Highly performant reliable message storage using in-memory replication technology

A method for providing data storage in a computing environment, the method comprising:

• executing, at a plurality of processing nodes, wherein each processing node includes a processor and a physical memory, one or more application server instances, wherein each application server instance is associated with one or more replicated stores;
• providing a replicated store daemon cluster, the replicated store daemon cluster comprising a plurality of replicated store daemons, including a replicated store daemon operating at and associated with each processing node, wherein each replicated store daemon provides access to the physical memory of the processing node the replicated store daemon is operating at and associated with, wherein each replicated store daemon comprises a RDMA (remote direct memory access) interface;
• two or more replicated stores, each of the two or more replicated stores being associated with a processing node of the plurality of processing nodes;
• opening, by a replicated store of the two or more replicated stores at a first processing node of the plurality of processing nodes, a fixed-size primary region on the physical memory on the first processing node, the fixed-size primary region being managed by a replicated store daemon associated with the first processing node, wherein the replicated store daemon associated with the first processing node places a lock on the fixed-size primary region;
• storing, by the replicated store at the first processing node, data in the fixed-size primary region;
• and
• persisting, by the replicated store daemon associated with the first processing node, the data to a fixed-size secondary region in a physical memory on another processing node, the physical memory on the another processing node being managed by a replicated store daemon associated with the another processing node of the plurality of processing nodes, wherein the replicated store daemon associated with the another processing node places a lock on the fixed-size secondary region;
• wherein the data is stored in the fixed-size primary region in the physical memory on the first processing node simultaneously with the data being persisted in the fixed-sized secondary region in the physical memory on the another processing node;
• wherein the fixed size primary region opened by the replicated store at the first processing node is unique to the replicated store and is not accessible by any other of the two or more replicated stores at the first processing node;
• and
• wherein the lock placed by the replicated stored daemon associated with the first processing node and the lock placed by the replicated store daemon associated with the another processing node are stored in the global directory.

A non-transitory computer readable storage medium, including instructions stored thereon for providing data storage in a computing environment which when read and executed by one or more computers cause the one or more computers to perform the steps comprising:

• executing, at a plurality of processing nodes, wherein each processing node includes a processor and a physical memory, one or more application server instances, wherein each application server instance is associated with one or more replicated stores;
• providing a replicated store daemon cluster, the replicated store daemon cluster comprising a plurality of replicated store daemons, including a replicated store daemon operating at and associated with each processing node, wherein each replicated store daemon provides access to the physical memory of the processing node the replicated store daemon is operating at and associated with, wherein each replicated store daemon comprises a RDMA (remote direct memory access) interface;
• two or more replicated stores, each of the two or more replicated stores being associated with a processing node of the plurality of processing nodes;
• opening, by a replicated store of the two or more replicated stores at a first processing node of the plurality of processing nodes, a fixed-size primary region on the physical memory on the first processing node, the fixed-size primary region being managed by a replicated store daemon associated with the first processing node, wherein the replicated store daemon associated with the first processing node places a lock on the fixed-size primary region;
• storing, by the replicated store at the first processing node, data in the fixed-size primary region;
• and
• persisting, by the replicated store daemon associated with the first processing node, the data to a fixed-size secondary region in a physical memory on another processing node, the physical memory on the another processing node being managed by a replicated store daemon associated with the another processing node of the plurality of processing nodes, wherein the replicated store daemon associated with the another processing node places a lock on the fixed-size secondary region;
• wherein the data is stored in the fixed-size primary region in the physical memory on the first processing node simultaneously with the data being persisted in the fixed-sized secondary region in the physical memory on the another processing node;
• and
• wherein the fixed size primary region opened by the replicated store at the first processing node is unique to the replicated store and is not accessible by any other of the two or more replicated stores at the first processing node;
• and
• wherein the lock placed by the replicated stored daemon associated with the first processing node and the lock placed by the replicated store daemon associated with the another processing node are stored in the global directory.