| wo 2019/155251 Claims PCT/IB2018/050749 Following is the claim for this invention: - 1> In this invention we have huge number of incoming writes(inserts or updates) to a database like document oriented database MongoDB. We here use a non-blocking queue for threads or processes to insert the objects(inserts or updates) ready to be pushed to the database. Then different worker processes or threads dequeue or remove objects from the non-blocking queue typically in batches of size allowed for bulk operations depending upon the database used and do bulk updates or inserts to the database. Also different worker processes or threads pushing bulk inserts or updates to the database will use different connections(one unique connection per worker process or thread) to the database server from the connection pool made available to it. Typically some databases spawn one process or thread per connection request and hence the above novel technique which helps us to handle bulk operations like inserts or updates for a database is the claim for this invention. |
In this invention we have huge number of incoming writes(inserts or updates) to a database like document oriented database MongoDB. We here use a non-blocking queue for threads or processes to insert the objects(inserts or updates) ready to be pushed to the database. Then different worker processes or threads dequeue or remove objects from the non-blocking queue typically in batches of size allowed for bulk operations depending upon the database used and do bulk updates or inserts to the database. Also different worker processes or threads pushing bulk inserts or updates to the database will use different connections(one unique connection per worker process or thread) to the database server from the connection pool made available to it. Typically some databases spawn one process or thread per connection request and hence the above architecture helps us to handle bulk operations like inserts or updates for a database.