摘要:
A method and apparatus for providing buffer state accounting at a link level, otherwise known as link flow control, in addition to flow control at the virtual connection level. Link flow control enables receiver cell buffer sharing while maintaining per-connection bandwidth with lossless cell transmission. High link level update frequency is enabled without a significant sacrifice in overall link forward bandwidth. A higher and thus more efficient utilization of receiver cell buffers is achieved.
摘要:
A method and apparatus for providing buffer state accounting at a link level, otherwise known as link flow control, in addition to flow control at the virtual connection level. Link flow control enables receiver cell buffer sharing while maintaining per-connection bandwidth with lossless cell transmission. High link level update frequency is enabled without a significant sacrifice in overall link forward bandwidth. A higher and thus more efficient utilization of receiver cell buffers is achieved.
摘要:
A method and apparatus for providing buffer state accounting at a link level, otherwise known as link flow control, in addition to flow control at the virtual connection level. Link flow control enables receiver cell buffer sharing while maintaining perconnection bandwidth with lossless cell transmission. High link level update frequency is enabled without a significant sacrifice in overall link forward bandwidth. A higher and thus more efficient utilization of receiver cell buffers is achieved.
摘要:
Permission based flow control is implemented in a computer network having at least a downstream, intermediate and upstream network device by receiving credits at the intermediate network device from the downstream network device and granting credits from the intermediate network device to the upstream network device based at least in part upon the credits received at the intermediate network device from the downstream network device. Credit chaining as described above is employed to permit the granting of the right to transmit downstream to be predicated upon buffer availability downstream of the next downstream network device. Via the use of credit chaining, high utilization of network resources is achieved with minimal loss of data traffic.
摘要:
Method and apparatus for extending lower-layer flow control end-to-end, transparent to network host systems, between network hosts attached to legacy networks without requiring any changes to the host networks or hosts themselves. In a first embodiment, a source host and a destination host, each residing on a respective LAN, communicate via an intermediate network supporting the lower-layer flow control. A higher-layer, end-to-end flow control protocol exists between the hosts. The lower-layer flow control protocol accounts for loss of bandwidth between each hop within the intermediate network. To extend the lower-layer flow control protocol to the hosts, an intelligent edge device of the intermediate network adjusts observed higher-layer flow control parameters based upon the state of the lower-layer flow control. No modifications to the legacy network or hosts is required, and the extension of flow control is transparent to the legacy networks.
摘要:
A flow control technique for wide area ATM networks is disclosed in which allocation of buffers in a receiver switch is controlled by a transmitter switch. The receiver switch periodically transmits feedback messages to the transmitter switch indicatative of the state of fullness of the receiver switch buffers. The transmitter switch calculates updated receiver buffer fullness based upon the feedback message and the number of cells transmitted from the transmitter switch to the receiver switch since the feedback message was sent. Transmission of cells from the transmitter switch to the receiver switch is then controlled in accordance with an allocation technique, thereby allocating buffers in the receiving switch. The technique may be a roll-off technique in which the number of buffers available to each flow in the transmitter switch is reduced geometrically as the updated receiver buffer state is calculated to be more full. The flow control technique may be applied on either a per link or a per flow basis.
摘要:
A system and a method are disclosed for supporting multiple application traffic types over connection oriented networks. Data is received from multiple applications, for example in the form of data units, and is associated with a traffic type. The traffic type associated with a data unit indicates whether the data unit is inelastic traffic or elastic traffic. Data associated with inelastic traffic is mapped to at least one connection of a first connection type which provides for some amount of committed bandwidth, and limits on delay variation. Data associated with elastic traffic is mapped to at least one connection of a second connection type, which does not provide any committed bandwidth. Connections of the second connection type may ensure that elastic traffic data units that are accepted for transfer by the underlying network are delivered without loss due to network congestion, for example, by use of a flow control technique. A hierarchical scheduler may be employed to schedule transmissions of data units across connections organized into virtual trunks. The hierarchical scheduler may be used to ensure a high multiplexing gain with respect to connections within a single virtual trunk, as well as across multiple virtual trunks.
摘要:
A method is disclosed for facilitating multicast operation in a network in which a data unit is multicast from a root node to a plurality of leaves via a plurality of branching point nodes in response to feedback processed at each branching point node. At least one cell forwarding technique is selected from a plurality of cell forwarding techniques at the respective branching point nodes. The cell forwarding techniques facilitate multicast operation by controlling forwarding and discard of multicast cells. The forwarding techniques are realized via use of a buffer ring in which cells are stored prior to forwarding. Manipulating head and tail pointers associated with the buffer ring allows for a plurality of desirable forwarding techniques.