GB2322763A

GB2322763A - Regulating cell intervals in an ATM network

Info

Publication number: GB2322763A
Application number: GB9725053A
Authority: GB
Inventors: Deog-Nyoun Kim
Original assignee: Daewoo Electronics Co Ltd
Current assignee: WiniaDaewoo Co Ltd
Priority date: 1996-12-26
Filing date: 1997-11-26
Publication date: 1998-09-02
Anticipated expiration: 2017-11-26
Also published as: KR19980053025A; KR100221323B1; GB2322763B; GB9725053D0; JPH10200551A

Abstract

A method for regulating cell intervals of a sequence of cells arriving in a connection of a network by assigning theoretical arrival times to the cells, in which the theoretical arrival time of a cell is determined using the average cell interval, the theoretical arrival time of a previous cell or its actual arrival time if and an interval between the actual and/or theoretical arrival times of two previous cells. If the actual arrival time of a previous cell is later than its theoretical arrival time that actual arrival time is used in place of its theoretical arrival time, in a manner which brings forward the theoretical arrival time and thus reduces the unnecessary delay imposed upon the current cell.

Description

METHOD FOR REGULATING A CELL INTERVAL IN ATM NETWORKS The present invention relates to a method for regulating cell intervals in ATM networks; and, more particularly, to a method capable of efficiently utilizing an allocated bandwidth by regulating the cell intervals based on calculated theoretical arrival times of cells.

ITU-TS(international telecommunication union telecommunication standardization sector) adopted an ATM(asynchronous transfer mode) as a basic technology in order to efficiently establish the B-ISDN(broadband - integrated service digital network). Since an ATM network statistically multiplexes and transmits information in unit of a limited size, i.e., packets or cells, it is possible to flexibly hold various services and efficiently use a bandwidth. However, if the ATM network does not provide a connection with a bandwidth corresponding to a peak bit rate thereof, the allocated bandwidth may be temporarily exceeded.

Consequently, the ATM network could fall into congestior especially in the case of a motion picture or high-speed data which are prone to burst. In terms of traffic control in the network, it may be desirable to prevent congestion before it happens by regulating input bit or cell rate from a signal source at the entrance of the network. In order to explain a conventional input bit rate regulation scheme, a connection establishment process is described first as follows.

First, a signal source which requires a connection establishment to a network make a request for a new connection by transmitting such traffic parameters as an average bit or cell rate from the signal source, required delay and delay jitter, etc, which represent statistical characteristics of the signal source and a QOS(quality of service) that the signal source expects from the network. Then, a connection admission control unit in the network examines whether the new connection can be established without depreciating the QOS of pre-established connections, based on a current condition of the network and the traffic parameters of the signal source which requires the connection establishment; and permits the connection if it is admissible.

During establishing the new connection, the network makes a contract with the signal source and sets up a path in order to link the signal source to its destination so that the new connection is established between the signal source and th destination. The contact specifies the QOS reserved for the connection and the characteristics of the signal source, e.g., the average cell rate or the average cell interval, upon which the QOS can be provided by the network. Once the connection is established, the signal source transmits cells to the network and a usage parameter control is carried out at the entrance of the network, wherein the usage parameter control signifies a process of regulating or shaping intervals of cells from the signal source according to the average cell interval specified in the contract when the signal source violates the contract.

Referring to Fig. 1, there is illustrated a network which includes a signal source 1, a usage parameter control(UPC) unit 2, a plurality of nodes A to D and a destination 7. A connection is established between the signal source 1 to the destination 7 through a path including the signal source 1, UPC unit 2, nodes A - D and the destination 7. The UPC unit 2 regulates cell intervals if the signal source 1 destroys the contract, wherein theoretical arrival times of cells are calculated and the cells are transmitted to a next node based on the calculated theoretical arrival times thereof.

Referring to Fig. 2, there is provided a timing chart illustrating a conventional usage parameter control scheme.

