Distributed data grid routing for clusters managed using container orchestration services

A method comprising:

• subscribing, by a reverse proxy server using an event driven architecture, to receive a current data grid topology of a container-based cluster of compute nodes;
• receiving, by the reverse proxy server executing on the container-based cluster of compute nodes managed using a container orchestration service, a first data grid topology of the container-based cluster of compute nodes at a first time;
• generating, by the reverse proxy server, a first segment-to-node map in response to receiving the first data grid topology;
• receiving, by the reverse proxy server from a client device, a first request to retrieve first data from the container-based cluster of compute nodes, the first request comprising a key of a key-value pair;
• computing, by the reverse proxy server, a hash value by inputting the key into a consistent hashing function executed by the reverse proxy server, wherein the consistent hashing function partitions a hash value space into a plurality of segments, where the segments are mapped to current nodes identified using the first segment-to-node map;
• determining, by the reverse proxy server, at least one node among the container-based cluster of compute nodes associated with the hash value using the first segment-to-node map;
• sending, by the reverse proxy server, the first request to the at least one node;
• receiving, by the reverse proxy server, the first data from the at least one node in response to the first request;
• sending the first data to the client device;
• receiving, at a second time using the event driven architecture, a second data grid topology of the container-based cluster of compute nodes;
• and
• generating, by the reverse proxy server, a second segment-to-node map in response to receiving the second data grid topology.

A non-transitory machine readable medium storing a program which, when executed by a processor, causes a reverse proxy server executing on a container-based cluster of compute nodes to:

• subscribe, using an event driven architecture, to receive a current data grid topology of the container-based cluster of compute nodes;
• receive, at a first time using the event driven architecture, a first data grid topology of the container-based cluster of compute nodes;
• generate a first segment-to-node map in response to receiving the first data grid topology;
• receive, from a client device, a first request to retrieve first data from the container-based cluster of compute nodes, the first request comprising a key of a key-value pair, wherein the container-based cluster of compute nodes is managed using a container orchestration service;
• compute a hash value by inputting the key into a consistent hashing function executed by the reverse proxy server, wherein the consistent hashing function partitions a hash value space into a plurality of segments, where the segments are mapped to current nodes identified using the first segment-to-node map;
• determine at least one node among the container-based cluster of compute nodes associated with the hash value using the first segment-to-node map;
• send the first request to the at least one node;
• receive the first data from the at least one node in response to the first request;
• send the first data to the client device;
• receive, at a second time using the event driven architecture, a second data grid topology of the container-based cluster of compute nodes;
• and
• generate a second segment-to-node map in response to receiving the second data grid topology.