Abstract: A cache management system in a multiprocessing computing system avoids blocking subsequent memory requests to access data in the cache after a previous memory request to access the data in the cache generates a cache miss and while the cache is being updated with the data. The previous memory request and subsequent memory requests are stored in a piggyback FIFO while the data is retrieved from a memory device. The cache is then updated with the data and the previous memory request and subsequent memory requests are processed on the cache.

Abstract: An interconnect switch stores data messages received from one or more source devices and prioritizes the data messages received from each source device based on the order that the data messages were received from the source device. For each available destination device associated with the interconnect switch, the interconnect switch identifies the data messages with the highest priority that are to be routed to the available destination device and selects one of the identified data messages for the available destination device. The interconnect switch then routes the selected data messages to the available destination devices.

Abstract: A system and method for data compression using a hybrid coding scheme includes a dictionary, a statistical model, and an encoder. The dictionary is a list containing data patterns each associated with an index. The indices of received data patterns are sent to the statistical model and to the encoder. The statistical model gathers statistical information for the indices and sends it to the encoder. The encoder uses the statistical information to encode the indices from the dictionary. The encoder is preferably an arithmetic encoder.

Abstract: A method of filtering packets having a class. The method includes the steps of accessing a table comprised of entries that include an indication to which class a given entry applies. Then there is the step of searching the entries until an entry whose class indication corresponds to a current class of the packet. Next there is the step of forwarding a frame associated with the entry whose class indication corresponds to the class of the packet. A data communication apparatus. The apparatus includes a mechanism for routing or switching packets. The apparatus includes a memory having information for routing or switching the packets. The apparatus includes a mechanism for compaction of information in the memory which removes redundancies in the information.

Abstract: A method and apparatus for layer 3 switching packets between locally attached virtual local area networks without using a routing protocol are provided. A learning internetwork switch is connected between a router and a plurality of virtual local area networks. Communications between devices on the virtual local area networks and the router pass through the learning internetwork switch. By inspecting certain packets that flow between the devices and the router, the learning internetwork switch learns the location of the devices without having to use a routing protocol. The learning internetwork switch learns the network layer and the data link layer addresses of the various devices. Once the learning internetwork switch has learned the location, the network layer address and data link layer address of a device, the learning internetwork switch can forward packets between devices on different virtual local area networks using layer 3 switching without involving the router.

Abstract: A packet processing method and apparatus efficiently process a binary data packet based upon information contained in the header portion of the packet. The method and apparatus employ a binary tree search method for determining ranges of key elements of the packet and associating with each of the ranges a user supplied data and filter mask. The binary search process employs a unique binary tree storage structure which both conserves storage memory and enables efficient processing to reach a final node of the binary tree decision table. Nodes can be added or deleted from the table by the user as the data packet processing needs change.

Abstract: A packet processing method and apparatus efficiently process a binary data packet based upon information contained in the header portion of the packet. The method and apparatus employ a binary tree search method for determining ranges of key elements of the packet and associating with each of the ranges a user supplied data and filter mask. The binary search process employs a unique binary tree storage structure which both conserves storage memory and enables efficient processing to reach a final node of the binary tree decision table. Nodes can be added or deleted from the table by the user as the data packet processing needs change.