User-defined lightweight application checkpointing system

A method comprising:

• accessing a library that has been implemented in a process, wherein the library facilitates checkpoint-and-resume functionality to enable the process to checkpoint itself via the implemented library;
• using a set of tags to annotate code of the process, the set of tags defining which data of the process is to be saved in an event in which the process terminates in an unexpected manner, wherein the set of tags define blocks within the code, identify variables at a beginning of each block to be saved, and define checkpoint locations within the blocks;
• execute a preprocessing operation that converts the tags into executable code that maintains a log describing progress of the process through the blocks and checkpoints including an identification of a current block and a most recently passed checkpoint within that block;
• after the data has been saved, detecting that the process has terminated in the unexpected manner;
• restarting the process from a beginning state;
• from the beginning state of the process, causing the process to progress through the code, including skipping execution of code for which data was previously saved;
• continue progressing and skipping through the code until said progression reaches the defined block;
• and
• causing the process to resume at a beginning of the defined block corresponding to a most recently passed checkpoint, wherein values of variables identified by the tags are restored and other data is reset through execution of the process.

A computer system comprising:

• one or more processors;
• and
• one or more hardware storage devices that store instructions that are executable by the one or more processors to cause the computer system to:
• access a library that has been implemented in a process, wherein the library facilitates checkpoint-and-resume functionality to enable the process to checkpoint itself via the implemented library;
• use a set of tags to annotate code of the process, the set of tags defining which data of the process is to be saved in an event in which the process terminates in an unexpected manner, wherein the set of tags define blocks within the code, identify variables at a beginning of each block to be saved, and define checkpoint locations within the blocks;
• executing a preprocessing operation that converts the tags into executable code that maintains a log describing progress of the process through the blocks and checkpoints including an identification of a current block and a most recently passed checkpoint within that block;
• after the data has been saved, detect that the process has terminated in the unexpected manner;
• restart the process from a beginning state;
• from the beginning state of the process, cause the process to progress through the code, including skipping execution of code for which data was previously saved;
• continue to progress and to skip through the code until said progression reaches the defined block;
• and
• cause the process to resume at a beginning of the defined block corresponding to a most recently passed checkpoint, wherein values of variables identified by the tags are restored and other data is reset through execution of the process.