Abstract: Methods and apparatus provide an adaptive load balancer that presents a virtual data system to client computer systems. The virtual data system provides access to an aggregated set of data, such as files or web service objects, available from a plurality of server data systems respectively operating within a plurality of server computer systems. The adaptive load balancer receives a client data access transaction from a client computer system that specifies a data access operation to be performed relative to the virtual data system presented to the client computer system. The adaptive load balancer processes the client data access transaction in relation to metadata associated with the virtual data system to provide access to the file or service object within a server computer system, or to access the metadata.

Abstract: Each time data, in the form of data blocks protected by code checks, must be reformatted, the original data is broken into new data blocks and a new code check is calculated from, and combined with, each new data block, but the new data blocks and new code checks are both reconstituted versions of the original data blocks and the original code checks. Consequently, the data is never left without protection. In one embodiment, an ingress encoder recomputes an ingress code check from an original data block and its associated header. An egress encoder computes an egress code check from the egress header for an outgoing data block reformatted from the original data block and the ingress code check. The outgoing information is then assembled from the egress header, the outgoing data block and the newly computed egress code check.

Abstract: Congestion management is provided using one or more receive queues organized into a queue list. Each queue is associated with a high water mark and a low water mark. The high water mark indicates a length at which the queue is considered congested. The low water mark is a length at which a congested queue is considered no longer congested. When a queue reaches or exceeds its high water mark, newly arriving frames destined for that queue which are eligible for discard are dropped. Frames arriving not marked as eligible for discard are enqueued at the queue's tail. Queues in the queue list are organized by relative priority. When a single queue above a predetermined priority is determined to be congested, all queues in the queue list are treated as congested. Congestion in lower priority queues results in only the individual congested queue being treated as congested.

Abstract: A system for temporarily storing and dequeuing received data units including searching a prioritized queue list for a highest priority queue having a head entry associated with a received data unit which is part of the guaranteed bandwidth of the connection on which it was received, and for which there are sufficient transmission credits available, based on a flow control mode associated with the data unit. If no such guaranteed bandwidth data unit is found, the system dequeues a highest priority queue having a head entry associated with an available bandwidth data unit, for which there are sufficient transmission credits available to begin transmission based on a flow control mode associated with the data unit. If no such available bandwidth data unit is found, then no data unit is dequeued during that traversal of the queue list.

Abstract: A system and method for performing event-based rate policing using varying window start times. Rate policing overhead, including counter and timer monitoring and resetting, is performed only as needed and in response to actual received traffic on each connection. As the last bit of a data unit is received from an external network, an “event time stamp” is generated and associated with the data unit, for example as part of an internal header or trailer attached to the data unit. To determine if a rate policing window was active when the frame was received, the event time stamp is compared with a sum of a window start time and a window period value stored in association with the connection on which the data unit was received. If the associated event time stamp indicates a time prior to the sum of the associated window start time and window period, then a rate policing window is determined to have been active when the frame was received.

Abstract: A system and method for providing QoS levels to received data units, such as ATM cells and/or Frame Relay Frames. A QoS level associated with a virtual connection is supported through allocation and management of resources used to process data units received over the connection. Resources used to support QoS levels are assigned to groups of connections referred to as “QoS groups”. Resources used to support one group of connections may be isolated from resources supporting other connections. When a connection is established, if a QoS group exists having the QoS level of the new connection, the connection may be added to that QoS group. Similarly, where the QoS levels for multiple connections must be modified, if those connections are within a single QoS group, then modification of the QoS level for that group would advantageously result in modification of the QoS for each virtual connection within the QoS group.

Abstract: The invention comprises a system bus apparatus and method for a multi-arm, multiprocessor computer system having a main memory and localized buffer cache memories at each processor. Each block of data in a cache includes tag bits which identifies the condition of the data block in relation to the corresponding data in main memory and other caches. The system bus (SYSBUS) comprises three subparts; 1) a MESSAGE/DATA bus, 2) a REQUEST/GRANT bus and 3) a BCU bus. The MESSAGE/DATA bus is coupled to every device on the system and is used for transferring messages, data and addresses. The REQUEST/GRANT bus couples between every device on an arm of the system and that arm's bus control unit (BCU). The BCU bus couples between the various BCUs.