The upper timing diagram indicates actual arrival times of 5 cells at the UPC unit 2 shown in Fig. 1, the cells being transmitted from the signal source 1 to the UPC unit 2. The lower timing diagram shows theoretical arrival times of the 5 cells determined by the conventional UPC scheme at the UPC unit 2. The theoretical arrival times are calculated as follows: TAT(i) = max[TAT(i-l) + T, AAT(i)] ----- Eq. (1) wherein TAT(i) is a theoretical arrival time of an ith cell qj; T is an average cell interval or an average inter-arrival time of cells specified in the contract, wherein two adjacent cells in a connection are apart from each other with the average cell interval on the average so that a bandwidth allocated to the connection is not exceeded; AAT(i) is an actual arrival time of the ith cell qi at the UPC unit 2; and i is a positive integer index representing a sequence of cells.

With reference to Fig. 2, an actual arrival time AAT(1) of a 1st cell q1 is assumed to be 3; an actual arrival time AAT(2) of a 2nd cell q2, 4; an actual arrival time AAT(3) of a 3rd cell q3, 11; an actual arrival time AAT(4) of a 4th cell q4, 29; and an actual arrival time AAT(5) of a 5th cell q5, 30. The predetermined average cell interval T is assumed to be 6, and then the theoretical arrival time TAT(i) of the ith cell qj is calculated as follows: TAT(1) = max[0, 3] = 3; TAT(2) = max[3 + 6, 4] = 9; TAT(3) = max[9 + 6, 11] = 15; TAT(4) = max [15 + 6, 29] = 29; and TAT(5) = max[29 + 6, 30] = 35 wherein the value of [ TAT(0) + T ] is set to 0.

As can be seen from the above example, the theoretical arrival time TAT(S) of the 5th cell q5 is determined to be the sum of the theoretical arrival time TAT(4) of the 4th cell q4 and average cell interval T even though the 5th cell q5 need not be retarded as much as the average cell interval T because the interval between the theoretical arrival time of the 3rd cell q3 and the theoretical arrival time of the 4th cell q4 is considerably longer than the average cell interval T.

In other words, in the conventionaL UPC scheme a current cell interval is regulated simply based on an interval between a current and its previous cells without due regard to be paid to arrival times of preceding cells; and, as a result, theoretical arrival times of the cells may be unnecessarily retarded, resulting in inefficient usage of the allocated bandwidth.

It is, therefore, a primary object of the invention to provide a method for regulating cell intervals by calculating theoretical arrival times of the cells to thereby efficiently utilize an allocated bandwidth.

In accordance with the present invention, there is provided a method for regulating cell intervals of a sequence of cells arriving in a connection of a network by assigning theoretical arrival times to the cells, wherein the theoretical arrival times are calculated based on actual arrival times of the cells and a predetermined average cell interval, comprising the steps of: (a) detecting an actual arrival time of a current cell; (b) comparing a previous cell interval of the current cell with the average cell interval, wherein the previous cell interval is an interval between a theoretical arrival time of a previous cell which arrived in the connection just before the current cell and a theoretical arrival time of a preceding cell which arrived at the connection just before the previous cell; (c) determining a theoretical arrival time of the current cell based on a comparison result obtained in the step (b) by using the actual arrival time of the current cell, the average cell interval, the theoretical arrival time of the previous cell, and the previous cell interval; and (d) repeating the steps (a) to (c) for each cells arriving after the current cell.

The above and other objects and features of the present invention will become apparent from the following description of preferred embodiments given in conjunction with the accompanying drawings, in which: Fig. 1 depicts a conventional network including therein a usage parameter control unit for regulating cell intervals; Fig. 2 provides a timing chart illustrating a conventional usage parameter control scheme; Fig. 3 shows a flow chart for regulating a cell interval in accordance with a first embodiment of the present invention; Fig. 4 offers a flow chart for regulating a cell interval in accordance with a second embodiment of the present invention; Fig. 5 represents a timing chart illustrating cell interval regulation in accordance with the first embodiment of the present invention; and Fig. 6 depicts a timing chart illustrating cell interval regulation in accordance with the second embodiment of the present invention.

Referring to Figs. 3 and 4, there are illustrated usage parameter control schemes for regulating cell intervals in accordance with a first and a second preferred embodiments of the present invention, respectively.

