Failure transparency for update applications under single-master configuration

A method for eliminating programming by a programmer to rectify an error condition by masking data failures, when a master copy of data is unavailable, said method comprising:

• storing a single master copy of data and at least one read-only copy of said data;
• performing write operations to said master copy using a middleware component,
• wherein when said master copy is available, said write operations comprise any of updates, deletions, and insertions to said master copy;
• performing read operations from one of said master copy and said read-only copy using said middleware component, when said master copy is available;
• storing said write operations to be made to said master copy, when said master copy is unavailable, in a table located in said middleware component as stored write requests until said master copy becomes available;
• wherein said middleware component implements said stored write requests in said table, when said master copy is unavailable;
• performing said read operations, when said master copy is unavailable, from said table and said read-only copy;
• propagating said stored write requests to said master copy, when said master copy becomes available, after being unavailable;
• identifying a port to send a notification to, when said stored write requests cannot be propagated to said master copy due to an integrity constraint;
• and
• outputting said notification.

A program storage device readable by machine, tangibly embodying a program of instructions executable by said machine to perform a method for eliminating programming by a programmer to rectify an error condition by masking data failures, when a master copy of data is unavailable, said method comprising:

• storing a single master copy of data and at least one read-only copy of said data;
• performing write operations to said master copy using a middleware component,
• wherein when said master copy is available, said write operations comprise any of updates, deletions, and insertions to said master copy;
• performing read operations from one of said master copy and said read-only copy using said middleware component, when said master copy is available;
• storing said write operations, when said master copy is unavailable, to be made to said master copy in a table located in said middleware component, only as stored write requests until said master copy becomes available;
• wherein said middleware component implements said stored write requests in said table, when said master copy is unavailable;
• performing said read operations, when said master copy is unavailable, from said table and said read-only copy;
• propagating said stored write requests to said master copy, when said master copy becomes available, after being unavailable;
• identifying a port to send a notification to, when said stored write requests cannot be propagated to said master copy due to an integrity constraint;
• and
• outputting said notification.