JP3298172B2 - Access arbitration method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is applied to an information processing apparatus, a communication exchange, in particular, an ATM exchange, and other systems having a plurality of resources commonly connected by a bus.
Here, the resource refers to a memory device, an output device, a display device, an input device, a printing device, a functional unit, and other facilities including hardware for operating the system.

[0002] The present invention relates to arbitration of a plurality of access requests generated within a system with respect to a plurality of shared resources arranged in one system.

[0003]

2. Description of the Related Art The arbitration of an access request according to the prior art will be described with reference to a model shown in FIG. In this model, there are eight access request sources R (1) to R (8) and eight resources S (1) to S (8) for receiving the access requests.
These are connected to the input terminal group and the output terminal group of the matrix switch means, respectively. The matrix switch means may be provided with hardware having a matrix configuration as shown in the figure, or may be a virtual one constituted by a logic circuit having a matrix function as shown and capable of transmitting an access request. The number of access request sources and the number of resources need not necessarily be equal.

The access request sources R (1) to R (8) generate access requests one after another. Since an access request is generated according to the circumstances of the source, a plurality of access requests may simultaneously compete for one resource. The contention is arbitrated by a control circuit that controls the matrix switch means, and the access request is transmitted from one access request generation source to one request destination resource. At this time, each of the ring arbiters RA (1) to RA (8) sweeps an access request as shown by an arrow in FIG. 6 and, when the corresponding access request is reached, via the intersection C (n, m). Only one access request is transmitted to the resource side, and the resource accepts the access request.

When one access request is accepted,
Information indicating that the request has been accepted is sent to the request source that sent the access request. By receiving one access request, another access request sent to the ring arbiter is waited, and the other access request is detected at the next timing or at the next timing. After the access request is accepted, processing such as transfer of necessary data via the matrix switch means or another bus signal line is executed.

[0006]

In such contention arbitration of access requests, it is good when the number of access request sources and the number of resources are small, but as the number increases, the number of contention arbitrations increases. N is N + ΔN, M is M +
By increasing to ΔM, the size of the matrix switch is N × M, but becomes (N + ΔN) × (M + ΔM) = NM + ΔNM + NΔM + ΔNΔM. The increase in the size of the matrix switch directly leads to an increase in the contention arbitration operation by the ring arbiter. That is, in FIG. 4, the request source R (1) and the resource S (8)
In order to establish a connection with the ring arbiter, each of the ring arbiters must perform contention arbitration at eight intersections. Further, as the size of the matrix switch increases, the number of intersections that the ring arbiter should go around increases. In addition, it is necessary to transmit and receive connection availability information for each access request. That is, the operation time becomes long, and expensive elements that can execute high-speed operation must be used.

An object of the present invention is to improve the contention arbitration by changing the logic of contention arbitration and apparently reducing the number of access request sources or the number of resources to reduce the number of contention arbitrations to be executed. And SUMMARY OF THE INVENTION It is an object of the present invention to provide a method capable of shortening the contention arbitration time for access request arbitration and performing the contention arbitration with a low-speed element.

[0008]

According to the present invention, there are provided N (N is an integer of 2 or more) request sources for generating access requests;
M resources (M is an integer of 2 or more) shared by the access request sources and N request sources connected to the input terminal, and N resources connected to the output terminal are connected to the N resources. Matrix switch means having a capacity of × M;
The present invention is applied to an apparatus having a control circuit for controlling the connection / disconnection of the intersection of the matrix switch means.

The access request sources are resources in a broad sense, and may be one of the M resources depending on the device configuration. The matrix switch means may be a matrix switch in which switches composed of hardware are arranged at intersections of a matrix, and a virtual circuit that substantially realizes a circuit for connecting or disconnecting an access request at the intersection by a program control circuit. Matrix switch.

The features of the present invention are as follows. A control circuit for controlling the connection / disconnection of the intersection of the matrix switch means, the arbitration means for executing arbitration when an access request is issued from a plurality of the access request sources for one of the resources, and the arbitration A connection means for controlling the connection / disconnection of the intersection of the matrix switch means according to the result of the above, wherein the arbitration means of the control circuit has a plurality of intersections of the matrix switch means smaller than the number (N × M); Means for at least partially dividing the output terminal into groups including a plurality of intersections, means for focusing on one of the divided groups, and executing the arbitration within the focused group, Means for circulating all of the divided groups over time.