The procedure in accordance with the first as well as th second embodiments of the present invention starts with step S1, wherein an initial value of i is set to 1, i being a positive integer index representing a sequence of cells; and a predetermined average cell interval T is inputted.

At step S2, an actual arrival time AAT(i) of the ith cell is is detected. At step S3, it is checked whether i is 1; and the procedure goes to step S4 if i equals to 1 or to steps S5 if otherwise. Since the value of i is 1 for the first cell, the procedure goes to step S4, wherein the actual arrival time AAT(1) of the 1st cell q1 is determined to be a theoretical arrival time TAT(1) thereof. Then the procedure goes to step S12. At step S12, the value of i is increased by 1, and theb value of a previous cell interval PreT(i) for an ith cell is set to [TAT(i-l)-TAT(i-2)], wherein for a second cell q2, TAT(i-2), i.e., TAT(0), is set to be 0.

At step S13, it is checked whether an end signal indicating the termination of cell transmission is inputted from a signal source or no more cell arrives for a predetermined time: if the end signal is inputted or no more cell is inputted for the predetermined time, the procedure terminates; if otherwise, the procedure goes back to step S2, wherein an actual arrival time AAT(i) of the ith cell qj is inputted. At step S3, if i is not 1, the procedure proceeds to step S5.

At step S5, the previous cell interval PreT(i) for the ith cell qj is compared with the average cell interval T. If the PreT(i) is smaller than or equal to T, the procedure goes to step S6, wherein it is checked if the sum of the theoretical arrival time TAT(i-1) of the (i-l)st cell qi1 and the average cell interval T is greater than the actual arrival time AAT(i) of the ith cell qj. If the checked result is affirmative, [TAT(i-1) + T] is determined to be the theoretical arrival time TAT(i) of the ith cell qj at step S8.

If the checked result at step S6 is negative, the procedure goes to step S9. At step S9 the actual arrival time AAT(i) of the ith cell qj is determined to be the theoretical arrival time TAT(i) thereof.

In accordance with the present invention, if, at step S5, the previous cell interval PreT(i) for the ith cell qj is found to be greater than the average cell interval T, the theoretical arrival time TAT(i) of the ith cell q is determined such that the average value of the PreT(i), i.e., [TAT(i-l) - TAT(i-2)] and the difference between TAT(i) and TAT(i-1), i.e., [TAT(i) - TAT(i-l)], rather than [TAT(i) TAT(i-1)] itsel,f is equal to or greater than the average cell interval T.

In accordance with the first embodiment of the present invention as shown in Fig. 3, at step S7-1, the theoretical arrival time TAT(i-1) of the (i-l)st cell will is added to a value which is obtained by subtracting the previous cell interval PreT(i) for the ith cell qj from twice the average cell interval T; and the added result is compared with the actual arrival time AAT(i) of the ith cell q . If the added result [TAT(i-1) + (2T - PreT(i)}] is decided to be larger than ART(i), the procedure goes to step S10-1, wherein the added result [TAT(i-1) + {2T - PreT(i)}] is selected as the theoretical arrival time TAT(i) of the ith cell qj. If the added result [TAT(i-1) + {2T - PreT(i)}] is determined not to be greater than AAT(i) at step S7-1, the procedure goes to step Sli, wherein AAT(i) is set to be the theoretical arrival time TAT(i) of the ith cell In the second embodiment of the present invention as shown in Fig. 4, step S7-2 replaces steps S7-1 of the first embodiment shown in Fig. 3, wherein step S7-2 is composed of two substeps S71 and S72. A cell interval parameter K(i) for the ith cell qi is calculated at substep S71, by dividing the previous cell interval PreT(i) for the ith cell qj by the average cell interval T. Then, at substep S72, the theoretical arrival time TAT(i-1) of the (i-l)st cell q; t is added to a value which is obtained by dividing the average cell interval T by the cell interval parameter K(i); and the added result is compared with the actual arrival time AAT(i).

