摘要:
A method and apparatus for buffering data units in a communication switch that allows for configurable monitoring of the buffer contents is presented. Such an apparatus includes a context table that stores a plurality of independent group identifiers for each connection. Although the group identifiers may include a partition group identifier and a loss group identifier that is dependent on the partition group identifier, additional group identifiers are included in the context table for each connection that are independent of other group identifiers in the context table. Such a context table may be dynamically reconfigured in order to group connections for buffer monitoring operations related to congestion detection, traffic shaping, and data admission with respect to buffering. When a data unit is received corresponding to a particular connection, the context table is referenced to retrieve the set of group identifiers corresponding to that connection. As a cell is stored in the buffer, count values corresponding to at least some of the group identifiers included in the set of group identifiers for that connection are incremented. Similarly, when data is dequeued from the buffer for forwarding, the set of group identifiers for that particular data unit is determined and the counters corresponding to at least some of those groups are decremented. As such, the count value corresponding to each of the groups can be referenced to determine the number of data units corresponding to that group that are currently stored within the buffer.
摘要:
A method and apparatus for buffering data units in a communication switch that allows for configurable monitoring of the buffer contents is presented. Such an apparatus includes a context table that stores a plurality of independent group identifiers for each connection. Although the group identifiers may include a partition group identifier and a loss group identifier that is dependent on the partition group identifier, additional group identifiers are included in the context table for each connection that are independent of other group identifiers in the context table. Such a context table may be dynamically reconfigured in order to group connections for buffer monitoring operations related to congestion detection, traffic shaping, and data admission with respect to buffering. When a data unit is received corresponding to a particular connection, the context table is referenced to retrieve the set of group identifiers corresponding to that connection. As a cell is stored in the buffer, count values corresponding to at least some of the group identifiers included in the set of group identifiers for that connection are incremented. Similarly, when data is dequeued from the buffer for forwarding, the set of group identifiers for that particular data unit is determined and the counters corresponding to at least some of those groups are decremented. As such, the count value corresponding to each of the groups can be referenced to determine the number of data units corresponding to that group that are currently stored within the buffer.
摘要:
A method and apparatus includes processing for combining a plurality of virtual circuits into a combined virtual circuit, where such processing begins by buffering cells of each virtual circuit into a corresponding buffer. The processing then continues by obtaining priority information for each virtual circuit and obtaining logical buffer de-queuing information. The priority information, for example, may equate to priorities established via the varying levels of ATM services. The logical buffer de-queuing information corresponds to an access sequence for a plurality of logical ring buffers that are comprised of the buffers, or buffer identifiers. The processing then continues by generating the combined virtual circuit based on the logical buffer de-queuing information and the priority information.
摘要:
The present invention provides a system and method for providing a different quality of service to a flow associated with an aggregate flow is provided. For an embodiment, the method comprises scheduling a plurality of data traffic flows in a communication network, the method comprising: (i) scheduling data traffic flows in a first category of data traffic, said first category of data traffic having a first bandwidth capacity associated therewith; (ii) determining whether any portion of said first bandwidth capacity is unused by data traffic flows in said first category of data traffic; and (iii) scheduling data traffic flows in a second category of data traffic providing said unused portion of said first bandwidth capacity for use for said data traffic in said second category.
摘要:
A multi-protocol switch that supports both ATM and IP traffic and method for use is presented. The multi-protocol switch associates certain input connections, as identified by virtual path identifier/virtual connection identifier (VPI/VCI), with either IP or ATM traffic. When the connection identifier for a cell received indicates that the cell is an ATM cell, the multi-protocol switch forwards the cell through the switch, where the forwarding is based on the connection identifier for the cell. If the connection identifier for a cell indicates that the cell is IP traffic, the cell is stored with other cells included in the IP packet to which the cells correspond within the ingress line card of the switch. The destination address included in the packet is used to determine a forwarding decision for the cells included in the packet. After at least partial reassembly, the packet is segmented and forwarded through the switch based on the forwarding decision determined from the destination address.
摘要:
A method and apparatus for merging a plurality of virtual connections to form a merged virtual connection is presented. As cells for each of the plurality of virtual connections are received, they are buffered into a corresponding plurality of cell buffers, where each virtual connection has a corresponding cell buffer. When enough cells to constitute a complete packet are buffered for a particular virtual connection in its corresponding cell buffer, the identity of that virtual connection is queued into a queue, which may take the form of a linked list. A plurality of linked lists may be included such that differentiation between various virtual connections based on class is possible, where the class division may be based on a number of different factors. Prioritization information for the merged virtual connection is then obtained, where in the case where a plurality of classes are supported, the prioritization information includes class prioritization information. A cell stream for the merged virtual circuit is then generated based on the prioritization information and the virtual connection identities currently stored in the queue structure. The cell stream is generated such that entire packets are included in the cell stream in a continuous manner and no intermingling of cells corresponding to different packets occurs.
摘要:
A method and apparatus for scheduling and servicing events using a calendar structure is described. In accordance with one preferred embodiment of the present invention, a calendar structure is provided to implement work-conserving methods (for example, queuing, such as fair queuing, or, as one specific example, weighted fair queuing (WFQ)). Such a calendar structure preferably provides two slots per tier and uses a temporal pointer based on virtual time. In accordance with another preferred embodiment of the present invention, a calendar structure is provided to implement shaping of flows of information. Such a calendar structure preferably provides one slot per tier and uses a temporal pointer based on real time. For scheduling, a preferred occurrence time at which an event is preferred to occur is calculated. Events having preferred occurrence times farther from a current time value denoted by the temporal pointer are scheduled on a calendar tiers of lower resolution, while events having preferred occurrence times nearer to the current time value denoted by the temporal pointer are scheduled on calendar tiers of higher resolution. For servicing, the events are selected from slots to which the temporal pointer is pointing. If a slot is being used to schedule an event pending servicing, the slot is considered to be an occupied slot. Occupied slots at higher resolution calendar tiers are serviced exhaustively over occupied slots at lower resolution calendar tiers.