In the means for dividing into groups, the logic of the division can be changed according to the frequency of occurrence of the access request, and the logic of the division can be changed according to the priority of the access request. Can also.

[0012]

According to the method of the present invention, the range in which the contention arbitration is performed is divided into small groups, and the contention arbitration is executed by parallel processing for each group. The number of transfers for contention arbitration is reduced.

Therefore, as the size of the matrix switch increases, the control time required for contention arbitration is at worst the same as in the conventional example, and is further improved.

Also, priorities can be set according to the frequency of occurrence of access, grouping can be changed, and priorities can be assigned to each group by an external priority control signal.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The configuration of a first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a configuration diagram of the first embodiment of the present invention.

According to the present invention, eight request sources R (1) to R (8) for generating access requests and eight request sources R (1) to R (8) shared by the access request sources are provided. Resources S (1) ~
S (8) and eight request sources R (1) to R (8) are connected to input terminals ₁₁ to ₁₈ and eight resources S (1) to S (8)
Matrix switch and the resource S (1) to S one request source of access for the (8) R (1) ~R (8 having but the capacity of the output terminal 2 ₁ to 2 ₈ is connected to the 8 × 8 ), An arbitration unit 5 that performs arbitration when an access request is issued from a plurality of the above, and a matrix switch that is a disconnection unit that controls disconnection of an intersection of the matrix switch 3 based on the result of the arbitration. An access arbitration system including a control circuit 4 including a disconnection unit 6.

Here, the feature of the present invention is that the arbitration means of the control circuit 4 has a plurality of intersections of the matrix switches 3 smaller than the number (8 × 8), and at least a part of the output terminals 2 About ₂₁ to ₈ means for dividing into groups including a plurality of intersection points G (1,1) ~G (4,4) , one of the divided groups G (1,1) ~G (4,4) Means for executing the arbitration in the groups G (1,1) to G (4,4) focused on one of the groups G (1,1) to G (4,4)
Means for circulating 4) over time for all of the divided groups G (1,1) to G (4,4). In the first embodiment of the present invention, the circulation is performed in synchronization with the transmission timing of the time slot.