If the added result [TAT(i-1) + T/K(i)] is larger than the actual arrival time AAT(i) of the ith cell qi, the procedure goes to step S10-2, wherein the added result [TAT(i-1) + T/K(i)] is selected as the theoretical arrival time TAT(i) of the ith cell q . If the added result [TAT(i-1) + T/K(i)] is smaller than or equal to AAT(i), the procedure goes to step Sll. At step S11, the actual arrival time AAT(i) of the ith cell qj is selected as the theoretical arrival time TAT(i) of the ith cell In accordance with the preferred embodiments of the present invention, the previous cell interval PreT(i) for the ith cell qi, i.e., [TAT(i-1) - TAT(i-2)], as well as the theoretical arrival time TAT(i-1) of the (i-l)st cell qi1 is taken into account in calculating the theoretical arrival time TAT(i) of the ith cell. When the previous cell interval PreT(i) is considerably longer than the average cell interval T, the theoretical arrival time TAT(i) of the ith cell qi is determined such that an interval between the (i-2)nd cell q; 2 and the ith cell qi is greater than or equal to twice the average cell interval T.

Specifically, when the previous cell interval PreT(i) is longer than the average cell interval T, the theoretical arrival time TAT(i) calculated in accordance with the first embodiment of the present invention is given as: TAT(i) = max[TAT(i-1)+(2T-PreT(i)), AAT(i)] = max[iTAT(i-2)+PreT(i))+(2T-PreT(i)), AAT(i)] = max[TAT(i-2)+2T, AAT(i)].

----- Eq. (2) Since the theoretical arrival time TAT(i) is given by the larger one of [TAT(i-2)+2T] and AAT(i), the interval between the (i-2)nd cell q; 2 and the ith cell qi is greater than or equal to twice the average cell interval T.

Further in accordance with the second embodiment of the present invention, when the previous cell interval PreT(i) is greater than the average cell interval T, the theoretical arrival time TAT(i) may be defined as: TAT(i) = max[TAT(i-l)+T/K(i), AAT(i)] = max[{TAT(i-2)+PreT(i)}+T/K(i), AAT(i)] = max[(TAT(i-2)+K(i).T}+T/K(i), AAT(i)] = max[TAT(i-2)+(K(i)+1/K(i)}eT, AAT(i)].

----- Eq. (3) Because K(i) is positive, [K(i)+ 1/K(i)] is greater than or equal to 2, which in turn gives rise to [TAT(i-2) {K(i)+l/K(i)}T] greater than or equal to [TAT(i-2)+2T].

Since the theoretical arrival time TAT(i) is given by the larger one of [TAT(i-2)+fK(i)+l/K(i)}oT] and AAT(i), the interval between the (i-2)nd cell qi-2 and the ith cell qj is greater than or equal to twice the average cell interval T.

Consequently, the theoretical arrival time TAT(i) of the ith cell q, is determined such that the interval between the (i- 2)nd cell qi-2 and the ith cell qj is greater than or equal to twice the average cell interval T.

After the theoretical arrival time TAT(i) of the ith cell qj is determined in accordance with the first or the second embodiment of the present invention, the procedure goes to step S12, wherein i is increased by 1 and the previous cell interval PreT(i) for the ith cell qj is updated with a value obtained by subtracting the theoretical arrival time TAT(i-2) for the (i-2)nd cell qiz from the theoretical arrival time TAT(i-l) of the (i-l)st cell qj1. At a next step S13, it is checked if the end signal is inputted or no more cell is inputted for the predetermined time, wherein the end signal indicates that there are no more cells to be transmitted. If the end signal or no more cell is inputted, the procedure ends; and if otherwise, the procedure will be repeated by returning to step S2.

Consider the case shown in Fig. 2 where the AAT(1) is 3; the AAT(2), 4; the AAT(3), 11; the AAT(4), 29; and the AAT(S), 30 with T and TAT(0) being set to 6 and 0, respectively.

Theoretical arrival times TAT(i)'s of the ith cell calculated in accordance with the preferred embodiments of the present invention are depicted in Figs. 5 and 6.

Referring to Fig. 5, there is depicted a timing chart illustrating theoretical arrival times TAT's calculated in accordance with the first embodiment of the present invention, wherein the upper and the lower diagrams show actual arrival and theoretical arrival times AAT's and TAT's, respectively.

