Reliable replication mechanisms based on active-passive HFI protocols built on top of non-reliable multicast fabric implementations

An apparatus comprising:

• a host fabric interface (HFI) including,
• a transmit port, configured to send data onto an non-reliable fabric;
• a receive port, configured to receive data from the non-reliable fabric;
• wherein the HFI further is configured to,
• send a multicast message to a switch in the non-reliable fabric, wherein the multicast message is configured to cause the switch to generate and send a unicast message corresponding to the multicast message to each of the plurality of target nodes via the non-reliable fabric;
• maintain indicia identifying which target nodes the multicast message is to be delivered to;
• receive one or more reply messages from one or more of the plurality of target nodes via the non-reliable fabric, the one or more reply messages indicating that the target node sending the reply message has successfully received the unicast message corresponding to the multicast message sent to the target node;
• determine one or more reply messages have yet to be received from one or more of the target nodes within a timeout period;
• and
• in response thereto,
• generate and send a unicast message corresponding to the multicast message to each of the one or more target nodes that did not return a reply message within the timeout period.

A method for reliably delivering a multicast message from an originator node including a Host Fabric Interface (HFI) to a plurality of target nodes over a non-reliable fabric to which each of the originator node and the plurality of target nodes is coupled, comprising:

• generating the multicast message via a network software stack operating on the originator node and sending the multicast message from the HFI of the originator node to a switch in the non-reliable fabric, wherein the multicast message is configured to cause the switch to generate and send a unicast message corresponding to the multicast message to each of the plurality of target nodes, each unicast message including indicia identifying at least one of the HFI and the originator node;
• maintaining indicia at the originator node identifying which target nodes the multicast message is to be delivered to;
• receiving, at the HFI for the originator node, one or more reply messages from one or more of the plurality of target nodes, the one or more reply messages indicating that the target node sending the reply message has successfully received the unicast message corresponding to the multicast message sent to the target node;
• for each reply message received at the HFI, notifying the network software stack operating on the originator node;
• determining one or more reply messages have yet to be received from one or more of the target nodes within a timeout period;
• and
• in response thereto, for each target node for which a reply message has yet to be received,
• generating a unicast message corresponding to the multicast message via the network software stack and sending the unicast message via the HFI to each of the target node.

An apparatus, configured to be implemented as a node in a network including a plurality of nodes coupled in communication via a non-reliable fabric, the apparatus comprising:

• a host processor, coupled to memory;
• a storage device storing software instructions comprising a plurality of software modules including a network software stack;
• a host fabric interface (HFI) including,
• a transmit port, configured to send data onto the non-reliable fabric;
• a receive port, configured to receive data from the non-reliable fabric;
• wherein execution of the software instructions on the host processor or a virtual machine running on the host processor causes the apparatus to,
• generate a multicast message to be delivered to a plurality of target nodes and forward the multicast message to the HFI;
• and
• maintain indicia identifying which target nodes the multicast message is to be delivered to;
• wherein the HFI is configured to,
• send the multicast message to a switch in the non-reliable fabric, wherein the multicast message is configured to cause the switch to generate and send a unicast message corresponding to the multicast message to each of the plurality of target nodes;
• receive one or more reply messages from one or more of the plurality of target nodes, each of the one or more reply messages indicating that the target node sending the reply message has successfully received the unicast message corresponding to the multicast message sent to the target node;
• and
• for each reply message received at the HFI, notify the network software stack;
• wherein execution of the software instructions further causes the apparatus to,
• determine one or more reply messages have yet to be received from one or more of the target nodes within a timeout period;
• and
• in response thereto, for each target node for which a reply message has yet to be received,
• generate a unicast message corresponding to the multicast message via the network software stack and send the unicast message via the HFI to the target node.