illustrates the operation of the node (for example node N2 and N3) in the network.Among this figure, shown in the bracket in different step with reference to S _n(n=1 ... 7), their representation nodes use and at Figure 10, Figure 11 a of back ... the table of describing among the 11g.

Node receives certain cell (step 80) at first and reads its title (step 81).Upgrade the service counting relevant (step 82, table S1) then with this VPI/VCI identifier.Next step, node judges whether relate to default channel, that is, enter the VPI/VCI identifier and whether indicate into default channel (step 83, table S1).

If do not relate to default channel, next step judges that whether this stream be for its stream (step 84 is shown S1) of setting certain dedicated channel, that is, enter VPI/VCI and whether be marked as dedicated channel.If this situation is directly taken out from table S3 and is entered the relevant Route Selection mark (step 86) of VPI/VCI identifier with this.

On the other hand, if as yet not to this flow distribution dedicated channel, this method enters step 88, reads the IP destination-address from cell in this step; As seen from Figure 3, this address always is mapped in cutting apart on the fixed position among the payload bytes 13-17 of cell.According to the destination-address of reading, from routing table, search the output port relevant (table S4) and can use Route Selection mark (table S5) accordingly with this address.At this output port, be spendable VPI/VCI identifier of this Route Selection label lookup (table S7); This identifier is arranged on (table S6) in the Route Selection mark conversion table of this output port.Simultaneously, linkage flag is become special-purpose, and upgrade the data in each table (table S1, S3, S4).

If in step 83, find to relate to default channel, upgrade with this IP address to or the relevant service counting (step 85, table S2) of IP destination-address, and judge according to count value whether this stream satisfies special-purpose connection criterion (step 87).If do not satisfy the special-purpose criterion that connects, route cell is in a usual manner promptly searched the route (output port) that is used for this specific IP destination-address and corresponding Route Selection mark (step 89, table S4).On the other hand, if satisfy special-purpose criterion, the process and the relevant process identical (step 88) of setting up a new dedicated channel of connecting.

After above steps, node finds correct routing label, and can be routed to correct output port (step 90) to cell.Before cell is sent to next link, remove the Route Selection mark and the setting VPI/VCI identifier (table S6) of going out to title.

From the flow chart of top oblatio as can be seen, when from several sources or the parallel same destination of chain road direction multiplexing when professional, still might for example set up dedicated channel to the node in the network according to destination-address.If adopt multiplexing on dedicated channel, it must be finished under the order of protection grouping.

Fig. 9 is with the possible implementation of the formal specification gateway node of functional block diagram.

Be the grouping that temporarily storage arrives, buffer 91 be set at the input end of node.The title that measurement and controll block 92 are read to divide into groups, and utilize whether source/destination-address is that described address is to establishing dedicated channel parameters to checking from its table (T1).

The IP address that retrieval piece 93 receives from measurement and controll block, and search the Route Selection mark that this retrieval piece offers block 95 according to this address from routing table (T2), carry out in this block that cell is cut apart and the Route Selection mark attached to cutting apart on the good cell.

In addition, the input side at node is provided with bookkeeping device 98.To follow the tracks of and each input, the relevant Route Selection the used mark of output port (table T3).

Cell and send to ATM switch 96 together from the Route Selection mark of cutting unit, the latter is connected to correct output port 97 to cell according to the Route Selection mark.Here, the Route Selection mark is an identifier in the switch, and it does not send in the network.

Measure and controll block is measured traffic carrying capacity in every stream, and in the given time cycle, do not detect relevant with this stream parameter of deleting certain dedicated channel when professional, for example Route Selection mark and VPI/VCI.

Figure 10 schematically shows the structure of network node.Each functional block of this intranodal is with top identical, but the table that this node uses is different from the table that gateway node uses.For this reason, each piece indicates with the same reference number of corresponding blocks among Fig. 9.The table S1 that different masses uses ... S7 is respectively at Figure 11 a ... describe among the 11g.

Measurement and controll block 92 are that still (table S1, Figure 11 a) for the information of dedicated channel for default channel for each VPI/VCI to keeping a counter and relevant channel.In addition, can be transformed into dedicated channel to the business on the default channel in order to make it, it be the IP address to or the IP destination-address keep a counter (table S2, Figure 11 b).In network, the right business in source/destination for example may be from the parallel link multiplexing, and perhaps business may be set up dedicated channel from several sources to certain given destination-address multiplexing for this reason.

Dedicated channel Route Selection mark corresponding to VPI/VCI is set in an independent table (S3, Figure 11 c), and they can be used by the Route Selection piece.In addition, for all connections are provided with a routing table, in this table IP address related with output port and Route Selection mark (table S4, Figure 11 d).The grouping of crossing for all routes in the default channel and be essential for setting up this table of new dedicated channel.

The Route Selection the used mark that bookkeeping device 98 is followed the tracks of with each is imported, output port is relevant (table S5, Figure 11 e).

The ATM output port has a conversion table, and the Route Selection mark of arrival cell substitutes (table S6, Figure 11 f) with the VPI/VCI of the cell of going out in the table.In addition, output port is followed the tracks of spendable VPI/VCI identifier (table S7, Figure 11 g).

Though top reference example has with reference to the accompanying drawings been explained the present invention, should be appreciated that the present invention is not subject to this, but revisable in the scope of the inventive concept of in the appended claims book, setting forth.For example, according to top described, stream can by identical sources/destination between business professional or that be tied to same destination constitute.

Claims

1. a method that transmits grouping in the packet-switched telecommunications net wherein is sent to another node to the node of grouping from network according to certain address of containing in each grouping, according to this method:

On the node that is positioned at the network boundary place (N1), the packet that transmit (30) is divided into low protocol layer data cell (50),

Transmit data cell in the default channel of this network and dedicated channel, the data cell of the grouping that destination-address is different transmits in the default channel of separating, and the packet relevant with same stream transmits in the dedicated channel of separating,

It is characterized in that

In a single day node on the network boundary (N1) monitors the grouping relevant with same stream, and detects the fact that satisfies predetermined criterion to constant current, distribute the low protocol layer connection identifier (CID of a special use to use for this stream,

Node (N2) in the network monitors the connection identifier (CID of arrival packet, and in case detect and default channel or the incoherent new connection identifier (CID of dedicated channel, destination-address according to this grouping distributes an output port and the new low protocol layer connection identifier (CID of going out to data cell, and data cell is routed to relevant output port

In case in the data cell that arrives in succession, detect the low protocol layer connection identifier (CID of the identical described new connection identifier (CID of conduct, only according to should low protocol layer connection identifier (CID in the enterprising walking along the street of this node (N2) by selection, and

After cycle preset time does not detect business in certain stream with special-purpose connection identifier (CID, on node, stop the special-purpose use that connects.

2. the described method of claim 1 is characterized in that, adopts similar out-of-date control on each node.

3. the described method of claim 1 is characterized in that, the node (N2) in the network also monitors the business on the default channel, and in case certain that detects certain default channel is distributed special-purpose low protocol layer connection identifier (CID to connecting when satisfying predetermined criterion to constant current.

4. the described method of claim 1 is characterized in that, the traffic carrying capacity that receives in given measuring period as described predetermined criterion.

5. the described method of claim 1 is characterized in that, their desired service quality information that contained in the packet as described predetermined criterion.