Theoretical arrival times of the cells are calculated with reference to Fig. 3 as follows: for i = 1, TAT(1) = AAT(1) = 3; for i = 2, PreT(2) = TAT(1) - TAT(O) = 3 - 0 = 3 < T = 6, TAT(1) + T = 9 > AAT(2) = 4, TAT(2) = TAT(1) + T = 9; for i = 3, PreT(3) = TAT(2) - TAT(1) = 6 = TAT(2) + T = 15 > AAT(3) = 11, TAT(3) = TAT(2) + T = 15; for i = 4, PreT(4) = TAT(3) - TAT(2) = 6 = TAT(3) + T = 21 < AAT(4) = 29, TAT(4) = AAT(4) = 29; and for i = 5, PreT(S) = TAT(4) - TAT(3) = 14 > T = 6, TAT(4) + {2T-PreT(S)) = 29 + (2(6 - 14) = 27 < AAT(S) = 30, TAT(S) = TAT(4) + (2T-PreT(S)) = 30.

Referring to Fig. 6, there is represented a timing chart.

depicting cell interval regulation in accordance with the second embodiment of the present invention wherein the upper and the lower diagrams show actual arrival and theoretical arrival times AAT's and TAT's, respectively. Theoretical arrival times of the cells are calculated in accordance with Fig. 4 as follows: for i = 1, TAT(1) = AAT(1) = 3; for i = 2, PreT(2) = TAT(1) - TAT(O) = 3 - 0 = 3 < T = 6, TAT(1) + T = 9 > AAT(2) = 4, TAT(2) = TAT(1) + T = 9; for i = 3, PreT(3) = TAT(2) - TAT(1) = 6 = TAT(2) + T = 15 > AAT(3) = 11, TAT(3) = TAT(2) + T = 15; for i = 4, PreT(4) = TAT(3) - TAT(2) = 6 = TAT(3) + T = 21 < AAT(4) = 29, TAT(4) = AAT(4) = 29; and for i = 5, PreT(S) = TAT(4) - TAT(3) = 14 > T = 6, K(5) = PreT(5)/T = 14/6, TAT(4) + {l/K(i)}.T = 29 + (6/14)-6 = 31.57 > AAT(5) = 30, TAT(S) = TAT(4) + {1/K(i)}.T = 31.57.

Although the theoretical arrival time TAT(S) of the 5th cell q5 in accordance with the second embodiment of the present invention is about 31.57, the 5th cell q5 is considered to arrive at 32 since cells are transmitted in time slots, wherein a time slot corresponds to a time during which one cell is transmitted and is taken as a time unit.

Consequently, when the previous cell interval PreT(S) for the 5th cell q5, i.e., a theoretical arrival time interval between the 3rd cell q3 and the 4th cell q4, is considerably greater than the average cell interval T, a theoretical arrival time interval between the 4th cell q4 and the 5th cell qS in accordance with the first and the second embodiments of the present invention is determined to be 1 and 2.57 respectively, both of which are much smaller than the average cell interval T = 6. However, the interval between the 3rd cell q3 and the 5th cell qS in accordance with the first and the second embodiments of the present invention becomes 15 and 16.57 respectively, both of which are greater than twice the average cell interval 2T = 12. Therefore, the allocated bandwidth of a connection can be utilized without suffering an unnecessary delay.

While the present invention has been described with respect to certain preferred embodiments only, other modifications and variations may be made without departing; from the scope of the present invention as set forth in the following claims.

Claims

Claims:

1. A method for regulating cell intervals of a sequence of cells arriving in a connection of a network by assigning theoretical arrival times to the cells, wherein the theoretical arrival times are calculated based on actual arrival times of the cells and a predetermined average cell interval, comprising the steps of: (a) detecting an actual arrival time of a current cell; (b) comparing a previous cell interval of the current cell with the average cell interval, wherein the previous cell interval is an interval between a theoretical arrival time of a previous cell which arrived in the connection just before the current cell and a theoretical arrival time of a preceding cell which arrived at the connection just before the previous cell; (c) determining a theoretical arrival time of the current cell based on a comparison result obtained in the step (b) by using the actual arrival time of the current cell, the average cell interval, the theoretical arrival time of the previous cell, and the previous cell interval; and (d) repeating the steps (a) to (c) for each cell arriving after the current cell.

