JPH0348553A - Communication resource allocating system for packet switching network - Google Patents

Abstract

PURPOSE:To guarantee uniform communication quality for respective communication by dividing a communication resource with respect to a packet switching machine into plural independent resources for communication classified appropriately, and controlling the communication quantity of the resource at every divided resource. CONSTITUTION:The communication resource R with respect to the packet switching machine 3 such as the band area of a communication channel or buffer capacity in a packet switching network is divided into the plural independent resources ri (i=1, 2, 3...N) for the communication classified appropriately. And the communication quantity of the resource is controlled so as to satisfy the requested communication quality of the communication housed in the communication resource ri at every divided resource ri. Therefore, each resource ri (r1, r2,...rN) obtained by dividing all the communication resources Ri can be set independently for another resources ri, and also, since the requested quality of communication can be satisfied at every class, the uniform communication quality can be guaranteed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to a communication resource allocation method for a packet-switched network.

(Prior art) The communication resource allocation method of the conventional packet-switched network uses the so-called statistical multiplexing effect to efficiently manage the communication resources within the network by sharing the communication resources with multiple communications. In order to achieve this, in most cases, communication resources within the network were allocated completely equally to each type of communication accommodated by the network.

That is, in the packet switch 1 shown in FIG.
The resource R is shared by communication 2, which is appropriately classified into classes 1 to N, and each resource rl (
i=1. 2. 3...) were to be shared by different classes, as shown in FIG.

In addition, priorities have been assigned to each communication class according to the degree of cell discard rate requirement, for example, and so-called priority control has been performed to satisfy the cell discard rate requirement of each communication to some extent.

(Problem to be Solved by the Invention) However, in the conventional communication resource allocation method in the bucket switching network as described above, each resource is allocated completely equally to each communication, so that packets exceeding the processing capacity of the network If this was done manually, there was a problem in that the communication quality, such as the packet loss rate and transfer delay, would uniformly deteriorate for each communication.

In addition, when classifying each communication and performing priority control according to the request for the cell discard rate, communication resources are shared between communications with different packet discard rate requirements, so in this case the actual discard rate changes depending on the relative value of the communication amount of each request class, so it is almost impossible to accurately satisfy the cell loss rate. Further, in this case, since the communication speed of each communication is not taken into account, it is difficult to reduce the cell discard rate for each communication to a certain value or less.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a communication resource allocation method for a packet switching network that can uniformly guarantee communication quality for each communication.

[Structure of the Invention] (Means for Solving the Problems) The communication resource allocation method in a packet switching network of the present invention for solving the above problems is based on packet switching as shown in FIG. Communication resources R related to the packet switch 3, such as bandwidth or buffer capacity of communication lines in the network
multiple independent resources r for appropriately classified communications.
l (1=1.2, 3...N), and for each divided resource r1, control the amount of communication so as to satisfy the required communication quality of the communication accommodated by that communication resource r. It is characterized by

(Operation) In the communication resource allocation method of the packet switching network of the present invention,
The total resource R is divided into appropriately classified communications, the communication resource r1 accommodates each divided resource Kri, and the amount of communication is controlled so as to satisfy the required communication quality of the communication.

Therefore, as shown in FIG. 2, each resource r+ (r+ 1 '2...
・rN) is independent from other resources r1 and satisfies the required communication quality for each class, so that communication quality can be uniformly guaranteed.

For example, for each class of cell discard size required by each call, the corresponding communication resource r is allocated, and as many calls as can be accommodated by that resource can be accommodated while satisfying the required quality. , the communication quality that can be guaranteed by that resource can be easily predicted. Furthermore, since the communication resources r divided into classes are used independently of each other, even if a certain communication resource r1 has more communication companies than its processing capacity, other communication resources r1 will have a certain level of capacity. Communication quality can be guaranteed. Note that the same applies to control channels.

(Example) Examples of the present invention will be described below.

FIG. 3 shows that in the communication resource allocation method of the packet communication network shown in FIGS. 1 and 2, within each class C,
This is an advantage of not performing priority control.

That is, in the figure, each class C communication is individually communicated to the core switch via a buffer, and each communication is controlled completely independently with respect to other communications.

On the other hand, FIG. 4 shows an example in which priority control is performed within each class Ct.

That is, in the figure, communication of each class CI is communicated to the core switch section via the priority control buffer 4, and priority control is performed by the operation of the priority control buffer 4.