Next, the operation of the first embodiment of the present invention will be described with reference to FIG. FIG. 2 is a flowchart showing the operation of the first embodiment of the present invention. The intersections of the matrix switches 3 are divided into groups G (1,1) to G (4,4),
When (1,1) to G (4,4) are represented by a set of intersections C (1,1) to C (8,1), C (1,1), C (1,2), C (2 , 1), C (2,2) ⇔G (1,1) C (1,3), C (1,4), C (2,3), C (2,4) ⇔G (1,2) ) C (1,5), C (1,6), C (2,5), C (2,6) ⇔G (1,3) C (1,7), C (1,8), C (2,7), C (2,8) ⇔G (1,4) C (3,1), C (3,2), C (4,1), C (4,2) ⇔G (2 , 1) C (3,3), C (3,4), C (4,3), C (4,4) ⇔G (2,2) C (3,5), C (3,6) , C (4,5), C (4,6) ⇔G (2,3) C (3,7), C (3,8), C (4,7), C (4,8) ⇔G (2,4) C (5,1), C (5,2), C (6,1), C (6,2) ⇔G (3,1) C (5,3), C (5,2) 4), C (6,3), C (6,4) ⇔G (3,2) C (5,5), C (5,6), C (6,5), C (6,6) ⇔G (3,3) C (5,7), C (5,8), C (6,7), C (6,8) ⇔G (3,4) C (7,1), C ( 7,2), C (8,1), C (8,2) ⇔G (4,1) C (7,3), C (7,4), C (8,3), C (8,2) 4) ⇔G (4,2) C (7,5), C (7,6), C (8,5), C (8,6) ⇔G (4,3) C (7,7), C (7,8), C (8,7), C (8,8) ⇔G (4,4) (S 1). Of these, groups G (1,1) to G (4,4) that can arbitrate for competition at one access opportunity are G (i1, j1)
, G (i2, j2), G (i3, j3), G (i4, j4) [i1 ≠ i2
{I3Ｇi4 and j1 ≠ j2 ≠ j3 ≠ j4] are four groups G (1,1) to G (4,4). These are referred to as arbitration right groups G (i1, j1) to G (i4, j4) in distinction from other groups G (1,1) to G (4,4). The arbitration right groups G (i1, j1) to G (i4, j4) in FIG. 1 are groups G (1,1) to G (4,4) surrounded by solid lines, and R (1), R (2) ), ⇔S (3), S (4) [Arbitration Group G
(1,2)] R (3), R (4), ΔS (5), S (6) [Arbitration group G
(2,3)] R (5), R (6), ΔS (7), S (8) [Arbitration group G
(3,4)] R (7), R (8), ΔS (1), S (2) [Arbitration right group G
(4,1)] (S2). This indicates a range in which competitive arbitration is performed at a certain opportunity. For example, one arbitration right group G (3,4) has resources S (1) to S (8) which may be accessible to request sources R (5) and R (6) are resources S (1) to S (8). 7) or S (8). Arbitration right group G (i1, j1) within one combined group
Each of the groups G (1,1) to G that constitute G (i4, j4)
Mixing and arbitration are performed in parallel for (4,4) (S3).
Here, in comparison with the conventional example, the contention arbitration for the resource S (8) needs to be performed between all the request sources R (1) to R (8), so that a large-scale control data transfer is required. Although it was necessary to select an access candidate, in the case of the first embodiment of the present invention, only contention arbitration between the request sources R (5) and R (6) is required. That is, the arbitration range is 4 compared to the conventional example.
Each arbitration right group G (i1, j1) to G (i4, j
4) is performing parallel arbitration, which can speed up the arbitration.

While the groups G (1,1) to G (4,4) are fixed and the combination of the arbitration right groups G (i1, j1) to G (i4, j4) is circulated, any request source R (1)-R (8)
Can access any of the resources S (1) to S (8).

The arbitration right groups G (i1, j1) to G (i4,
When the competitive arbitration in j4) is completed (S4), the following arbitration groups G (i1, j1) to G (i4, j4) are set as R (1), R (2), ⇔S (1), S (2) [Arbitration Group G
(1,1)] R (3), R (4), ⇔S (3), S (4) [Arbitration group G
(2,2)] R (5), R (6), ΔS (5), S (6) [Arbitration group G
(3,3)] R (7), R (8), ΔS (7), S (8) [Arbitration right group G
(4,4)], a competitive arbitration is performed in the combination of the arbitration right groups G (i1, j1) to G (i4, j4) (S5), and the next arbitration right group G (i1, j1) To G (i4, j4). All combinations are circulated (S6), and the competition arbitration ends.

Next, a second embodiment of the present invention will be described with reference to FIG. FIG. 3 is a configuration diagram of the second embodiment of the present invention. The feature of the second embodiment of the present invention is that the group G (i, j)
i = 1,2,3,4, j = 1,2,3,4], and monitoring data CL (i) of CT (i) or an external control signal. An arbitration right group selection unit GCT for changing the grouping of groups G (1,1) to G (4,4) and controlling priority by PR is provided.
Upon receiving the monitoring data CL (i), the arbitration right group selecting unit GCT determines the groups G (1,1) to G based on the frequency of occurrence.
Prioritize (4,4). Alternatively, the grouping is changed. For example, when the arbitration right group selection unit GCT detects that the request sources R (1) to R (4) and the resources S (1) to S (4) are frequently accessed, the group G ( A priority is assigned to (1,1), G (1,2), G (2,1), G (2,2), or a group G (1,1) as shown by a dashed line in FIG.
, G (1,2), G (2,1), G (2,2) to form a group LG, and the request sources R (1) to R (4) and the resource S
It corresponds to frequent access between (1) and S (4). If a large grouping such as the group LG is used, contention arbitration in the group can be performed frequently, and the number of data transfers for contention control decreases.
The round time of G (4,4) can also be reduced. Also, the arbitration right group selecting unit G can externally set the priority by the external control signal PR.
Can be input to CT.

