Abstract: Systems and methods for multicast send duplication instead of replication in a high performance computing environment. A method can provide a plurality of switches, a plurality of hosts, the plurality of hosts being interconnected via the plurality of switches, wherein a host of the plurality of hosts comprises a multicast sender node, the sender node comprising a system image generation module and a current message sequence module. The method can organize the plurality of switches into two rails, the two or more rails providing redundant connectivity between the plurality of hosts. The method can send two or more duplicate multicast packets on different rails. Upon a receiving node receiving at least two versions of the same multicast packet, only one will be delivered to the communication stack/clients above the layer that handles the encapsulation header.

Abstract: Systems and methods for multicast send duplication instead of replication in a high performance computing environment. A method can provide a plurality of switches, a plurality of hosts, the plurality of hosts being interconnected via the plurality of switches, wherein a host of the plurality of hosts comprises a multicast sender node, the sender node comprising a system image generation module and a current message sequence module. The method can organize the plurality of switches into two rails, the two or more rails providing redundant connectivity between the plurality of hosts. The method can send two or more duplicate multicast packets on different rails. Upon a receiving node receiving at least two versions of the same multicast packet, only one will be delivered to the communication stack/clients above the layer that handles the encapsulation header.

Abstract: Systems and methods for multicast send duplication instead of replication in a high performance computing environment. A method can provide a plurality of switches, a plurality of hosts, the plurality of hosts being interconnected via the plurality of switches, wherein a host of the plurality of hosts comprises a multicast sender node, the sender node comprising a system image generation module and a current message sequence module. The method can organize the plurality of switches into two rails, the two or more rails providing redundant connectivity between the plurality of hosts. The method can send two or more duplicate multicast packets on different rails. Upon a receiving node receiving at least two versions of the same multicast packet, only one will be delivered to the communication stack/clients above the layer that handles the encapsulation header.

Abstract: Systems and methods for multicast send duplication instead of replication in a high performance computing environment. A method can provide a plurality of switches, a plurality of hosts, the plurality of hosts being interconnected via the plurality of switches, wherein a host of the plurality of hosts comprises a multicast sender node, the sender node comprising a system image generation module and a current message sequence module. The method can organize the plurality of switches into two rails, the two or more rails providing redundant connectivity between the plurality of hosts. The method can send two or more duplicate multicast packets on different rails. Upon a receiving node receiving at least two versions of the same multicast packet, only one will be delivered to the communication stack/clients above the layer that handles the encapsulation header.

Abstract: One embodiment of the present invention provides a distributed computer system comprising a group of nodes, said group of nodes having a first and a second node adapted to act as current monitor nodes, each node of the group of nodes having an identifier and a management layer being informed which node currently acts as the current monitor node. The first and second node each include a protocol server adapted to associate and to send an address to a node requesting an address in the group of nodes, and a memory adapted to store an association address-identifier of a node requesting an address. The protocol server is further capable of requesting the management layer of a node for an indication informing if said node acts as the current monitor node, and attributing, to the current monitor node, the ownership of a list of associations, said list comprising associations of the memory.

Abstract: A method for obtaining a quorum vote by a first node using a quorum cable, wherein the quorum cable comprises a first end connected to the first node and a second end connected to a second node, including determining whether the quorum cable is reserved by the second node using a remote reservation input (RRI) on the first end, if the quorum cable is not reserved by the second node asserting a local reservation input (LRI) on the first end by the first node, monitoring a local reservation output (LRO) on the first end by the first node to determine whether the LRO on the first end is asserted in response to asserting the LRI on the first end, and obtaining the quorum vote by the first node, if the LRO on the first end is asserted.

Abstract: One embodiment of the present invention provides a distributed computer system comprising a group of nodes, said group of nodes having a first and a second node adapted to act as current monitor nodes, each node of the group of nodes having an identifier and a management layer being informed which node currently acts as the current monitor node. The first and second node each include a protocol server adapted to associate and to send an address to a node requesting an address in the group of nodes, and a memory adapted to store an association address-identifier of a node requesting an address. The protocol server is further capable of requesting the management layer of a node for an indication informing if said node acts as the current monitor node, and attributing, to the current monitor node, the ownership of a list of associations, said list comprising associations of the memory.

Abstract: The present invention describes a computer network including a network membership manager. In particular, a group of nodes on a computer network are managed by a distributed membership manager. Nodes of the computer network contain membership managers that manage the interaction between the nodes. Management of the computer network includes propagating configuration data to the nodes, providing an election process for determining the master node within a group of nodes, and monitoring the health of each node so that a change in the configuration and/or management structure can be accommodated by the nodes of the network.