Method and apparatus for checkpointing in computer system

Claim 3

The method according to claim 1, wherein any one of the processes in which a fault has occurred is a fault process, and wherein the monitoring step includes the step of checking whether the fault process has performed inter-process communication since a most recent one of the global and local checkpoints in the fault process.

Claim 4

The method according to claim 2, wherein the rolling back step includes the step of determining to roll back only the fault process to a most recent checkpoint in the fault process, when the checking indicates that the fault process has performed no inter-process communication since the most recent checkpoint.

Claim 5

The method according to claim 2, wherein the rolling back step includes the step of determining to roll back the processes to the respectively corresponding global checkpoints, when the checking indicates that the fault process has performed inter-process communication since the most recent checkpoint in the fault process.

Claim 6

The method according to claim 1, further comprising the step of storing a record of the monitored inter-process communication and of the generated global and local checkpoints for each of the processes.

Claim 7

The method according to claim 5, wherein the rolling back step includes the step of determining processes to be rolled back and appropriate checkpoints to which each of the processes is to be rolled back, based on the stored record.

Claim 8

The method according to claim 6, wherein the determining step includes the step of determining to roll back only a fault process in which the fault has occurred to a most recent one of the global and local checkpoints in the fault process, when the stored record indicates that the fault process has performed no inter-process communication since the most recent checkpoint.

Claim 9

The method according to claim 6, wherein the determining step includes the step of determining, when the stored record indicates that a fault process in which the fault has occurred has performed inter-process communication since the most recent one of the global and local checkpoints in the fault process, to roll back to the appropriate checkpoints the processes on which the inter-process communication performed by the fault process had an effect, based on the stored record.

Claim 10

The method according to claim 6, wherein the determining step includes the step of determining the appropriate checkpoints such that the rolled back processes can be restarted without a contradiction.

Claim 11

The method according to claim 6, wherein the determining step includes the step of limiting the appropriate checkpoints to a most recent global checkpoint.

Claim 12

The method according to claim 1, further comprising the step of restarting the at least one process from the global or local checkpoint to which it is rolled back.

Claim 13

A computer system, comprising: means for executing a plurality of processes performing inter-process communication; means for monitoring the inter-process communication for each of the processes; means for generating a global checkpoint for each of the processes by synchronizing the checkpointing of the processes; means for generating a local checkpoint independently from the generation of the global checkpoint, at an arbitrary time determined respectively for each of the processes; and means for rolling back at least one of the processes to one of the global and local checkpoints determined based on the monitored inter-process communication when a fault occurs.

Claim 14

A computer system, comprising: means for executing a plurality of processes performing inter-process communication; means for monitoring the inter-process communication for each of the processes; means for generating a global checkpoint for each of the processes by synchronizing the checkpointing of the processes; means for generating a local checkpoint independently from the generation of the global checkpoint, at an arbitrary time determined respectively for each of the processes; and means for rolling back at least one of the processes to one of the global and local checkpoints determined based on the monitored inter-process communication when a fault occurs.

Claim 15

An article of manufacture, comprising: a computer-usable medium having computer-readable program code means embodied therein for causing a computer to function as a checkpointing manager for managing checkpointing for a plurality of processes performing inter-process communication, the computer-readable program code means including: first computer- readable program code means for causing the computer to receive a result from monitoring the inter-process communication for each of the processes; second computer readable program code means for causing the computer to command to generate a global checkpoint for each of the processes for synchronizing the checkpointing of the processes; third computer readable program code means for causing the computer to receive information of a local checkpoint for each of the processes, the local checkpoint being generated independently from the generation of the global checkpoint with a timing determined respectively for each of the processes; and fourth computer readable program code means for causing the computer to determine to roll back at least one of the processes to one of the global and local checkpoints based on the result of the monitoring, when a fault occurs.

Claim 16

An article of manufacture, comprising: a computer-usable medium having computer-readable program code means embodied therein for causing a computer to function as a checkpointing manager for managing checkpointing for a plurality of processes performing inter-process communication, the computer-readable program code means including: first computer- readable program code means for causing the computer to receive a result from monitoring the inter-process communication for each of the processes; second computer readable program code means for causing the computer to command to generate a global checkpoint for each of the processes for synchronizing the checkpointing of the processes; third computer readable program code means for causing the computer to receive information of a local checkpoint for each of the processes, the local checkpoint being generated independently from the generation of the global checkpoint with a timing determined respectively for each of the processes; and fourth computer readable program code means for causing the computer to determine to roll back at least one of the processes to one of the global and local checkpoints based on the result of the monitoring, when a fault occurs.