Here, a procedure for assigning priorities to the groups G (1,1) to G (4,4) will be described. Arbitration right groups G (i1, j1) to G (i4, j4) in FIG. 3 are portions surrounded by solid lines, and R (1), R (2), ⇔S (3), S (4) [Mediation Group G
(1,2)] R (3), R (4), ΔS (5), S (6) [Arbitration group G
(2,3)] R (5), R (6), ΔS (7), S (8) [Arbitration group G
(3,4)] R (7), R (8), ΔS (1), S (2) [Arbitration right group G
(4,1)]. This indicates a range in which competitive arbitration is performed at one opportunity. For example, one mediation right group G (3,4)
Are the resources S (1) to S (8) that the request sources R (5) and R (6) may be able to access, and the resources S (7)
Alternatively, it is indicated that S (8). The arbitration right groups G (i1, j1) to G (i4, j4) sequentially circulate as described in the first embodiment of the present invention to perform competitive arbitration.

In the second embodiment of the present invention, the monitoring unit CT (i) receives the monitoring data CL (i) from the request sources R (1) to R (8) and transfers it to the arbitration right group selecting unit GCT. . In the arbitration right group selection unit GCT, the monitoring data CL
(i) is deciphered, and arbitration right groups G (i1, j1) to G (i4, j4) to be given the highest priority are set from the access occurrence frequency. Also, the arbitration right group G (i1, j1)
ＧG (i4, j4) can be arbitrarily assigned a priority.

Here, if the priority of the access requests to the resources S (3) and S (4) of the request sources R (1) and R (2) is more prominent than the others, the arbitration right group Selector GC
T is R (1), R (2), ⇔S (3), S (4) [arbitration group G
(1,2)] ... (priority) is suspended. As a result, the next mediation right group G (i1,
j1) to G (i4, j4) are originally R (1), R (2), ⇔S (1), S (2) [arbitration right group G
(1,1)] R (3), R (4), ⇔S (3), S (4) [Arbitration group G
(2,2)] R (5), R (6), ΔS (5), S (6) [Arbitration group G
(3,3)] R (7), R (8), ΔS (7), S (8) [Arbitration right group G
(4,4)] is represented by R (1), R (2), {S (3), S (4) [arbitration right group G (1,2)] ... (priority) R (3), R (4), ⇔S (1), S (2) [Arbitration group G (2,1)] R (5), R (6), ⇔S (5), S (6) [Arbitration group G (3,3)] R (7), R (8), {S (7), S (8) [arbitration right group G (4,4)]. This is the first priority of arbitration group G (1,2)
In this example, the arbitration right group G (2,3), which should originally be transferred to the group G (2,2), is caused to jump to the group G (2,1). In this manner, the positions of the highest-priority arbitration right groups G (i1, j1) to G (i4, j4) are determined first, and the other arbitration right groups G (i, j4) are logically inconsistent with each other. i1, j1) to G (i4, j4) are determined.

As another example, the request source R (3)
If the priority of the access request to the resources S (1) and S (2) of R (4) is more prominent than others, the arbitration right group selecting unit GCT determines R (3) and R (4) , ⇔S (1), S (2) [Arbitration group G
(2,1)]… (priority), R (1), R (2), ⇔S (1), S (2) [arbitration right group G
(1,1)] R (3), R (4), ⇔S (3), S (4) [Arbitration group G
(2,2)] R (5), R (6), ΔS (5), S (6) [Arbitration group G
(3,3)] R (7), R (8), ΔS (7), S (8) [Arbitration right group G
(4,4)] where R (1), R (2), ⇔S (7), S (8) [arbitration group G (1,4)] R (3), R (4) , ⇔S (1), S (2) [arbitration right group G (2,1)] ... (priority) R (5), R (6), ⇔S (5), S (6) [arbitration right group G (3,3)] R (7), R (8), ⇔S (3), S (4) [Arbitration group G (4,2)] from group G (2,3) The arbitration right groups G (i1, j1) to G (i4, j4) are jumped to the group G (2,1). Other mediation groups G (1,4), G (3,3), G (4,2)
Are determined sequentially after the arbitration right group G (2,1) is determined.

