Multiplex communications – Data flow congestion prevention or control – Control of data admission to the network
Reexamination Certificate
2000-02-16
2004-08-31
Kizou, Hassan (Department: 2662)
Multiplex communications
Data flow congestion prevention or control
Control of data admission to the network
C370S412000, C370S429000
Reexamination Certificate
active
06785235
ABSTRACT:
The present invention relates to a data communication system, and more specifically to frame relay multiplexing in such a system, and more specifically to priority controller for frame relay multiplexing.
TERMS AND DEFINITIONS
In order to better understand the technical background of the present invention, the following terms and definitions should be observed:
LAPD
Link Access Protocol D channel ref.:
ITU Q.921
OSI model
Open System Interconnection Model
divide a communication process in
layer 1-7
Physical layer
The lowest OSI model layer 1 respon-
sible for activation, maintaining and
deactivation of physical circuits
Data link layer
The OSI model layer 2 responsible for
the transfer of data across a link
with recovery from lost data and data
with errors
Frame relay
Data communication where just the
lower part of layer 2 is performed in
nodes between the nodes with full
layer 2 termination with error recovery.
TECHNICAL BACKGROUND
In data communication frame relay is often used. In this case the data frames are forwarded by many nodes before the layer
2
protocol is terminated. If one of these nodes does a multiplexing and concentration of the traffic from many sources, there is a possibility of communication overflow in some time intervals. If the sum of the data rates from the sources is greater than in the destination direction, a frame queue will build up at the output port. The frames may be delayed too much in this queue to be acknowledged before the retransmission timer of the layer
2
source expires.
THE PROBLEM AREA
The layer
2
protocol like LAPD has by default a retransmission time of T
1
second. The sum of the transmission time, the queue delay and transmission time of the acknowledge frame must not exceed these T
1
seconds. If this sum is greater than T
1
, the frame will be retransmitted from the source. The frame will probably be placed in the queue once more. This in turn means that traffic from other sources may be delayed too much and the useful throughput will decrease.
PRIOR ART AND KNOWN SOLUTIONS
One way to solve this problem will be to measure the real output queue and give all sources the same probability to enter the queue based on a statistical measurement of the sources. This solution requires a lot of computations.
From U.S. Pat. No. 5,231,633 (Hluchyj et al.) is known a method for prioritising, selectively discarding and multiplexing different traffic type fast packets, and more specifically a queuing and dequeuing mechanism for use in an integrated fast packet network.
According to this prior art fast packets from differing traffic types are multiplexed with one another through use of a weighted round-robin bandwidth allocation mechanism. Although this prior art technique has to do with packet communication and multiplexing associated therewith, this prior art is silent as regards any suggestion of using fuzzy logic for the decision of when a data frame has to be deleted from the data stream in question.
OBJECTS OF THE INVENTIONS
The main object of the present invention is to provide an appropriate solution to the queuing of data frames in a communication network.
Another object of the present invention is to provide a method by which data frames which by fuzzy logic are estimated to have a small possibility to be transmitted in time, are deleted and not sent to the output queue, preferably by letting frames from sources with high traffic be deleted first.
Still another object of the invention is to provide a method by which the size of the output data queue is affected by using fuzzy logic.
SUMMARY OF THE INVENTION
In other words, the invention solves the above problem by an output queue length simulation, an input source traffic estimator and fuzzy logic for the decision of when a data frame has to be deleted from the data stream. The main advantages with this solution are low processor load and simplicity by a combination of a few principles.
Further advantages of the present invention will appear from the following description taken in connection with the enclosed drawing, as well as from the appended patent claims.
REFERENCES:
patent: 4630261 (1986-12-01), Irvin
patent: 5231633 (1993-07-01), Hluchyj et al.
patent: 5444820 (1995-08-01), Tzes et al.
patent: 5521923 (1996-05-01), Willmann et al.
patent: 5546326 (1996-08-01), Tai et al.
patent: 5566163 (1996-10-01), Petit
patent: 5764641 (1998-06-01), Lin
patent: 5812526 (1998-09-01), Chang et al.
patent: 6018516 (2000-01-01), Packer
patent: 19609081 (1997-06-01), None
patent: 0 667570 (1995-08-01), None
patent: 0 705006 (1996-04-01), None
patent: 0 774846 (1997-05-01), None
patent: WO 96/15599 (1996-05-01), None
Thomas D. Ndousee: Fuzzy Neural Control of Voice Cells in ATM Networks, IEEE Journal On Selected Areas in Communications, vol. 12, No. 9, Dec., 1994, pp. 1488-1494.*
Rydman, P.; International Search Report; International Application No. PCT/NO98/00240; Mar. 11, 1999; pp-1-2.
Kizou Hassan
Lee Timothy
Telefonaktiebolaget LM Ericsson (publ)
LandOfFree
Priority control of queued data frames in frame delay... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Priority control of queued data frames in frame delay..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Priority control of queued data frames in frame delay... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3333469