Automated document replication in a distributed computing system
View Patent ↗In a method of automated document replication in a distributed computing system, a document at a first node of a distributed computing system is analyzed. Based on the analysis, it is determined whether the document has a multi-node affinity. In response to determining that the document has a multi-node affinity, at least one document is automatically replicated between nodes of the distributed computing system. The indicated nodes are indicated based on context of the multi-node affinity of the analyzed document.
1. A method of automated document replication in a distributed computing system, said method comprising:
analyzing, by a processor, references within data content of a first document at a first node of a distributed computing system;
determining, based on said analysis, whether said first document references documents stored external to said first node; and
in response to determining that said first document references documents external to said first node, automatically replicating at least one document between indicated nodes of said distributed computing system so as to facilitate local query processing of any joins to said first document, said indicated nodes indicated based on external references of said first document.
2. The method as recited in claim 1 , wherein said analyzing references within a first document at a first node of a distributed computing system comprises: analyzing said first document upon initial storage at said first node.
3. The method as recited in claim 1 , wherein said analyzing references within a first document at a first node of a distributed computing system comprises:
analyzing said first document as a part of a rebalancing of said distributed computing system.
4. The method as recited in claim 1 , wherein said analyzing references within a first document at a first node of a distributed computing system comprises:
analyzing said first document as part of a bootstrapping of said first node.
5. The method as recited in claim 1 , wherein said automatically replicating at least one document between indicated nodes of said distributed computing system so as to facilitate local query processing of any joins to said first document comprises:
automatically replicating, at said first node, all documents referenced by said first document which are not already stored at said first node.
6. The method as recited in claim 1 , wherein said automatically replicating at least one document between indicated nodes of said distributed computing system so as to facilitate local query processing of any joins to said first document comprises:
automatically replicating said first document on nodes, other than said first node, where documents referenced in said first document are located.