JP2001333105A - Communication quality assurance method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a communication quality assurance method that provides communication quality required for each application flow so as to prevent the quality from being deteriorated at path revision without increasing the number of queues for quality assurance. SOLUTION: Routers R1, R2,... manage a communication quantity by each combination of arrival interfaces receiving packets and transmission interfaces sending the packets, a reserved resource quantity is managed in the unit of each of autonomous systems AS1, AS2,... and when a path to an autonomous system is revised, reduction in a band is reported to an old path addressed to the system, the increase in the band is reported to a new path, and using a degree of branch and confluence confirms the arrival of a message about the band reduction and the band increase to a required router and then revises the path.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for guaranteeing communication quality in a communication network.

[0002]

2. Description of the Related Art With the development of communication networks represented by the Internet in recent years, various applications such as mail, file transfer, the World Wide Web, and live broadcasting have been used. With the further expansion of the user base in the future, the use of communication networks for higher quality video and audio applications, applications closely related to corporate activities, and applications requiring more certainty such as medical care and plant control will be increased. It is expected to expand.

[0003] Some of these applications are sensitive to the communication quality provided by the network, and the applications are provided to users due to fluctuations in bandwidth, absolute communication delay time, and delay time. The quality of moving images and voices deteriorates, companies cannot provide services promised to be contractually provided, and in the field of telemedicine, problems such as a more difficult diagnosis cannot be performed in an emergency.

[0004] The present invention relates to a communication quality assurance method for using an application that is more sensitive to communication quality and requires predictable quality among applications using such a communication network.

[0005] Conventional techniques of this type include:-IntServ (S. Shenker et al., "Integrated services i
n the Internet arechitecture: an overvier ", IETF RF
C 1663)-DiffServ (Y.Bernet et al., "A framework for diffe
rentiated services ", IETF Internet draft, draft-iet
f-diffserv-framework-02.txt) ・ SCORE (I.Stoica et al., "Providing Guaranteed Ser
vices Without Per Flow Management ", Proceedings of
1999 ACM SIGCOMM Conference, pp.81-94).

[0006] IntServ is characterized by reserving network resources for each flow used by each application. Usually, in the case of a protocol used in the Internet, each flow is identified by a transmission address, a reception address, a transmission port, and a reception port.

[0007] One of the protocols for making this reservation is R
SVP (Resource Reservation Protocol: R. Braden et
al., "Resource Reservation Protocol", RFC 2205, 19
In 97), each router manages the location where resource reservation should be made as a path state by transmitting a message called a path message from the transmission side to the reception side, and sends a resource reservation request from the reception side to the transmission side. I will go. The router that has made the resource reservation identifies the packet for each session and guarantees the communication quality of the reserved resource by using a queue prepared for each flow.

[0008] DiffServ is characterized in that a packet is divided into a relatively small number of classes, and there is a difference in how to handle the packet. The policy of this handling differs depending on the administrator of the network, and there are domains in the network that make the policy the same. When a packet enters a certain domain, the router at the boundary of the domain polices based on the class attached to the packet,
Change classes, control priority, etc.

[0009] In SCORE, there exist a domain called SCORE of the network and other places. Outside SCORE, like IntServ, packets are distinguished by session. MPLS (Multiprotoco
l Label Switching: R. Callonet al., "A framework for
multiprotocol label switching ", Internet draft, dr
aft-ietf-mpls-framework-02.txt, 1997), etc., and the packet is transferred outside the SCORE using resources reserved for each session.

[0010]

[Problems to be Solved by the Invention] The problems to be solved by the present invention are: provision of necessary communication quality to each flow, reduction of the number of queues for guaranteeing communication quality, and quality at the time of route change. This is to prevent deterioration, and each will be described below.

<Provision of Necessary Quality for Each Flow> As described above, the present invention provides, among applications using a communication network, applications that are more sensitive to communication quality and require predictable quality. It is for. However, DiffServ is easily affected by other traffic because communication resources allocated to each class and a policy for classifying are individually performed in each domain.
Therefore, depending on the situation of other traffic, a situation occurs in which communication quality required by a certain application cannot be provided.

