Efficient and reliable host distribution of totally ordered global state

A distributed computing system, comprising:

• a plurality of computing nodes that are coupled to one another by at least one electronic data network, the plurality of computing nodes including at least a first computing node and a second computing node,
• each one of the plurality of computing nodes including a processing system that includes a plurality of hardware processing cores that are coupled to common electronic memory that is accessible by each of the plurality of hardware processing cores on a respective computing node,
• the processing system of the first computing node configured to:
• receive an electronic data message;
• generate a global state electronic data message based on the received electronic data message, the global state electronic data message including a logical clock sequence identifier of a logical clock for the distributed computing system;
• and
• send, by using the at least one electronic data network, the annotated global state electronic data message to other ones of the plurality of computing nodes,
• the processing system of the second computing node configured to:
• execute a plurality of client computer processes, where different hardware cores execute different ones of the plurality of client computer processes;
• receive the global state electronic data message that was sent from the first computing node;
• store the global state electronic data message in the common electronic memory;
• and
• selectively deliver messages regarding the global state electronic data message to the plurality of client computer processes of the second computing node.

A distributed computing system, comprising:

• an electronic data communications network;
• a sequencer computing node that includes a processing system with at least one processor and memory, the processing system configured to:
• maintain a logical clock for the distributed computing system,
• execute a sequencer computer process that annotates received messages with a sequence identifier derived from the maintained global logical clock for the distributed computing system, and
• transmit, via the electronic data communications network, messages that have been annotated with a corresponding sequence identifier;
• and
• at least one monitor computing node coupled to the electronic data communications network and including a processing system with a plurality of hardware processing cores coupled to common data storage accessible by each of the plurality of hardware processing cores, the common data storage configured to store a cache that includes at least a partial subset of the global state, the processing system configured to:
• execute, across the plurality of hardware processing cores, a plurality of client computer processes,
• receive, via the electronic data communications network, annotated messages that have been transmitted from the sequencer computing node,
• add the annotated messages to the at least a partial subset of the global state that is stored in the cache of the common data storage, and
• selectively deliver a notification to at least one of the plurality of client computer processes executing on at least one of the plurality of processing cores, the notification regarding the annotated message that has been received from the sequencer computing node.

A method of implementing a distributed computing system over a plurality of computing nodes, where one of the plurality of computing nodes is a sequencer computing node and other ones of the plurality of computing nodes are monitor computing nodes, the method comprising:

• maintaining a logical clock for the distributed computing system;
• generating, on the sequencer computing node, an annotated electronic data message based on a received electronic data message, the annotated electronic data message including an identifier that is based on the maintained logical clock;
• publishing, using the electronic data communications network, the annotated electronic data message;
• receiving, on a first monitor computing node of the monitor computing nodes, the annotated electronic data message, the first monitor computing node including a plurality of hardware processing cores coupled to common data storage that is accessible by each of the plurality of hardware processing cores, the first monitor computing node having a plurality of local client computer processes executed thereon across the plurality of hardware processing cores;
• storing, to the cache located in the common data storage of the first monitor computing node, the annotated electronic data message as part of a local copy of a global state of the distributed computing system;
• responsive to reception of the annotated electronic data message, selectively notifying at least one local client computer process of the plurality of local client computer processes about the received annotated electronic data message.