FIG. 5 is a flowchart of processing when resource R is statically divided into classes C.

That is, in this example, if there are N communication quality requirement classes C1 in a certain packet switching network, the network divides the communication resources R such as bandwidth, buffer amount, or both in the network into N overlapping classes. It is physically or logically divided into independent communication resources (sub-resources) r+ without resources.

Therefore, when a communication request accompanied by an application for communication request quality occurs in step 501, the communication class is analyzed in step 502, and in step 503, the communication when the requested communication is accommodated in the divided resource (subresource) r+ Evaluate quality.

At this time, the network has a certain communication quality requirement class 1.
When such a communication request occurs, the state of the currently used communication resource r1 is known, and when the communication of the call making the communication request is added to the traffic in that state, in step 504 Determine whether class 1 communication request quality can be guaranteed, and if it is guaranteed ↑
Only when the request call is disconnected will the communication of the requested call be accommodated.

However, when dividing the communication resource R into N sub-resources r, the size of each sub-resource r1 is a predetermined value. That is, the size of each sub-resource r1 is set in advance depending on the amount of communication of each class that the communication resource under consideration accommodates or will accommodate.

Therefore, when it is determined in step 504 that the communication quality can be satisfied, the process moves to step 505, and the communication requesting the communication is accommodated. Furthermore, if it is determined in step 504 that the communication quality cannot be satisfied, the process moves to step 506, where the communication requested is rejected.

In this example, examples of priority control when performing priority control based on the communication speed class of the communication accommodated by the sub-resource r1 include so-called non-interrupt type priority control and interrupt type priority control methods. . For example, when performing non-interrupt type priority control in this embodiment, separate buffers are prepared according to each communication speed class, and if there is a packet in the buffer used by high-speed communication, Outputs unconditionally, but if there are packets in a buffer that is being used by a faster communication than the one being focused on, packets will be output from the focused buffer until all packets are exhausted. Just make sure it doesn't happen.

FIG. 6 is a flowchart of processing when at least one of the divided resources is used as a spare resource.

In this case, when there are N communication quality request classes, the communication resources R in the network are divided into (N+1) subresources. Of these, N are fixedly allocated to the corresponding communication quality requirement classes, and the remaining one sub-resource is prepared as a reserve resource rN+, so that each sub-resource responds to new communication requests by allocating its own resources. In other words, when a communication request (class I) newly occurs for all subresources r, it is possible to accommodate the communication request. Try not to let the river flow.

Therefore, in steps 601 to 605, processes similar to steps 501 to 504 in FIG. 5 are executed, and step 60
If it is determined in step 4 that the communication quality cannot be satisfied, the insufficient resource amount is calculated in step 606, and it is determined in step 607 whether the insufficient resource amount can be secured from the reserve resource rN+1.

Then, in step 607, if it is possible to secure 'I'l1
If the communication is disconnected, the process moves to step 608 to secure insufficient resources, and then, in step 609, the requesting communication is accommodated. Thereafter, when the amount of communication resources requested for class Ci decreases, the resources temporarily taken in from the reserve resource rN+1 are released.

If it is determined in step 607 that insufficient resources cannot be secured, the process moves to step 610 and the communication requesting is rejected.

The sub-resource r, uses the communication resources for the shortage as a reserve resource l! i.
The algorithm for temporarily fetching from rN+1 is completely unlimited, that is, a method in which each sub-resource r1 attempts to fetch spare resources independently.
In principle, there is a method in which a certain sub-resource r can take in all the communication resources of reserved resources rN+1, and each sub-resource r1 has a maximum value of the amount of resources that can be taken in from the reserved resources. , an example is a method that does not attempt to take in spare resources above that value.

Priority control is also similar to the previous example.

FIG. 7 is a flowchart of processing when the boundaries of each sub-resource r1 are dynamically varied.

In this example, in the case of a packet switching network with N communication quality request classes, the amount of communication resources corresponding to each class is not determined in advance, and when a communication request for a certain class occurs, The necessary communication resources are incorporated into the communication resources of that class. However, in this case as well, similar to the examples shown in FIGS. 5 and 6, communication resources (subresources) of different classes are prevented from sharing resources. When managing communication resources like this,
When a new call must be accommodated, an attempt is made to acquire the communication resources necessary to guarantee communication quality for that class.