<Reduction of Queues for Assurance of Communication Quality>
The mechanism for providing the reserved communication resources in the router is realized by a queue called a packet scheduler. 2. Description of the Related Art In recent years, with the increase in bandwidth of communication lines, there is a tendency that packet transfer of routers is implemented by hardware. Therefore,
To get sufficient router performance, the packet scheduler needs to be implemented in hardware. However, since the packet scheduler requires a large amount of hardware, the cost becomes enormous in a network where sufficient packet transfer capability is required, and it is difficult to implement a large number of packet schedulers in practice. is there.

[0013] In IntServ, the maximum number of schedulers is the number of flows passing through the router, and the number of schedulers becomes enormous especially in a core network where a large amount of traffic flows. In SCORE, by increasing the domain of MPLS, it is possible to reduce the number of networks in which a packet scheduler must be prepared for each flow. However, conversely, as the domain of the MPLS increases, the number of labels increases, and many schedulers must be prepared.

<Prevention of Quality Degradation at Route Change> In the Internet, the route currently being used is changed due to a failure of a router or a line, a stop of a network device for maintenance, addition of a new network, or the like. There is. However, in IntServ and SCORE, even if the current route is available, after the route change is made, until the receiving or transmitting computer executes the resource reservation process for the new route and is accepted. Communication quality cannot be guaranteed or provided. Further, in RSVP, there is a method called route pinning. When this is used, the route cannot be changed while an application using the route exists.

As described above, conventionally, it has been difficult to simultaneously solve the above three problems.

[0016]

According to the present invention, in order to solve the above-mentioned problems, in each router, an interface from which a packet arrives (hereinafter, referred to as an arrival interface) and an interface for transmitting a packet (hereinafter, a transmission interface) Each router manages the amount of reserved resources for each partial network.-A router that changes the route to a certain partial network has resources on the old route addressed to that partial network. Notify the amount of decrease, notify the increase of the amount of resources on the new route, and use the degree of branching and the degree of confluence to confirm that all the decrease and amount of resources messages have arrived at the required router. Use means.

According to the present invention, the resources of the flows using the route from the incoming interface to the outgoing interface are collectively managed by the above configuration. This firstly reduces the number of queues that each router must implement. Now, if there are n interfaces in the router, the number of queues is n (n-1)
It is. Second, communication resources provided for a flow from a certain application on the transmission side to an application on the reception side are hierarchically managed based on the chain of the combination of the incoming interface and the outgoing interface. In this way, routers in the network share communication resources for each application flow, and provide the required communication quality for each flow.

Next, with the above configuration, it is possible for a router that changes a route to a partial network that is a destination when a route is changed to move a communication resource affected by the changed route from an old route to a new route. And Then, according to the above configuration, after confirming that all the changes have been distributed to the routers requiring the resource change accompanying the route change, the route is changed, thereby making the reserved resources the same before and after the route change, thereby preventing quality deterioration. I do.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION

[0020]

First Embodiment <Premise> In this embodiment, an embodiment using the Internet as a communication network described in the claims will be described. In this case, the partial network is an autonomous system, and the reserved bandwidth is used as the reserved resource amount.

FIG. 1 shows an example of an autonomous system, a router, and a user communication device (terminal) which are partial networks in the Internet. Here, the autonomous system A
There are S1, AS2, AS3 and AS4. A router and a user communication terminal are connected to each autonomous system.
In the figure, the ones indicated by 印 are routers, and the ones indicated by □ are user communication terminals S1, S2, ..., D1, D2.
... Each autonomous system is connected by its own router. In this figure, communication from the user communication terminal S1 of the autonomous system AS1 to the user communication terminal D3 of the autonomous system AS3, and from the user communication terminal S3 of the autonomous system AS3 to the user communication terminal D1 of the autonomous system AS1
Communication to is shown.

Normally, the calculation of the route of the Internet is divided into the autonomous system and between the autonomous systems, and is calculated by different routing protocols. The calculated route is managed by each router in the form of a routing table. The routing table contains the destination address, mask and information of the next router on the route to the destination. These may be within an autonomous system and between autonomous systems, and there may be at least one next router for one destination address. If each record on the routing table is expressed as a directional branch from one router to the next router, for a certain destination, a node that is a DAG (Directed Acyclic Graph) and has an out-degree of 0 There will be only one.

