Traffic and geography based cognitive disaster recovery

A system for optimizing disaster recovery migration comprising:

• at least two data storage and processing sites each capable of alternatively serving as a primary data storage site and a backup data storage site;
• an analysis module at each of said sites configured to cognitively analyze at least one system related parameter during system operation, wherein said at least one system related parameter includes an occurrence of a critical data transaction;
• a machine learning module at each site configured to predict, using the analyzed system related parameter, one or more patterns of said at least one system related parameter, wherein said machine learning module is configured to predict a time required to complete a currently active data transaction;
• and
• a control module configured to perform disaster recovery migration procedures based on said one or more predicted patterns, wherein the control module is configured to delay disaster recovery migration procedures until the predicted time of completion of the currently active data transaction and then perform the disaster recovery migration procedures.

A computer program product for optimizing disaster recovery migration in a system having first and second remote data storage and processing sites each capable of alternatively serving as a primary data storage and processing site and a backup data storage and processing site, the computer program product comprising one or more computer readable storage media collectively having program instructions embodied therewith, the program instructions executable by processors at each data storage and processing site to cause the system to:

• cognitively analyze at least one system related parameter during system operation, wherein said at least one system related parameter includes data traffic at said first and second sites;
• from the analyzed system related parameter predict, using machine learning, one or more patterns of said data traffic;
• and
• perform disaster recovery migration procedures based on said one or more predicted patterns, wherein performing said disaster recovery migration procedures include:
• determining that said one or more predicted patterns indicate greater data traffic near said second site serving as a backup site of said first site;
• determining a next time slot for performing a migration based on said one or more predicted patterns and a time of completion of all current transactions;
• and
• performing said migration at said next time slot by switching said second site from the backup site of said first site to a primary site and switching said first site from a primary site to a backup site of said switched second site.

A computer program product for optimizing disaster recovery migration in a system having at least two data storage and processing sites each capable of alternatively serving as a primary data storage site and a backup data storage site, the computer program product comprising one or more computer readable storage media collectively having program instructions embodied therewith, the program instructions executable by processors at each data storage and processing site to cause the system to:

• cognitively analyze at least one system related parameter during system operation, wherein said at least one system related parameter includes an occurrence of a critical data transaction;
• from the analyzed system related parameter predict, using machine learning, one or more patterns of said at least one system related parameter, wherein said program instructions are executable to predict a time required to complete a currently active data transaction;
• and
• perform disaster recovery migration procedures based on said one or more predicted patterns, wherein said program instructions are executable to delay disaster recovery migration procedures until the predicted time of completion of the currently active data transaction and then perform the disaster recovery migration procedures.