Abstract: A cache may store critical cache lines and non-critical cache lines, and may attempt to retain critical cache lines in the cache by, for example, favoring the critical cache lines in replacement data updates, retaining the critical cache lines with a certain probability when victim cache blocks are being selected, etc. Criticality values may be retained at various levels of the cache hierarchy. Additionally, accelerated eviction may be employed if the threads previously accessing the critical cache blocks are viewed as dead.

Abstract: A cache may store critical cache lines and non-critical cache lines, and may attempt to retain critical cache lines in the cache by, for example, favoring the critical cache lines in replacement data updates, retaining the critical cache lines with a certain probability when victim cache blocks are being selected, etc. Criticality values may be retained at various levels of the cache hierarchy. Additionally, accelerated eviction may be employed if the threads previously accessing the critical cache blocks are viewed as dead.

Abstract: A cache may store critical cache lines and non-critical cache lines, and may attempt to retain critical cache lines in the cache by, for example, favoring the critical cache lines in replacement data updates, retaining the critical cache lines with a certain probability when victim cache blocks are being selected, etc. Criticality values may be retained at various levels of the cache hierarchy. Additionally, accelerated eviction may be employed if the threads previously accessing the critical cache blocks are viewed as dead.

Abstract: A cache may store critical cache lines and non-critical cache lines, and may attempt to retain critical cache lines in the cache by, for example, favoring the critical cache lines in replacement data updates, retaining the critical cache lines with a certain probability when victim cache blocks are being selected, etc. Criticality values may be retained at various levels of the cache hierarchy. Additionally, accelerated eviction may be employed if the threads previously accessing the critical cache blocks are viewed as dead.

Abstract: A switch device can identify when a loss of signal event occurs on a communication link connecting a local link partner and a remote link partner. The switch device may automatically perform a link restart process to restore communication with the remote link partner without performing a speed negotiation with the remote link partner. The link restart process may include disabling a transmitter for a recovery duration and configuring a transmitter and a receiver to an initial configuration state. The link restart process may also synchronize the start of a training protocol communication between the local link partner and remote link partner. The switch device may also perform the link restart process when traffic received from the remote link partner fails a link criteria.

Abstract: A switch device can identify when a loss of signal event occurs on a communication link connecting a local link partner and a remote link partner. The switch device may automatically perform a link restart process to restore communication with the remote link partner without performing a speed negotiation with the remote link partner. The link restart process may include disabling a transmitter for a recovery duration and configuring a transmitter and a receiver to an initial configuration state. The link restart process may also synchronize the start of a training protocol communication between the local link partner and remote link partner. The switch device may also perform the link restart process when traffic received from the remote link partner fails a link criteria.

Abstract: When cipher block chaining encryption/decryption is employed, write fault recovery is accomplished by storing information useful for the writing of cipher block chaining processed data before it is written to storage cells. Hence when write failure is discovered, this information stored can be retrieved for rewriting the data to the cells. Preferably, the information stored includes security configuration information for cipher block chaining processing a unit of data.

Abstract: A system and method is disclosed for providing real time signal to noise computation for a 100 Mb Ethernet physical layer device. A slicer module obtains a slicer error signal in the physical layer device and generates a signal that represents a square of the slicer error signal. A variance computation unit uses the signal that represents the square of the slicer error signal to calculate a value of variance of a histogram of the physical layer device by continuously accumulating a plurality of values of the square of the slicer error signal for a selected period of time. A real time signal to noise ratio is then computed using the calculated value of the variance of the histogram of the physical layer device.

Abstract: A method or corresponding apparatus in an exemplary embodiment of the present invention monitors an operational status of network links and retrieves or receives the operational status. The operational status is stored (e.g., within 50 msec) on an ongoing basis. The operational status corresponds to an operational LSP from among a primary LSP and secondary LSP(s) used to service traffic via the network links to maintain network connectivity via the LSPs. The secondary LSP(s) may be either a bypass LSP or a backup LSP, where each LSP is Fast Re-route (FRR) enabled. Further, correspondence between the primary LSP, secondary LSP(s), and the network links may be tracked. In order to deliver network traffic via the operational LSP known to service traffic between the source and destination, a link identifier may be used. The link identifier is determined from the source and destination information within the network traffic.

Abstract: A method or corresponding apparatus in an exemplary embodiment of the present invention determines how many copies of a multicast packet are to be sent, as copies, to multiple destinations based on information in the multicast packet and a group (e.g., bridge node) with which the packet is associated. The copies of the multicast packets are then generated. After generating the copies, an action to take for each copy is determined. This determination is made using the information in the multicast packet and based on the group with which the packet is associated. After the action is determined for each copy, the action is performed on each copy. Through use of an embodiment of the present invention, memory space is conserved, allowing for continued use of a device in a multicast environment of an ever growing network, such as the Internet.

Abstract: A method or corresponding apparatus in an exemplary embodiment of the present invention monitors an operational status of network links and retrieves or receives the operational status. The operational status is stored (e.g., within 50 msec) on an ongoing basis. The operational status corresponds to an operational LSP from among a primary LSP and secondary LSP(s) used to service traffic via the network links to maintain network connectivity via the LSPs. The secondary LSP(s) may be either a bypass LSP or a backup LSP, where each LSP is Fast Re-route (FRR) enabled. Further, correspondence between the primary LSP, secondary LSP(s), and the network links may be tracked. In order to deliver network traffic via the operational LSP known to service traffic between the source and destination, a link identifier may be used. The link identifier is determined from the source and destination information within the network traffic.

Abstract: A method or corresponding apparatus in an exemplary embodiment of the present invention determines how many copies of a multicast packet are to be sent, as copies, to multiple destinations based on information in the multicast packet and a group (e.g., bridge node) with which the packet is associated. The copies of the multicast packets are then generated. After generating the copies, an action to take for each copy is determined. This determination is made using the information in the multicast packet and based on the group with which the packet is associated. After the action is determined for each copy, the action is performed on each copy. Through use of an embodiment of the present invention, memory space is conserved, allowing for continued use of a device in a multicast environment of an ever growing network, such as the Internet.