Nortel Networks Corporation

Ciena Corporation

Alan Stanley John Chapman

Hsiang-Tsung Kung

Elwyn Brian Davies

Hsiang T Kung

The present invention relates to a transmission device implementing a flow control mechanism for aggregate trunks. The transmission device can be implemented as a router that includes an input for receiving aggregate traffic streams, an output for releasing the aggregate traffic streams to a destination point and a control unit capable to regulate the rate of release of packets from the output. Specifically, the flow control operation effected by the control unit is dependent on receipt of acknowledgement messages issued at the destination point, each acknowledgement message confirming the receipt of one or more particular packets at the destination point. The control unit will continuously increase the packet sending rate until a packet is lost in the network between the transmission device and the destination point. On detection of packet loss, based on the lack of one or more corresponding acknowledgement messages from the destination point, the control unit will reduce the packet sending rate.

Method and apparatus for bandwidth management of aggregate data flows

Transport networks are facing new challenges and opportunities because of the explosive growth of data traffic. Besides having to meet the ever increasing bandwidth demand, transport networks need to provide new functionalities for the support of data applications. Novel enhanced transport systems with inherent packet multiplexing to meet these challenges are described. It uses Internet protocols and technologies to implement these packet transport systems, and thereby re-use much of the existing Internet infrastructure already widely deployed.

Efficient packet data transport mechanism and an interface therefor

The present invention relates to a switch for processing data units, such as IP data packets. The switch can be implemented as a router that includes a plurality of input ports, a plurality of output ports and a switch fabric capable of establishing logical pathways to interconnect a certain input port with a certain output port. A characterizing element of the router is its ability to control bandwidth usage on a basis of a logical pathway. This prevents congestion to occur in the switch fabric and also at the level of the output ports. For every active logical pathway the router sets-up a bandwidth control mechanism including at least one queue to hold data units received at an input port. The bandwidth control mechanism performs an accounting operation to determine the average bandwidth usage and if less than a threshold requests for releasing data units in the switch fabric are sent to the switch fabric controller. When the threshold is exceeded the bandwidth control mechanism stops sending requests for service to the switch fabric controller. The system also supports priorities. Instead of stopping the issuance of signals requesting release of data packets the priority level of the queue (reflected in a priority field in the signal requesting release of a data packet) changes from HI to LO. The switch fabric controller is designed to recognize priority requests and will accept LO priority requests only when there are no other HI priority requests pending. The present invention also provides a data transport device which includes a plurality of nodes interconnected by a physical link that establishes two ring-shaped paths on which data is transported on opposite directions. Each node is capable of sending a control message to the upstream node to advise the upstream node of the data carrying requirements of the node at which the message has originated. The upstream node can then throttle the introduction of data in the paths to preserve capacity that can be used by the downstream nodes.

Method and apparatus for simple IP-layer bandwidth allocation using ingress control of egress bandwidth

Flow control of packet based traffic by window is known. Novel modification is described which causes the flow control mechanism to reduce sending rate to some configured number rather than just reducing it by a fixed amount such as one half. The description also shows how the flow control mechanism can be constrained to a maximum rate. The configured numbers will assure that the connection can always run at a minimum rate but not more than a maximum rate. If the guaranteed minimum bandwidth is known and the round trip time between the end points is known or has been calculated, then the sender node needs only reduce its window to that which corresponds to a sending rate equal to that configured number. In this way the protocol will still probe for extra, opportunistic bandwidth but will be able to maintain the minimum rate. In a similar way a window that corresponds to the maximum rate can be calculated and used to constrain the maximum rate of sending.

Apparatus for and method of managing bandwidth for a packet based connection

A method of monitoring telecommunications network traffic comprising the steps of: receiving a packet stream comprising packets each identified as belonging to one of at least three classes; calculating a difference between the numbers of packets received identified as belonging to a first and a second of said classes; and deriving a measure of traffic load on the network responsive to said difference. The invention also relates to a method for admission control based on the above method of monitoring and a method for overcoming admission control avoidance. It also relates to apparatus embodying these methods.

Internet differentiated services service for transaction applications

Congestion at a network node can be aggravated by having too many TCP connections. A simple method of avoiding the bad effects of too many TCP connections is to limit the number of connections. Limiting the number of connections is achieved by an admission control which delays or even discards the connection set-up packets. TCP traffic flows are monitored to generate packet loss characteristics and when a certain condition is met, a connection request queue is disabled.

TCP admission control

The present invention relates to a lossy switch for processing data units, for example IP data packets. The switch can be implemented as a contained network that includes a plurality of input ports, a plurality of output ports and a lossy switch fabric capable of establishing logical pathways to interconnect a certain input port with a certain output port. A characterizing element of the switch is its ability to control the discard of data packets at a transport point within the switch. This control mechanism prevents and reduces congestion which may occur within the switch fabric and at the level of the input and output ports. The system also supports priorities, routing HI priority request data packets over the switch fabric before LO priority request data packets, and discarding LO priority data packets first when controlling congestion.

Memory and apparatus for input based control of discards in a lossy packet network

The invention provides a method and apparatus for re-ordering data traffic units, such as IP data packets, that may have been miss-ordered during a transmission over a multi-pathway link between a source node and a destination node in a network. The re-ordering apparatus includes a storage medium for intercepting the IP data packets and holding the IP data packets to allow IP data packets delayed on slower pathways to catch-up. The IP data packets in the storage medium are re-ordered based on their sequence number in an attempt to restore the original order of the IP data packets. A maximal time delay variable determines how long a certain IP data packet can be held in the storage medium. The TP data packet is released prior to the maximal time delay variable or as the maximal time delay variable is exceeded.

Method and apparatus for re-ordering data packets in a network environment

The present invention relates to the field of data communication networks. More specifically, it pertains to devices and methods for the management of bandwidth on links between routing nodes in data networks. The system is particularly useful for reducing congestion caused by high volume traffic streams. The invention provides a novel router that separates traffic on the basis of data type into separate queues buffers. The various queue buffers are associated with virtual output ports on a common physical link. A scheduler regulates the data release from the queue buffers into the physical link to control the bandwidth portion that is made available to each type of data.

Growing community of inventors

Kanata, Canada

Alan Stanley John Chapman