CPC H04L 63/029 (2013.01) [G06F 9/5083 (2013.01)] | 14 Claims |
1. A method for accessing data nodes of a data cluster, the method comprising:
obtaining, by a data access gateway (DAG), a request from a host; and
in response to the request:
obtaining bidding counters from data nodes, wherein the bidding counters comprise numerical representations of computational, network, and storage resources of the data nodes available to service the request;
obtaining metadata mappings from the data nodes, wherein:
a metadata mapping of the metadata mappings comprises a data node identifier associated with a data node of the data nodes corresponding to the metadata mapping and a plurality of chunk metadata; and
a chunk metadata of the plurality of chunk metadata may specify:
a chunk identifier associated with a corresponding chunk,
whether the chunk is a data chunk or a parity chunk,
whether the chunk is stored in an accelerator pool or a non-accelerator pool,
a storage location of the chunk, and
a data slice identifier associated with the chunk;
identifying, based on the bidding counters and metadata mappings, a data node of the data nodes associated with a highest bidding counter of the bidding counters and an appropriate metadata mapping of the metadata mappings, wherein the data node is associated with a highest amount of available computational, network, and storage resources, as specified by the highest bidding counter, compared to other data nodes of the data nodes; and
sending the request to the data node.
|