Abstract: A server computer that is connected over a network to one or more remote client computers includes specialized hardware for handling computationally-intensive signal processing operations involving video, audio, and/or other signals. As a result, users of the client computers experience little or no delay when accessing over the network applications that are hosted on the server.

Abstract: In one embodiment, a method is described for detecting an operational failure between the network unit and an adjacent network unit in the stack; controlling the switching engine to redirect packets which would otherwise be sent from a particular port to the adjacent network unit to be forwarded from another port to be sent to a different network unit in the stack; and entering the switching engine into a bypass mode in response to control data indicating an operational failure between at least two other network units in the stack to cause packets to be forwarded without being re-directed by the switching engine.

Abstract: Systems and methods of controlling and optimizing system latency in order to minimize system latency between remotely operating computing devices over a communications network are disclosed. A method of controlling and optimizing latency over a computer network from a first computing device to a second computing device includes identifying a condition associated with the transmission of a video signal, receiving the video signal at a computing device, determining whether the condition is met after receiving the video signal, and displaying the video signal on a display device when the condition is met.

Abstract: In one embodiment, a method is described for detecting an operational failure between the network unit and an adjacent network unit in the stack; controlling the switching engine to redirect packets which would otherwise be sent from a particular port to the adjacent network unit to be forwarded from another port to be sent to a different network unit in the stack; and entering the switching engine into a bypass mode in response to control data indicating an operational failure between at least two other network units in the stack to cause packets to be forwarded without being re-directed by the switching engine.

Abstract: A server computer that is connected over a network to one or more remote client computers includes specialized hardware for handling computationally-intensive signal processing operations involving video, audio, and/or other signals. As a result, users of the client computers experience little or no delay when accessing over the network applications that are hosted on the server.

Abstract: Systems and methods of controlling and optimizing system latency in order to minimize system latency between remotely operating computing devices over a communications network are disclosed. A method of controlling and optimizing latency over a computer network from a first computing device to a second computing device includes identifying a condition associated with the transmission of a video signal, receiving the video signal at a computing device, determining whether the condition is met after receiving the video signal, and displaying the video signal on a display device when the condition is met.

Abstract: A network stack includes a plurality of network units each of which includes a multiplicity of ports for receiving and forwarding addressed data packets, at least two cascade ports and a switching engine for forwarding received packets to at least one port in accordance with address data in the packets and a cascade connection including, for each of two opposite directions around the stack, at least one unidirectional path for data packets composed of links each between a respective cascade port on a network unit and a corresponding cascade port on the next network unit.

Abstract: Detection of a signature in a data packet comprises performing a pre-classification of the packet, using header information and particularly a 5-tuple access control list, into one of a multiplicity of flows and directing the payload of the packet to a respective one of a multiplicity of deterministic finite state machines each of which stores a plurality of signatures as a sequence of states and acts only on the respective flow.

Abstract: A network stack includes a plurality of network units each of which includes a multiplicity of ports for receiving and forwarding addressed data packets, at least two cascade ports and a switching engine for forwarding received packets to at least one port in accordance with address data in the packets and a cascade connection including, for each of two opposite directions around the stack, at least one unidirectional path for data packets composed of links each between a respective cascade port on a network unit and a corresponding cascade port on the next network unit.

Abstract: Network units such as switches for use in a cascaded stack are organised to provide a cascade connection in the form of a dual unidirectional connection so that, in its ordinary configuration, there is at least one and preferably more than one unidirectional ring for each direction around the cascade, each ring including a respective port on each unit. For each ring, each port on a unit is connected by a respective link to a corresponding port on the preceding unit and the following unit. The units provide a self-healing operation in the event of various kinds of operational failure. The self-healing operation includes loop-back of packets in units adjacent the failure and bypass of a packet switching process for other units. The units include control logic for passing control frames containing status information relating to the units and links between them and for co-operation with a CPU to control a switching engine to perform the self-healing operation in accordance with that status information.

Abstract: Cascade control logic for use in a switch or other network unit that can be used in a cascaded stack can maintain normally a point-to-point half-duplex connection for control data with each of the next preceding and next succeeding units in the cascade. Each cascade logic device is organised so that for one direction, conveniently called the up direction, a device is a master and in the other direction the device is a slave in respect of the control path. A control device will generate master control frames in the up direction and deliver slave control frames in the down direction. The control device is organised so that in the absence of reception of valid control frames on a control link control data which would otherwise be sent out on that link is looped back within the control device. In this manner the control device can maintain under normal circumstances two virtual control channels which can ‘self-heal’ notwithstanding the failure or powering-down of a unit in the cascade.