<Overview> In order to implement the present invention, each router section needs to: 1. Packet transfer function based on reserved resource amount 2. A function for making necessary resource reservations between autonomous systems. Introduce a function to change the reserved resources when the route is changed.

Hereinafter, each of them will be described.

<1. Packet Transfer Function Based on Reserved Resource Amount> In the present invention, as described above, a queue for managing a communication resource amount for each combination of an incoming interface and an outgoing interface is introduced. FIG. 2 shows an example of the relationship between a queue introduced into a router and an interface. In this example, the router has a total of four interfaces. Also, each interface has a queue for another interface.

That is, queues Q2, Q3, and Q4 for interfaces 2, 3, and 4 from interface 1; queues Q1, Q3, and Q4 for interfaces 1, 3, and 4 from interface 2; Queues Q1, Q for 1, 2, 4
2 and Q4, and the interface 4 has queues Q1, Q2 and Q3 for the interfaces 1, 2 and 3.

In each of these queue portions, a Virtual Clock
(L. Zhang, "A New Architcture for Packet Switched N
etwork Protocols ", Massachusetts Institute of Technology Doctoral Dissertation, 1989).

FIG. 3 shows how traffic to the AS 3 is managed by a combination of an incoming interface and an outgoing interface. The traffic from AS4 is R1 → R3 at R2, and R3 at R3.
2⇒R5, R5 R3⇒R6, R6 R5
⇒R8 is managed as R6 → R9 in R8. The traffic from AS1 is managed as R4 → R6 in R5, and is merged with the traffic from AS4 and managed.

<2. Function to make necessary resource reservation between autonomous systems> Using "resource reservation message" to reserve resources required for communication between autonomous systems, route to the destination autonomous system based on the policy of each router This message is propagated while being processed and changed along with it, and resources are reserved for necessary routers. Also, the reserved bandwidth is managed for each destination autonomous system.

The following describes the policy of the router, the format of this message, the message propagation method, and the resource reservation status managed by the router.

(Policy) The policy of each router is as follows: 1. Maximum transfer bandwidth It contains two parameters, the ratio between neighboring routers.

"Maximum transfer bandwidth" is the maximum value of the bandwidth transferred from a certain adjacent router to another adjacent router. Represents the maximum bandwidth allowed by various neighboring routers when a resource reservation request message is received.

The "ratio between adjacent routers" means that a plurality of routes are used for a certain partial network as a destination, that is, a route from a certain router to a partial network as a certain destination branches, If it is set to use, it indicates the ratio of allocating the reserved bandwidth to each route.

(Format of Resource Reservation Message) This message includes information such as: the address of the destination autonomous system and the reserved bandwidth.

(Resource Reservation Message Propagation Method) As an initial state, each router creates the resource reservation message for each destination autonomous system required by the user communication terminal directly connected to the router.

2. The router reserves the resources and sends the message to the next router on the route to each destination.

3. It is assumed that a certain router k receives a resource reservation message from the adjacent router i to the destination autonomous system j, and that the reserved bandwidth is r _i ^(k) (j). Also, the set of the next router and the previous router on the route to the destination autonomous system j in a certain router k is represented by N _j
^(k) and P _j ^(k), and the ratio between adjacent routers for the adjacent router i in the router k is t _i ^(k).
The reserved bandwidth of the resource reservation message that the destination autonomous system j sends to the next router l is

[0038]

(Equation 1)

It is assumed that

4. In router i, the reserved bandwidth of the queue for traffic with the interface that receives packets from neighboring router k as the incoming interface and the interface that sends out packets forwarded to neighboring router j as the outgoing interface.

[0041]

(Equation 2)

(Where 1 represents the destination autonomous system).

FIG. 4 shows a flow chart of the processing at the time of resource reservation.

(Resource Reservation Status Managed by Router) FIG. 5 shows the resource reservation status managed by the router. As shown in this figure, the table indicating the resource reservation status includes the autonomous system and the bandwidth.

