Transparent checkpointing and process migration in a distributed system

A node, comprising:

• a processor configured to:
• record a first set of memory locations to a buffer during a first checkpoint interval;
• and
• create a checkpoint from contents of the first set of memory locations, while recording a second set of memory locations to the buffer during a second checkpoint interval,
• wherein the processor is configured to create the checkpoint by a store of the contents of the first set of memory locations in a certain order and is configured to store the contents of a memory location from the first set of memory locations earlier than it would otherwise be stored when the processor needs to modify the memory location while the processor creates the checkpoint.

A node, comprising:

• a processor configured to:
• record a first set of memory locations modified by the processor to a buffer during a first checkpoint interval;
• create a checkpoint from contents of the first set of memory locations, while a second set of memory locations modified by the processor is recorded to the buffer during a second checkpoint interval;
• store in the memory a copy of each message output from the node during the process until an acknowledgement is received;
• output each message copied in the memory that does not receive an acknowledgement;
• and
• output an acknowledgement for each message received.

A node, comprising:

• a processor configured to store in a memory an identifier and a copy of each message of a plurality of messages output from the node until an acknowledgement is received,
• wherein the identifier and the copy of each message of a plurality of messages output are used to create a checkpoint;
• wherein a failure is determined to occur when the identifier differs from other identifiers on other nodes while processes are executed on the other nodes, and
• wherein when a failure occurs, the processor is further configured to output each message of a plurality of messages copied in the memory that does not receive an acknowledgement.

A node, comprising:

• a processor configured to:
• output an acknowledgement for each message of a plurality of messages received;
• create a checkpoint;
• and
• change an identifier upon the checkpoint being completed;
• wherein a failure is determined to occur when the identifier differs from other identifiers on other nodes, and
• wherein if a failure occurs, the processor is further configured to discard duplicate messages received by the node after the node is rolled back to the checkpoint.