2. The method as recited in claim 1, wherein an actual arrival time of the first cell in the sequence is set to a theoretical arrival time thereof.

3. The method as recited in claim 2, wherein a theoretical arrival time for a preceding cell of the second cell in the sequence is set to 0.

4. The method as recited in claim 3, wherein the determining step (c) includes the steps of: (cl) if the previous cell interval is smaller than or equal to the average cell interval, comparing a sum of the theoretical arrival time of the previous cell and the average cell interval with the actual arrival time of the current cell to select a larger one of the two as the theoretical arrival time of the current cell; and (c2) if the previous cell interval is greater than the average cell interval, determining the theoretical arrival time of the current cell such that an interval between the preceding cell and the current cell is greater than twice the average cell interval.

5. The method as recited in claim 4, wherein the step (c2) includes the steps of: (c21) comparing a sum of the theoretical arrival time of the previous cell and a value obtained by subtracting the previous cell interval from twice the average cell interval with the actual arrival time of the current cell; and (c23) selecting a larger one of the two compared values at the step (c21) as the theoretical arrival time of the current cell.

6. The method as recited in claim 4, wherein the step (c2) includes the steps of: (c22) defining a cell interval parameter by dividing the previous cell interval by the average cell interval; (c24) comparing a sum of the theoretical arrival time of the previous cell and a value obtained by dividing the average cell interval by the cell interval parameter with the actual arrival time of the current cell; and (c26) selecting a larger one of the two compared value at the step (c24) as the theoretical arrival time of the current cell.

7. A method for regulating cell intervals of a sequence of cells arriving in a connection of a network by determining theoretical arrival times for the cells, wherein the theoretical arrival times are calculated based on actual arrival times of the cells and a predetermined average cell interval, comprising the steps of: (a) detecting an actual arrival time of a first cell of the sequence; (b) setting the actual arrival time of the first cel as a theoretical arrival time thereof; (c) detecting an actual arrival time of a next cell of the sequence; (d) comparing a previous cell interval of the next cell with the average cell interval, wherein the previous cell interval is an interval between a theoretical arrival time of a previous cell which arrived in the connection just before the next cell and a theoretical arrival time of a preceding cell which arrived at the connection just before the previous cell, a theoretical arrival time for a preceding cell for the second cell in the sequence being set to 0; (e) determining a theoretical arrival time of the next cell based on a comparison result obtained in the step (d) by using the actual arrival time of the next cell, the average cell interval, the theoretical arrival time of the previous cell, and the previous cell interval; and (f) subjecting each of the remaining cells in th sequence to said steps (c) to (e).

8. The method as recited in claim 7, wherein the determining step (e) includes the steps of: (el) if the previous cell interval is smaller than o equal to the average cell interval, comparing a sum of the theoretical arrival time of the previous cell and the average cell interval with the actual arrival time of the next cell; (e2) setting, as the theoretical arrival time of the next: cell, the sum if the sum is greater than the actual arrival time of the next cell and the actual arrival time of the next cell if otherwise; and (e3) if the previous cell interval is greater than the average cell interval, determining the theoretical arrival time of the next cell such that an interval between the preceding cell and the next cell is greater than twice the average cell interval.

9. The method as recited in claim 8, wherein the step (e3) includes the steps of: (e31) comparing a sum of the theoretical arrival time of the previous cell and a value obtained by subtracting the previous cell interval from twice the average cell interval with the actual arrival time of the next cell; and (e33) selecting a not smaller one of the two compared values at the step (e31) as the theoretical'arrival time of the next cell.

10. The method as recited in claim 8, wherein the step (e3) includes the steps of: (e32) defining a cell interval parameter by dividing the previous cell interval by the average cell interval; (e34) comparing a sum of the theoretical arrival time of the previous cell and a value obtained by dividing the average cell interval by the cell interval parameter with the actual arrival time of the next cell; and (e36) selecting a not smaller one of the two compared value at the step (e34) as the theoretical arrival time of the next cell.

11. A method for regulating cell intervals substantially as hereinbefore described with reference to and shown in the accompanying Figures 3-6.