<3. Function to change reserved resources at the time of route change> When a certain router changes the route to a destination autonomous system, the router that changes the route sends the next router before the change and the next router after the change. A "reserved resource change request message" is transmitted and propagated along the path toward the destination autonomous system. It is determined that the propagation of the "reserved resource change request message" has been completed. To other routers that have received

The format of this message and the method of transmitting the reservation resource change request message will be described below.

(Format of Reserved Resource Change Request Message) This message includes information such as: the address of the destination autonomous system, the reserved change bandwidth, the degree of branching, and the degree of confluence.

(Reservation Resource Change Request Message Propagation Method) As an initial state, a router that changes the route to a destination autonomous system d sends a request for decreasing R to the old route and a request to increase R to the new route for the reserved bandwidth R to d. At this time, the branching degree and the merging degree are both 1.

2. When the router receives the reservation resource change request message, it locks the route change to the destination autonomous system included in the message.

3. Upon receiving the reservation resource change request message, each router changes the reservation change bandwidth in the same manner as the method of processing the reservation bandwidth of the resource reservation message, and creates a message to the next router.

4. At this time, the processing of the branching degree and the merging degree is performed as follows. Now, a reservation resource change request message i (1
≤ i ≤ N) the degree of branching and the degree of confluence are F _i and J _i , respectively. At this time, when the N messages are merged and sent, the branching degree and merging degree of the new message n are respectively

[0052]

(Equation 3)

And

[0054]

(Equation 4)

Assume that: Further, when there message i is branched into M of the plurality of paths merging degree unchanged, degree of branching and M · F _i.

5. Degree of branch of received message i is F
_{Let i} be the degree of confluence and J _i

[0057]

(Equation 5)

If so, it is determined that all reserved resource change request messages have been received.

6. The router that has determined that the reservation resource change request message has been completely received for the new route and the old route sends a resource reservation change request delivery completion message to the router that sent the reserved resource change request message.

7. The router that has received the reservation resource change request delivery completion message similarly sends a reservation resource change request delivery completion message to the router that sent the reservation resource change request message.

8. The router that changes the route instructs acquisition of resources for the new route. After completion, change the route.

9. After the route is changed, the router that changes the route instructs the resource change of the old route.

FIG. 6 shows a flow chart of the process when the route is changed.

(Example of Reservation Resource Change Request Message Propagation)
FIG. 7 shows how a reserved resource change request message is propagated. In this figure, the router R1 changes the route to the autonomous system including R10 from R2 to R12. At this time, R1⇒R12, R12⇒R13, R13⇒
R2, R13 ⇒ R5, R1 ⇒ R2, R2 ⇒ R3, R3 ⇒
R4, R3 ⇒ R5, R4 ⇒ R6, R5 ⇒ R7, R6 ⇒ R
7 resources are changed. Routers R7, R8, R9, R1
0 can determine that the reservation resource change request has been completed, and propagate the reservation resource change request delivery completion message to the router R1.

[0065]

Embodiment 2 This embodiment is different from Embodiment 1 in that
1. In the packet transfer function based on the reserved resource amount>, the time interval between each packet in the series of packet strings is adjusted, and the other parts are the same as in the first embodiment.

As shown in FIG. 8, the present packet transfer mechanism is realized by providing a packet interval adjusting unit ad in the incoming interface of the router and an aggregate flow scheduler sc in the outgoing interface.

Hereinafter, the packet interval used in the present packet transfer mechanism will be described, and then the packet interval adjusting unit and the aggregated flow scheduler unit will be described.

(Packet Interval) In the present packet transfer mechanism, a packet interval, which is a time interval from the immediately preceding packet, is inserted into each packet and used for a series of packets arriving through a certain incoming interface. As the packet interval here, the interval of the transmission start time of each packet is used.

(Packet Interval Adjustment Unit) For a packet arriving at a certain incoming interface of a router, the transmission interface is determined according to the destination of the packet. Generally, a packet sequence arriving at one incoming interface branches into a packet sequence for each sending interface. At this time, when a packet immediately before a certain packet has a different packet sequence toward a different transmission interface from that packet, the packet interval included in the packet differs from the packet interval set by the immediately preceding router. Value.