That is, in FIG. 7, a communication request is generated along with an application for communication request quality in step 701, the communication class is analyzed in step 702, the insufficient resource amount is calculated in step 703, and the insufficient resource amount is calculated in step 704. It is determined whether or not it is possible to secure resources from among the resources that are not used for communication.

If it is determined in step 704 that it can be secured, in step 705, the resource is secured from among the remaining resources,
Next, a step 706 accommodates the communication requesting the communication.

On the other hand, if it is determined in step 704 that the insufficient amount of resources cannot be secured from the remaining resources, the process moves to step 707, where the communication requested is rejected.

For information on algorithms for securing scarce resources, see Chapter 6.
It is similar to that shown in the figure.

Priority control is similar to that shown in FIG.

Possible parameters of the communication quality requirement class that can be used in the second embodiment include the packet/soto discard rate, transfer delay, and communication speed, and classification can be performed by combining these.

(2) Packet discard rate For example, a required quality standard such as 10-3.10-6.10-8 can be considered.

■Transfer delay Standards for the size of transfer delay or the size of delay fluctuation can be considered.

■Communication speed A method of expressing the communication speed class as a unique parameter in the network using the absolute value of the communication speed ([]ρS),
There is a method of expressing the communication speed as a ratio to the size of the resources (bandwidth, buffer capacity, etc.) that manage communication resources.

Regarding the latter, the ratio between the capacity and the operation speed of accommodation calls from the perspective of multi-city expansion, or the Banoff γ size in the area where multi-city expansion is carried out and the breadth that is considered necessary for accommodation There are classes of size ratios, etc.

Furthermore, this communication request class has two cases: when a certain terminal makes a call, the terminal itself interrupts the class itself, and the terminal itself uses the communication speed and the value of the cell discard request (absolute value or absolute value).
iv class (different from the communication request class)). Regarding the latter, the requested value is converted into a communication request class for the communication resource such as the multiplex line or buffer in question, and which subresource to use is determined.

In addition, as a method for changing the request class depending on the size of the communication resource, there are various sizes of communication resources R in the network, so when allocating the communication resource R to the subresource r, Even for the same communication request, different subresources (classes) can be allocated depending on the size of the communication resource r. this is,
Among the evaluation parameters for classifying communication requests mentioned above,
This is related to communication speed.

That is, for example, if we consider a multiplex line as a communication resource, and consider a resource rB with a large capacity of the multiplex line and a resource r5 with a small capacity, communication that belongs to a slow class when viewed from the resource rB with a large capacity. There are cases where a request is in a high-speed class when viewed from the resource r5, which has a small capacity.

Alternatively, communication requests req I and req 2 belonging to the slow class when viewed from the resource rB with a large capacity are request H when viewed from the resource r5 with a small capacity.
may be a high-speed class, but req 2 may be a slow class, and in this example, appropriate control can be performed according to these circumstances.

Next, an example in which a band is used as a communication resource will be described.

First, as shown in FIG. 8(a), exchanges A, C. C,B
．． A,B. A,D. Virtual paths forming links between D and B
) Band L1 divided into 1, 2, 3.4.5 respectively
, L2, L3, L4, and L5 will be explained. A logical line exists between the above-mentioned exchanges, and this line has certain bands LA, LB,
Lc and LD (communication capacity) are allocated. Bands LA, LBs t, Cs t, D and L1, L2, L3
- The relationship between L4 and L5 is such that calls such as the following are accommodated in the virtual path.

The virtual buses 1, 2.3.4.5 are managed by the communication resource allocation method of the packet switching network according to the present invention as follows.

That is, as shown in FIG. 8(b), each virtual path 1. 2. 3. 4. 5 is divided into non-overlapping bands and managed according to communication classes 11, 12, and 13 of communication accommodated by the line network. The band management method in this case is the same as in the above-described embodiment in which a buffer is used as a communication resource.

For example, when a communication request (class i) occurs between exchanges, request class 11 in the corresponding virtual path,
12. 13...i, it is determined whether this communication request can be accommodated in the band corresponding to i.

This embodiment describes band management when a single path is established from a source station to a destination station.

Therefore, even if a third exchange is placed in the middle between the exchanges, a band management configuration is described in which one path is established between the originating and destination exchanges by passing through the intermediate exchange.

Next, we will discuss an example of bandwidth management for each communication line between exchanges.
Shown in Figures (a) and (b).

