A distributed deep learning system includes a plurality of calculation nodes connected to one another via a communication network. Each of the plurality of calculation nodes includes a computation unit that calculates a matrix product included in computation processing of a neural network and outputs a partial computation result, a storage unit that stores the partial computation result, and a network processing unit including a transmission unit that transmits the partial computation result to another calculation node, a reception unit that receives a partial computation result from another calculation node, an addition unit that obtains a total computation result, which is a sum of the partial computation result stored in the storage unit and the partial computation result from another calculation node, a transmission unit that transmits the total computation result to another calculation node, and a reception unit that receives a total computation result from another calculation node.

A computer comprising a plurality of processing nodes is provided. Each processing node has at least one processor configured to process input data to generate an array of data items. The processing nodes are arranged in cliques in which each processing node of a clique is connected to each other processing node in the clique by first and second clique links. The cliques are inter-connected in rings such that each processing node is a member of a single clique and a single ring. The processing nodes of all cliques are configured to exchange in each exchange step of a machine learning collective via the respective first and second clique links at least two data items with the other processing node(s) in its clique, and all processing nodes are configured to reduce each received data item with the data item in the corresponding position in the array on that processing node.

Some embodiments provide a novel method for deploying different virtual networks over several public cloud datacenters for different entities. For each entity, the method (1) identifies a set of public cloud datacenters of one or more public cloud providers to connect a set of machines of the entity, (2) deploys managed forwarding nodes (MFNs) for the entity in the identified set of public cloud datacenters, and then (3) configures the MFNs to implement a virtual network that connects the entity's set of machines across its identified set of public cloud datacenters. In some embodiments, the method identifies the set of public cloud datacenters for an entity by receiving input from the entity's network administrator. In some embodiments, this input specifies the public cloud providers to use and/or the public cloud regions in which the virtual network should be defined. Conjunctively, or alternatively, this input in some embodiments specifies actual public cloud datacenters to use.

Some embodiments provide a novel method for deploying different virtual networks over several public cloud datacenters for different entities. For each entity, the method (1) identifies a set of public cloud datacenters of one or more public cloud providers to connect a set of machines of the entity, (2) deploys managed forwarding nodes (MFNs) for the entity in the identified set of public cloud datacenters, and then (3) configures the MFNs to implement a virtual network that connects the entity's set of machines across its identified set of public cloud datacenters. In some embodiments, the method identifies the set of public cloud datacenters for an entity by receiving input from the entity's network administrator. In some embodiments, this input specifies the public cloud providers to use and/or the public cloud regions in which the virtual network should be defined. Conjunctively, or alternatively, this input in some embodiments specifies actual public cloud datacenters to use.

An electronic control unit includes a relay device that is connected via a first network with a different relay device included in a different electronic control unit to relay a frame via the first network. It is determined whether a reception interruption has occurred. The reception interruption signifies that at least one predetermined frame scheduled to be transmitted from the different relay device is not received within a predetermined time via the first network. In response to the reception interruption being determined to have occurred, it is determined whether an abnormality has occurred in the first network based on at least one of (i) a presence or absence of reception of a state data representing a state of the different electronic control unit from the different electronic control unit via a second network within a fixed time, and (ii) a content of the state data.

A method for deploying different virtual networks over several public cloud datacenters for different entities. For each entity, the method (1) identifies a set of public cloud datacenters of one or more public cloud providers to connect a set of machines of the entity, (2) deploys managed forwarding nodes (MFNs) for the entity in the identified set of public cloud datacenters, and then (3) configures the MFNs to implement a virtual network that connects the entity's set of machines across its identified set of public cloud datacenters. In some embodiments, the method identifies the set of public cloud datacenters for an entity by receiving input from the entity's network administrator. In some embodiments, this input specifies the public cloud providers to use and/or the public cloud regions in which the virtual network should be defined. Conjunctively, or alternatively, this input in some embodiments specifies actual public cloud datacenters to use.

Some embodiments provide a novel method for deploying different virtual networks over several public cloud datacenters for different entities. For each entity, the method (1) identifies a set of public cloud datacenters of one or more public cloud providers to connect a set of machines of the entity, (2) deploys managed forwarding nodes (MFNs) for the entity in the identified set of public cloud datacenters, and then (3) configures the MFNs to implement a virtual network that connects the entity's set of machines across its identified set of public cloud datacenters. In some embodiments, the method identifies the set of public cloud datacenters for an entity by receiving input from the entity's network administrator. In some embodiments, this input specifies the public cloud providers to use and/or the public cloud regions in which the virtual network should be defined. Conjunctively, or alternatively, this input in some embodiments specifies actual public cloud datacenters to use.

Embodiments provide a multiaxial motor control system for controlling motors for a plurality of shafts included in a multiaxial machine, and including a plurality of motor control devices and a controller. The controller is connected with the motor control devices, and transmits a command signal to the motor control devices. Each motor control device includes a communication controller, a rotation controller, and a drive unit, and drives a motor of a corresponding shaft. The communication controller transmits and receives signals including the command signal, and determine whether the command signal is received normally. The rotation controller generates a torque command to operate the corresponding motor. The drive unit generates a drive voltage for electrification to drive the corresponding motor in accordance with the torque command. When a motor control device detects failure in reception, the motor control device outputs a torque command for braking torque to stop the corresponding motor.

Techniques are described for managing a plurality of requests for a change in bandwidth for a resource reservation label switched path (“LSP”) of a ring network. For example, a method may include receiving, by a network device and from an egress network device of a ring network, a resource reservation request message to establish a multipoint-to-point (MP2P) ring label switched path (LSP). The method may also include modifying, by the network device, the resource reservation request message to specify a request for a change in bandwidth for a segment from the network device to the egress network device of the MP2P ring LSP. The method may also include sending, by the network device and to a downstream network device along the MP2P ring LSP, the modified resource reservation request message.

A ring-based switch has nodes with a link management logic having forward and reverse link outputs couplable to other nodes, forward and reverse link inputs adapted couplable to other nodes, and memory coupled to the link management logic. The link management logic has a first mode where packet bursts are received through the forward link input and transmitted on its forward link output. The link management logic has a second mode where packet bursts are received through the forward link input and transmitted on its reverse link output; and a third mode where packet bursts are received through the reverse link input and transmitted on its forward link output. The node transmits test packets over the forward link output and, if no acknowledgment is received over the reverse link input within a predetermined test-time interval, the link management logic configures in the second mode.