In order to correct this, the packet interval adjusting unit has a table for managing the operation amount of the packet interval included in the packet destined for the interface for each transmission interface. This table is hereinafter referred to as a packet interval operation table. Each record of the packet interval operation table manages the correspondence between two values. The first one is,
The identifier of the transmission interface, and the second is the time to be added to the packet interval included in the packet.

Here, when the identifier of the incoming interface is j and the identifier of the outgoing interface is k, the adding time is expressed as a _{j, k} . The initial value of a _{j, k} is set to 0 for all j and k.

Now, when a packet including the packet interval g arrives at the incoming interface having the identifier j and the identifier of the transmission interface of the packet is k,
Packet interval operation table and packet interval g included in the packet

[0073]

(Equation 6)

The operation is performed as follows.

(Aggregated flow scheduler unit) The aggregated flow scheduler unit has queues (not shown) for storing packets sent from each incoming interface in a distinguished manner, and takes out packets from each queue. The time to be transmitted (hereinafter referred to as “eligible time for each incoming interface”), the time when the last packet was transmitted from the queue for each incoming interface (hereinafter referred to as “the last transmission time for each incoming interface”), and the last packet transmitted from the outgoing interface At this time (hereinafter, referred to as final transmission time for each transmission interface), the available bandwidth is managed for each combination of the arrival interface and the transmission interface.

Now, the band which is the communication resource allocated from the arrival interface of the identifier j to the transmission interface of the identifier k is represented by r _{j, k} , and the qualifying time for each arrival interface for the arrival interface of the identifier j in the transmission interface of the identifier k the q _{j, k,} incoming interface every last transmission time of l _j for incoming interface identifier j in sending interface identifier _{k, k,} the transmission interface each last transmission time of the transmission interface identifier k a o _k. The size (length) of the _nk- th packet among the packets transmitted from the arrival interface of the identifier j to the transmission interface of the identifier k in the router i is

[0077]

(Equation 7)

Assume that

The aggregation flow scheduler unit transmits the packet by obtaining the qualifying time q _{j, k} for each incoming interface and sets the value of the packet interval in the packet by any of the following three methods.

[0080] First, in the first method (corresponding to claim 5), the initial value of q _{j, k} and o _k is the time at which the first packet has arrived. Next, the aggregation flow scheduler unit sends a packet every time a packet is transmitted or when a packet enters an empty queue.

[0081]

(Equation 8)

Is calculated for the incoming interface j whose queue is not empty, and j for which q _{j, k} is minimized is obtained.

Next, when q _{j, k} is equal to or smaller than the current time, the packet is transmitted. When transmitting a packet, the value of the packet interval in the packet is set to q _{j, k} −o _j
And then o _j = q _{j, k} .

[0084] In the second approach (corresponding to claim 6), the initial value of q _{j, k} and o _k is the time at which the first packet has arrived. Next, the aggregation flow scheduler unit sends a packet every time a packet is transmitted or when a packet enters an empty queue.

[0085]

(Equation 9)

Is calculated for the incoming interface j in which the queue is not empty, and j that minimizes q _{j, k} is _obtained .

Next, when q _{j, k} is equal to or smaller than the current time, the packet is transmitted. When transmitting a packet, the value of the packet interval in the packet is set to q _{j, k} −o _j
And then o _j = q _{j, k} .

[0088] Further, the third method (corresponding to claim 7), for q _j, the initial value of _k, the time at which the first packet has arrived, l _{j, k} and o _k is the identifier of the k In the transmission interface, the transmission time of the packet last transmitted from the queue for each arrival interface and the time of transmission of the packet from the transmission interface with the identifier k are set.

Next, the aggregate flow scheduler unit transmits a packet every time a packet is transmitted or when a packet enters an empty queue.

[0090]

(Equation 10)

Is calculated for the incoming interface j in which the queue is not empty, and j that minimizes q _{j, k} is _obtained .

