Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: A distributed system manager may identify updates to perform, identify devices to update, and perform the updates while maintaining service availability. The manager may consist of an orchestrator that uses a plug in architecture to interface with different update systems. The plugins may interface with an update system and respond to scan, download, and install commands from the orchestrator. The orchestrator may prepare each device within the distributed system for updating, then cause the updates to be performed via the plugins. In some embodiments, the logic for selecting devices to update and managing the workloads on the devices may be contained in a second type of plugin.

Abstract: Described is a technology by which a distributed system of computer nodes that provides services is self-updated via a component of the distributed system. Such a self-updating run applies one or more updates to each of the member computer nodes in the distributed system while maintaining the services provided by the distributed system. The updating run may include moving workloads from each computer node to be updated, updating that node, and reinstating workloads after updating. Performing the updating run may include moving update-related responsibility, which may be a workload, from one computer node to another computer node.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: The disclosed embodiments relate to an apparatus and method for acknowledging a data transfer. The first protocol may generate a request for a data transfer. The second protocol may receive the request for a data transfer from the first protocol. With the request, the second protocol may determine if the request for the data transfer contains a request for acknowledgement of completion of the data transfer. The second protocol may send a request corresponding to the request for data transfer to a third protocol. If the request for data transfer does contain a request for acknowledgement of completion of the data transfer, then the second protocol may set a variable in memory to wait for an event to correspond to the completion of the request and send an acknowledgement to the first protocol upon the occurrence of the event.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: Described is a technology by which a distributed system of computer nodes that provides services is self-updated via a component of the distributed system. Such a self-updating run applies one or more updates to each of the member computer nodes in the distributed system while maintaining the services provided by the distributed system. The updating run may include moving workloads from each computer node to be updated, updating that node, and reinstating workloads after updating. Performing the updating run may include moving update-related responsibility, which may be a workload, from one computer node to another computer node.

Abstract: A locking system may place a persistent object on each of the devices in a computer cluster. The objects may indicate that an update is being performed on the devices so that other systems may not also update the devices. When an update has completed, the objects may be removed. The objects may include an identifier for the system owning the objects, as well as other information regarding the update. In some cases, the objects may include executable code and perform some of the update functions on the various devices.

Abstract: The disclosed embodiments relate to an apparatus and method for managing flow control of a data transfer. An apparatus according to embodiments of the present invention may comprise a first protocol associated with a plurality of receive buffers, a second protocol adapted to manage the plurality of receive buffers for the first protocol, and a third protocol that determines whether one of the plurality of receive buffers is available for a data packet and (a) if one of the plurality of receive buffers is available, permits an acknowledgement packet to be sent to a node that sent the data packet, and (b) if one of the plurality of receive buffers is unavailable, drops the data packet, notifies the second protocol regarding the unavailability of the plurality of receive buffers, and withholds the acknowledgement packet.

Abstract: A system and method for detecting path failure in a network are described. The system and method include a network interface which receives packets over particular connections, and an agent able to monitor the network interface and start a no-traffic timer for each particular connection when a packet is received on the connection. The agent is further able, when the no-traffic timer expires, to check for outgoing data on the particular connection, and compare a count of previously sent test packets against a limit. If the count is below the limit the agent sends a new test packet on the particular connection, the test packet provides an indication of the state of a network path associated with the particular connection. If the count is above the limit the agent initiates failover procedures for the associated path.

Abstract: An apparatus employs a work request list to access a memory device. The apparatus comprises an upper layer protocol that generates the work request list comprising a plurality of work requests, the work request list having an attribute that indicates the number of the plurality of work requests in the work request list. The apparatus additionally comprises an interface that is adapted to receive the work request list and individually enqueue the plurality of work requests.

Abstract: The disclosed embodiments relate to an optimized memory registration mechanism that may comprise an upper layer protocol that associates I/O buffers with memory regions and that manages steering tags. The memory regions may be associated with a translation page table. The upper layer protocol may allocate one of the steering tags associated with at least one of the memory regions for a memory operation.

Abstract: An efficient technique for performing remote asynchronous mirroring includes receiving a stream of command requests, performing the requested commands, relaying the requested commands to a target while embedding an explicitly-defined or implied task precedence graph in the relayed commands to enable increased concurrency in tasks performed by the target.

Abstract: A system and method for detecting path failure in a network are described. The system and method include a network interface which receives packets over particular connections, and an agent able to monitor the network interface and start a no-traffic timer for each particular connection when a packet is received on the connection. The agent is further able, when the no-traffic timer expires, to check for outgoing data on the particular connection, and compare a count of previously sent test packets against a limit. If the count is below the limit the agent sends a new test packet on the particular connection, the test packet provides an indication of the state of a network path associated with the particular connection. If the count is above the limit the agent initiates failover procedures for the associated path.

Abstract: The disclosed embodiments relate to a queuing mechanism that may comprise a shared receive queue having a plurality of buffers. The queuing mechanism may also comprise a plurality of queue pairs, each of the plurality of queue pairs having a receive queue that comprises at least one of the plurality of buffers.

Abstract: An efficient technique for performing remote asynchronous mirroring includes receiving a stream of command requests, performing the requested commands, relaying the requested commands to a target while embedding an explicitly-defined or implied task precedence graph in the relayed commands to enable increased concurrency in tasks performed by the target.