Abstract: Detection of a signature in a data packet comprises performing a pre-classification of the packet, using header information and particularly a 5-tuple access control list, into one of a multiplicity of flows and directing the payload of the packet to a respective one of a multiplicity of deterministic finite state machines each of which stores a plurality of signatres as a sequence of states and acts only on the respective flow.

Abstract: A selected word is stored in a content addressable memory (CAM) by partitioning the word into at least two segments, the segments being individually lesser in width than the CAM but in aggregate greater than the width of the CAM. A first entry in the CAM comprises a predetermined prefix and a first of the segments and a second entry in the CAM comprises a second prefix, corresponding to the address of the first segment, and the second segment. A search key is similarly partitioned. In a first search cycle a first segment of the search key prefixed by the predetermined prefix is applied to the CAM and in the event of a matching entry a second segment of the search key, prefixed by a second prefix comprising an output address word identifying the matching entry, is applied to the CAM in a second search cycle.

Abstract: A FIFO store for data packets and their respective status words includes space for the writing of a predetermined sync word or one of a cyclic sequence of predetermined sync words with each status word. The sync word can be used to prevent forwarding of a packet and as an aid to fault diagnosis if on reading the status word the sync word does not match any of the predetermined sync words.

Abstract: A network router has a distributed processing scheme for multicast packets. The router has a look-up and forwarding engine operative in response to the reception of an internet protocol multicast packet to direct the packet to a multiplicity of transmit ports which may or may not be connected to the source network segment from which the multicast packet was received by the router. Each transmit port is associated with a predetermined default network segment and the engine directs the packet to the ports (e.g. across the router's backplane) with an identification of the source network segment and control flags which enable processing means associated with the ports to determine by comparison of the respective network segment and the source network segment whether the packet is bridged from the respective port or routed, with relevant modification of the packet, from that port.

Abstract: Controlling the allocation of priority for data packets within a multi-port switch includes detecting the commencement of an HTTP message; enabling the storage of an identification of the source port on which that message was received by the switch; finding correspondence between destination data of packets and the said source port; and adjusting a priority allotted to said packets in response to the correspondence. The adjustment of priority preferably increases the priority of packets forwarded to said the port relative to priority allotted packets received at the source port and thereby allots asymmetric priority to traffic between a server and a client.

Abstract: The present invention is based on a ring architecture wherein each chip in the multiplicity of chips is connected in a chain so that it can receive packets from the next previous chip in the chain and forward packets to the next following chip in the chain, whereby to form a ring. Before a packet is placed on the ring by a device it is provided with a ‘masking’ field which identifies each of the other devices and the ports within any such device to which the packet must be sent in order to achieve a multicast. When a device in its turn receives such a packet it will examine the masking field to determine whether the field identifies that device. If so, the packet, which may be held in temporary storage, may be processed for forwarding from a port of that device. The device will strip from the masking field the identification of the respective device and transmit the packet with a thus modified masking field to the next device in turn.

Abstract: Cascade control logic for use in a switch or other network unit that can be used in a cascaded stack can maintain normally a point-to-point half-duplex connection for control data with each of the next preceding and next succeeding units in the cascade. Each cascade logic device is organised so that for one direction, conveniently called the up direction, a device is a master and in the other direction the device is a slave in respect of the control path. A control device will generate master control frames in the up direction and deliver slave control frames in the down direction. The control device is organised so that in the absence of reception of valid control frames on a control link control data which would otherwise be sent out on that link is looped back within the control device. In this manner the control device can maintain under normal circumstances two virtual control channels which can ‘self-heal’ notwithstanding the failure or powering-down of a unit in the cascade.

Abstract: Network units such as switches for use in a cascaded stack are organised to provide a cascade connection in the form of a dual unidirectional connection so that, in its ordinary configuration, there is at least one and preferably more than one unidirectional ring for each direction around the cascade, each ring including a respective port on each unit. For each ring, each port on a unit is connected by a respective link to a corresponding port on the preceding unit and the following unit. The units provide a self-healing operation in the event of various kinds of operational failure. The self-healing operation includes loop-back of packets in units adjacent the failure and bypass of a packet switching process for other units. The units include control logic for passing control frames containing status information relating to the units and links between them and for co-operation with a CPU to control a switching engine to perform the self-healing operation in accordance with that status information.

Abstract: A FIFO store for data packets and their respective status words includes space for the writing of a predetermined sync word or one of a cyclic sequence of predetermined sync words with each status word. The sync word can be used to prevent forwarding of a packet and as an aid to fault diagnosis if on reading the status word the sync word does not match any of the predetermined sync words.