Next, when q _{j, k} is equal to or smaller than the current time, the packet is transmitted. When transmitting a packet, a p as the current time, set the value of the packet interval in the packet to _{p-o j, o j =} p and l _j, and _k = p.

[0093]

As described above, according to the present invention, a communication resource is managed for each incoming interface and a sending interface; each router manages a reserved resource amount for each partial network; The router that changes the route to the network notifies the resource decrease on the old route addressed to the partial network, notifies the resource increase on the new route, and uses the branching degree and the merging degree to calculate the total resource amount. As described above, by confirming that the decrease and resource increase messages have arrived at the necessary routers: Providing the required communication quality for each flow; Reducing the number of queues for guaranteeing the communication quality;・ The problem of preventing quality degradation when changing routes can be solved.

[Brief description of the drawings]

FIG. 1 is a diagram showing how a packet is delivered on the Internet.

FIG. 2 is a diagram showing the structure of a router according to the present invention.

FIG. 3 is a diagram showing an example of traffic management in the present invention.

FIG. 4 is a flowchart of processing at the time of resource reservation.

FIG. 5 is a diagram showing an example of a resource reservation state managed by a router.

FIG. 6 is a flowchart of a process when a route is changed;

FIG. 7 is a diagram showing an example of how a reserved resource change request message is propagated;

FIG. 8 is a diagram showing a structure of a router corresponding to the second embodiment.

[Explanation of symbols]

S1, S2, ..., D1, D2, ...: user communication terminals, R1, R2, ...: routers, AS1 to AS4: autonomous system, Q1 to Q4: queue, ad: packet interval adjustment unit, sc: Aggregation flow scheduler part.

Claims (7)

