System and method supporting highly-available replicated computing applications using deterministic virtual machines

A system comprising:

• multiple computing nodes configured to execute replicated virtual machines within a cluster in a distributed computing network, the multiple computing nodes configured to execute an application of the virtual machines, the application of each virtual machine comprising bytecode;
• wherein each computing node, when executing the associated virtual machine, is configured to:
• sequentially record, in a replicated transaction log used by the virtual machines, transactions of the virtual machine and one or more upgrades to the bytecode of the application;
• and
• generate, for the computing node, one or more snapshots and truncate the replicated transaction log, each snapshot capturing an application state and the bytecode of the associated virtual machine;
• wherein, in response to restoration of at least one of the computing nodes in the cluster or an addition of at least one additional computing node to the cluster, at least one of the computing nodes is configured to restore one or more of the virtual machines to a current application state of one or more others of the virtual machines using the transaction log and at least one of the one or more snapshots;
• and
• wherein the computing nodes are further configured to:
• determine that an effectful operation is to be performed on behalf of the virtual machines;
• write an effect intention into at least one memory associated with at least one of the virtual machines;
• obtain, via the transaction log, a result of the effectful operation as performed by an effectful subsystem in response to the effect intention;
• and
• use, at each of the virtual machines, the result of the effectful operation as contained in the transaction log.

A non-transitory computer readable medium containing instructions that when executed cause each of multiple computing nodes to:

• execute a replicated virtual machine within a cluster in a distributed computing network, the virtual machine including an application that is executed by the computing node, the application comprising bytecode;
• sequentially record, in a replicated transaction log used by the virtual machines, transactions of the virtual machine and one or more upgrades to the bytecode of the application;
• generate one or more snapshots and truncate the replicated transaction log, each snapshot capturing an application state and the bytecode of the virtual machine;
• in response to restoration of the computing node, restore the virtual machine to a current application state of one or more others of the virtual machines using the transaction log and at least one of the one or more snapshots;
• determine that an effectful operation is to be performed on behalf of the virtual machine;
• write an effect intention into at least one memory associated with at least one of the virtual machines;
• obtain, via the transaction log, a result of the effectful operation as performed by an effectful subsystem in response to the effect intention;
• and
• use, at the virtual machine, the result of the effectful operation as contained in the transaction log.