If the highest priority of access to the resources S (7) and S (8) of the request sources R (5) and R (6) is input by the external control signal PR, the arbitration right group is selected. Part G
CT: R (5), R (6), ΔS (7), S (8) [Arbitration group G
(3,4)]… (priority) is suspended. As a result, the next mediation right group G (i1,
j1) to G (i4, j4) are originally R (1), R (2), ⇔S (1), S (2) [arbitration right group G
(1,1)] R (3), R (4), ⇔S (3), S (4) [Arbitration group G
(2,2)] R (5), R (6), ΔS (5), S (6) [Arbitration group G
(3,3)] R (7), R (8), ΔS (7), S (8) [Arbitration right group G
(4,4)] where R (1), R (2), ⇔S (1), S (2) [arbitration group G
(1,1)] R (3), R (4), ⇔S (3), S (4) [Arbitration group G
(2,2)] R (5), R (6), ΔS (7), S (8) [Arbitration right group G
(3,4)] R (7), R (8), ΔS (5), S (6) [Arbitration right group G
(4,3)].

In the first and second embodiments of the present invention, eight request sources R (1) to R (8), eight resources S (1) to S (8), and eight request sources R (1) ) To R (8) is 4, the number of request sources R (1) to R (8) therein is 2, and the resource S (1)
ＳS (8) is divided into four resources S (1) to S (8)
Although the number is set to 2, these numbers are not limited, and may be configured with an arbitrary number and a group scale.

Further, a configuration can also be adopted in which priority can be set between a plurality of request sources R (1) to R (8) and a plurality of resources S (1) to S (8).

[0029]

As described above, according to the present invention, the number of contention arbitrations to be executed can be reduced by changing the logic of contention arbitration and apparently reducing the number of access request sources or the number of resources. According to the present invention, the contention arbitration time for access request arbitration can be shortened, and the contention arbitration can be performed by a low-speed element.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a first embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the first embodiment of the present invention.

FIG. 3 is a configuration diagram of a second embodiment of the present invention.

FIG. 4 is a configuration diagram of a conventional example.

[Explanation of symbols]

1 ₁ to 1 ₈ input terminals 2 ₁ to 2 ₈ output terminals 3 matrix switch 4 control circuit 5 arbitration unit 6 matrix switch disconnection unit LG group R (1) to R (8) Request generation sources S (1) to S ( 8) Resource Xp intersection

────────────────────────────────────────────────── ─── Continuing from the front page (56) References Kenichi Endo (3 others), High-speed ATM switch configuration method using hierarchical arbitration, IEICE Technical Report, SSE92-104, 67-72, 001, CSNG199900678011 ( 58) Field surveyed (Int. Cl. ⁷ , DB name) H04L 12/56 G06F 13/362 510

Claims (3)

(57) [Claims] 1. N (N is an integer of 2 or more) request sources that generate access requests, M (M is an integer of 2 or more) resources shared by the access request sources, Matrix switch means having an N × M capacity, wherein N request sources are connected to an input terminal, and the M resources are connected to an output terminal; and a plurality of the access request sources for one of the resources. Arbitration means for executing arbitration when an access request is issued from the network, and a control circuit including disconnection means for controlling disconnection of an intersection of the matrix switch means based on the arbitration result. In the arbitration means of the control circuit, the number of intersections of the matrix switch means (N × M)
Means for dividing the output terminal into groups each including a plurality of intersections at least partially, and means for focusing on one of the divided groups and executing the arbitration within the focused group And means for circulating the group of interest for all of the divided groups over time. 2. The access arbitration system according to claim 1, wherein said means for dividing into groups includes means for changing the logic of the division in accordance with the frequency of occurrence of access requests. 3. The access arbitration system according to claim 1, wherein said means for dividing into groups includes means for changing the logic of the division according to the degree of priority of an access request.