[Claims] In a communication network in which a user communication device and a router are connected, the communication network is divided into a plurality of partial networks and at least one route from a certain router to a destination exists. It manages the traffic for each combination of the incoming interface and the sending interface that sends out packets, manages the amount of reserved resources for each partial network, and changes the route to a certain partial network when changing the route to that partial network. Notify the resource decrease on the route and notify the resource increase on the new route, and use the degree of branching and confluence to arrive at all necessary resource decrease and resource increase messages at the required router. A communication quality assurance method characterized by changing a route after confirming the communication quality. 2. A communication device having a communication interface to which an address for uniquely identifying is assigned, and a user communication device for transmitting and receiving a packet via the communication interface, and an address for uniquely identifying the user interface. A communication network having at least one assigned communication interface and connected to a router that performs packet transfer via the communication interface and performs a route calculation for transferring the packet; The network is divided into partial networks, each of which has a partial network identifier and a partial address set to be used therein, and the route calculation for other partial networks is performed in units of partial address sets. Smaller subnetworks in the network Communication resources for guaranteeing the communication quality between user communication devices in a communication network that is divided into a backbone network for connecting networks and another network and has at least one route from a certain router to a destination. The reservation is separately performed in the partial network and between the partial networks.
A border router connected to another sub-network makes a resource reservation between the sub-networks, and a resource reservation in the sub-network is made by a router in the sub-network. The communication transmitted and received by the user communication devices within the network and the communication transferred between adjacent sub-networks are collectively performed for each communication addressed to the same sub-network. For communication between routers in the backbone network, A router in the network provides a queue for each combination of an incoming interface from which a packet arrives and a sending interface for sending the packet, manages the packet, and stores the transfer resource in the queue based on the amount of resources obtained by the resource reservation. Assign and come Monitor that the assigned packet is within the allocated transfer resource, discard the packet if the resource is exceeded, transmit the packet using the transmission interface within the allocated transfer resource, and For a packet transmitted by a user communication device in the network and arriving at a user communication device belonging to a partial network other than the partial network to which the user communication device belongs, the first user communication device in the route to which the user communication device belongs belongs Backbone of partial network
A router in the network manages the packet for each session to which the packet belongs, monitors that it is within the resource amount of the session allocated at the time of the resource reservation, discards the packet if the resource is exceeded, and allocates the packet. A communication quality assurance method, wherein a router within the backbone network manages the packet for each of the incoming interface and the outgoing interface so as to transmit the packet within the transfer resource. 3. The resource reservation performed by each of the partial networks is performed in a direction from a transmitting side to a receiving side or in a direction from a receiving side to a transmitting side. When the message including the reserved resource amount is sent from the receiving side to the transmitting side, the source partial network, the destination partial network, and the message including the reserved resource amount are transmitted along a path along which the packet is transferred, When the route to the partial network branches, each router divides the resource amount designated in the message into each route and transmits the divided messages as a plurality of messages, and the destination or the source partial network is the same. Change the message to an amount of resources that can satisfy the requirements of all messages, and combine them into one resource reservation request message. Communication quality assurance method according to claim 2, characterized in that the resource reservation by causing Umate propagate. 4. When a router changes a route to a destination partial network, a resource reservation is made from the transmitting side to the receiving side on the route before the change and the route after the change before actually performing the change. When performing the reservation resource change request message including the amount of change, the destination partial network, the degree of branching of the route to the destination partial network, and the degree of merging of the branched paths,
When resource reservation is performed from the receiving side to the transmitting side, a reservation resource change request message including a change amount, a destination partial network, a transmission source partial network, the branching degree and the merging degree is transmitted, and each router When the route to the network branches, the resource amount included in the reserved resource change request message is divided and transmitted as a plurality of messages, and the router that has received all the messages or the border router of the destination partial network cooperates. After confirming that all messages have been received, the completion of the change request delivery is notified in a direction opposite to the direction in which the reserved resource change request message was transmitted. The router which changes resources and changes the route changes the route after receiving the notification. Communication quality assurance methods. 5. A method for performing packet transmission by managing packets for each of an incoming interface and an outgoing interface, wherein a packet interval, which is a time interval between each packet of a series of packet streams flowing through a communication line and the immediately preceding packet, is used. When the series of packet sequences is branched from the incoming interface of the router to a transmission interface corresponding to the destination of each packet and becomes a packet sequence for each transmission interface different from the packet sequence, the packet interval is set to Change to the packet interval of the packet sequence for each transmission interface, and when transmitting packets from
A packet is stored in a queue for each incoming interface, an eligible time for each incoming interface that is a time at which a packet is to be taken out from each queue and transmitted, and a last transmission time for each outgoing interface that is the last time a packet was transmitted from the outgoing interface. The sum of the packet interval stored in the packet and the qualifying time for each arriving interface is defined as a new qualifying time for each arriving interface, and the queue in which the qualifying time for each arriving interface is equal to the current time or is in the past. Selecting a queue having the minimum qualifying time for each arriving interface, extracting the packet stored in the selected queue from the queue, and determining the packet interval in the fetched packet based on the qualifying time for each arriving interface and the final transmission for each sending interface; 3. The communication quality assurance method according to claim 2, wherein the time is reset to a difference from the time, and the qualified time for each incoming interface is set as the final transmission time for each transmission interface. 6. A method for calculating the qualifying time for each incoming interface, the qualifying time for each arriving interface, the bandwidth to be used by each queue and the time to transmit the next packet conforming to the bandwidth calculated from the size of the last transmission packet. The smaller of the packet interval stored in the packet and the sum of the qualifying time for each arriving interface is set as the qualifying time for each arriving interface.
Communication quality assurance method described. 7. A method for setting a qualified time for each incoming interface and a final transmission time for each outgoing interface, wherein the last transmission time for each incoming interface, which is the last time a packet was transmitted from a queue for each incoming interface, is managed. The qualifying time includes the last transmission time for each incoming interface, the bandwidth that can be used by each queue and the time to transmit the next packet conforming to the bandwidth calculated from the size of the last transmission packet, and the packet stored in the packet. The smaller of the interval and the sum of the last transmission time for each incoming interface is taken as the qualified time for each incoming interface, and the time when the packet was actually sent is used as the last transmission time for each sending interface, and the packet is reset when sending. Packet interval Using the difference between the last transmission time for each transmission interface and the time to start packet transmission, setting the time to start packet transmission to the last transmission time for each transmission interface and the last transmission time for each arriving interface. The communication quality assurance method according to claim 6.