Multiplex communications – Data flow congestion prevention or control – Control of data admission to the network
Reexamination Certificate
1998-10-07
2002-07-23
Chin, Wellington (Department: 2662)
Multiplex communications
Data flow congestion prevention or control
Control of data admission to the network
C370S253000, C370S400000, C709S235000
Reexamination Certificate
active
06424624
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to congestion detection and flow control in high speed packet switching networks and more particularly to methods and apparatus for implementing congestion detection and flow control for low priority traffic with optimized cost efficiency.
BACKGROUND ART
Modern digital networks often operate in a multimedia environment and interconnect, upon demand, very large numbers of users and applications through fairly complex digital communication network topologies.
Due to the variety of users' demands and the growth of distributed applications, network traffic is consuming more bandwidth, becoming more non-deterministic and requiring more connectivity. These changes have been the driver for the emergence of fast packet switching network architectures in which data, voice and video information are digitally encoded, chopped into fixed or variable length packets (also named “cells in ATM or Asynchronous Transfer Mode networks) and transmitted through a common set of nodes and links interconnected to constitute the network communication facilities.
The need for efficient transport of mixed traffic streams on very high speed lines (sometimes referred to as links or trunks), means, imposes a set of performance and resource consumption requirements including very high throughput, very short packet processing time, the flexibility to support a wide range of connectivity options and efficient flow and congestion control. Congestion is generally defined as a condition during which network performance is degraded due to saturation of network resources such as communication links, processor cycles, memory buffers, etc.
One of the key requirements for high speed packet switching networks is reduction of end to end delay in order to satisfy real time delivery constraints and to achieve the necessary high nodal throughput for the transport of voice and video. Increases in link speeds have not been matched by proportionate increases in the processing speeds of communication nodes. The fundamental challenge for high speed networks is to minimize the processing time and to take full advantage of the high speed/low error rate technologies. Most of the transport and control functions provided by the new high bandwidth network architectures are performed on an end to end basis.
One basic advantage of packet switching techniques (as opposed to so-called circuit switching techniques) is that it allows statistical multiplexing of the different types of data over a line, which optimizes the utilization of transmission bandwidth. One drawback, however, is that packet switching introduces delays and jitters which might be detrimental for transmission of isochronous data, like video or voice. Methods have been proposed to control the network in such a way that delays and jitters are bounded for every new connection that is set up across the packet switched network.
Such methods are described, for instance, in a published European Application number 0000706297 and include establishing a path through the network high speed lines and nodes, via an entry node port of said network, making optimal use of the available transmission bandwidth of the network along the path to the indicated destination.
Because different type of traffics need to be treated differently to maintain their usefulness at a destination, choices have to be made among the different types by assigning different specific priorities. In other words, when a source terminal requests a connection to a destination terminal via the network (i.e., a call is set-up), a quality of service (QoS) is assigned to the call in terms of maximum permissible delay (T
13
max) and packet loss probability (P_loss).
The QoS and traffic characteristics (e.g., peak data rate, mean data rate and average packet length) are used to compute the amount of bandwidth (i.e. equivalent capacity or Ceq) to be reserved on every line on the route or path assigned to the traffic between the source terminal and the destination terminal, in order to guarantee a packet loss probability which is smaller than the loss probability (P_loss) that has been specified for the connection. However, in operation, the network traffic must be controlled dynamically which means that some packets may have to be dropped or discarded within the network to avoid traffic congestion.
In practice, it is common to reserve bandwidth for high priority packets (e.g. so-called Real Time or RT traffic) allowing such packets are transmitted in preference to lower priority packets derived from discardable traffic (e.g. Non Real Time or NRT traffic or more particularly Non Reserved or NR traffic). Lower priority packets may be sent at rates greater than their declared rate to dynamically take advantage of any bandwidth remaining after all the higher priority traffic has been served. This remaining bandwidth can vary widely depending on the actual activity of the high priority traffic sources. It is therefore of considerable importance to manage the low priority traffic so as to optimize the use of the widely varying left-over bandwidth in the network while avoiding any congestion which would reduce network throughput. This obviously requires providing the network (and eventually also the sources) with congestion detection and flow control facilities.
Various mechanisms for controlling the flow of NR traffic have been proposed. In particular, an Available Bit Rate (ABR) flow control mechanism has been proposed for Asynchronous Transfer Mode (ATM) networks. ABR flow control is based on use of a particular flow control cell, the so-called Resource Management or RM cell. RM cells are used to collect congestion information from network node switches along connection paths and to send such information back to the traffic sources. While ABR flow control seems to be very efficient, it is complex to implement. End systems must generate RM cells periodically, provide scheduling for RM cells to be sent among data cells, and shape their traffic in response to congestion indications conveyed by received RM cells. Intermediate systems (switches along the paths) must be able to differentiate RM cells from regular data cells, extract RM cells and update these with congestion information. These complexities limit the cost effectiveness of this solution.
Moreover, the above solution requires that all so-called non-reserved (i.e. low priority) sources connected to a network be treated using the ABR mechanism. In fact, if a mix of ABR sources and non-ABR sources are connected to the network, the ABR sources will be disadvantaged as compared to the non-ABR sources which need not be capable of sending RM cells. The customer must therefore update the hardware of all the end systems before using ABR support, which is an additional drawback from an engineering standpoint.
Finally, there is no description, in the ABR standard, of a policing function which could be used to protect the network from misbehaving sources or from non ABR sources.
Other mechanisms have been proposed, with flow control which can be used on ATM or PTM (Packet Transfer Mode, including variable length packets) traffic and offer good performance. These flow control mechanisms add complexity to network equipment. Access nodes or ports need to store tables of values and must have the capability of adding a time-stamp to the data packets or to specific control packets. The overhead on the lines is also increased as at least a time stamp must be added to some transmitted packets.
A further improvement was disclosed in U.S. Pat. No. 5,313,454 which made the system transparent to the user (source) by providing an internal congestion avoidance method. To that end, congestion is identified throughout the network and transferred by setting an indicator in the packet header. Then congestion indications are used in the destination node to generate a rate control message which is fed back to the entry node. This prior art still adds overhead to the feedback flow if smooth and flexible congestion control is sought. Otherwise, the
Enger Marcus
Fichou Aline
Foriel Pierre-Andre
Galand Claude
Cesari and McKenna LLP
Chin Wellington
Cisco Technology Inc.
Tsegaye Saba
LandOfFree
Method and system for implementing congestion detection and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for implementing congestion detection and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for implementing congestion detection and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2891056