Processing method, and computer for fault recovery

A stream data processing method for use with a stream data processing system having a processing section for receiving first tuples, which are time-stamped data, and generating second tuples by performing a query, which is a data process on the first tuples, the stream data processing method comprising:

• giving, by the processing section, a tuple recovery point to the second tuples as an attribute value, the tuple recovery point being the information identifying the oldest one of the first tuples used for generating the second tuples managed by the stream data processing system or identifying a first tuple which is received by the stream data reception section earlier than the oldest one of the first tuples used for generating the second tuples;
• where the query execution section adds tuple recovery point attribute values to the generated second tuples, the tuple recovery point attribute values being the information identifying the oldest one of the first tuples used for generating the second tuples or identifying the first tuple which is received by the stream data reception section earlier than the oldest one of the first tuples used for generating the second tuples;
• and
• determining a recovery point for the stream data processing system by conducting a search on all the tuple recovery point attribute values managed by the stream data processing system to locate the oldest recovery point.

A computer for use with a system, the computer comprising:

• a processing section that receives first tuples, which are time-stamped data, and generates second tuples by performing a query, which is a data process on the first tuples;
• and
• a storage section;
• wherein the processing section includes:
• a stream data reception section that receives the first tuples,
• a query execution section that generates the second tuples by performing a data process on the first tuples,
• a stream data transmission section that transmits the second tuples generated by the query execution section, and
• a recovery point management section that determines a recovery point for the system, the recovery point being the information identifying the oldest one of the first tuples used for generating the second tuples or identifying a first tuple which is received by the stream data reception section earlier than the oldest one of the first tuples used for generating the second tuples;
• wherein the query execution section adds tuple recovery point attribute values to the generated second tuples, the tuple recovery point attribute values being the information identifying the oldest one of the first tuples used for generating the second tuples or identifying the first tuple which is received by the stream data reception section earlier than the oldest one of the first tuples used for generating the second tuples;
• and
• wherein the recovery point management section determines the recovery point for the stream data processing system by conducting a search on all the tuple recovery point attribute values managed by the stream data processing system to locate the oldest recovery point.