In the circuit network of FIG. 9(a), adjacent exchanges A and B
．． Communication lines 21, 22, 23 between B, C and B, D;
24, 25, 26; 27, 28, and 29, the bands LAB+LBC+L and D are managed independently, and this is different from the above embodiment in which a buffer is used as a communication resource for each communication line. A similar method is used. For example, when a communication request occurs between exchange A and exchange D, the communication is to be performed via exchange B. At this time, it is determined whether or not this communication request can be accommodated in the communication lines LAB and LIID, and it is determined whether or not to permit the communication request. At that time, the bandwidth of LAB and LBD is managed for each communication class, and it is determined whether the communication making the communication request can be accommodated in the bandwidth of the corresponding class. Note that the bandwidth of the above-mentioned communication line is managed for each communication class as shown in FIG. 9(b).

Furthermore, the above-mentioned bandwidth management method for virtual paths (
An embodiment of a management method that combines FIG. 8) with bandwidth management for communications for each exchange will be described.

In this embodiment, as shown in FIG. 10, there is a virtual path between exchanges AC, and a communication line between exchanges AB is
AB. VC and L sc. vc (LAB.
VC-Lac. The same management method as in the case of the configuration shown in FIG. 8 is applied to the communication lines 21 and 22, and the communication lines 24 and 25 between the exchange BC. 1 and LBC. 1. Band LBD is provided for communication lines 27, 28; 29, 30 between exchanges BD, and band L AB. VC and L Bc. vc
(LAB.Vc =LBc.vc) as shown in Fig. 8(a). Applying the same management method as shown in (b), the band LAB. 1 and L BC. The same management method as shown in FIGS. 9(a) and 9(b) is applied to LBD.

The present invention is not limited to the above-mentioned embodiments, but can be implemented in other suitable embodiments by making appropriate design changes.

[Effects of the Fourth Effect] As described above, since the present invention is a communication resource allocation method for a packet switching network as described in the claims, the communication accommodated by the communication resource is uniformly allocated. Makes it easy to guarantee quality.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of a packet switch for explaining the outline of the present invention, Fig. 2 is an explanatory diagram showing its operation, and Fig. 3 is an example of the configuration of a manual boat section when priority control is not performed. FIG. 4 is an explanatory diagram showing an example of the configuration of the input boat unit when performing priority control, FIG. 5 is a flowchart of processing related to the static partitioning method, and FIG. 6 is a flowchart of processing related to the spare resource application method. Flowchart, FIG. 7 is a flowchart of processing related to the method of dynamically changing the boundary, 1
8 8 Figure (a). (b) is a diagram showing the structure of an embodiment when performing band management for virtual paths;
Figures (a) and (b) are diagrams showing the configuration of an embodiment when performing bandwidth management on communication lines for each exchange, and Figure 10 shows bandwidth management on virtual paths between exchanges and communication lines between exchanges. FIG. 11 is an explanatory diagram of a conventional packet switch, and FIG. 12 is an explanatory diagram showing a conventional resource allocation method. R...Total resources r1...Divided resources

Claims (5)

[Claims] (1) Communication resources such as the bandwidth or buffer capacity of communication lines in a packet switching network are divided into multiple independent resources for appropriately classified communications, and the communication resources accommodate each divided resource. A communication resource allocation method for a packet-switched network characterized by controlling the amount of communication so as to satisfy the required communication quality of the communication. (2) In the communication resource allocation method for a packet switching network according to claim (1), at least one of the divided resources is used as a reserve resource to supplement the shortage of resources in other divided resources. A communication resource allocation method for a packet switching network characterized by: (3) In the communication resource allocation method for a packet switching network according to claim (1), each divided resource mutually supplements the amount of other divided resources. Communication resource allocation method for switched networks. (4) In the communication resource allocation method for a packet-switched network according to claim (1), the communication accommodated by each divided resource is further divided into classes, and higher classes are prioritized over lower classes. A communication resource allocation method for a packet-switched network characterized by frequency control. (5) In the communication resource allocation method for a packet-switched network according to claim (1), resource division is performed according to communication quality requirements such as packet discard rate requirements, transfer delay requirements, and control information, and each divided resource is A communication resource allocation method for a packet-switched network characterized in that communications accommodated by the network are divided into classes, and priority control is performed to give priority to high-speed communications over low-speed communications.