Mixed deployment architecture for distributed services

A method, comprising:

• accessing, by a data system, a computer application corresponding to a service for the data system, wherein the computer application is determined to be deployed on each of a plurality of distributed compute grid systems and not on a core compute grid system based on a first determination that the computer application is not a stateful application and that the computer application does not require synchronous communication with one or more applications provided via the plurality of distributed compute grid systems,
• wherein the data system comprises the core compute grid system and the plurality of distributed compute grid systems,
• wherein the plurality of distributed compute grid systems is configured to communicate with each other via a network,
• wherein a database application for providing data storage for the data system is determined to be implemented within the core compute grid system and not on the plurality of distributed compute grid systems based on a second determination that the database application is a stateful application, and
• wherein the plurality of distributed compute grid systems is configured to collectively provide the service based on one or more communications with the core compute grid system for accessing the database application;
• allocating, by the data system, a corresponding portion of processing capacity of each of the plurality of distributed compute grid systems to the computer application;
• receiving an indication that first stateful data associated with the computer application has been generated by a first distributed compute grid system of the plurality of distributed compute grid systems in response to processing an electronic transaction for a user of the service;
• determining, from a plurality of data replication routes, a ring data replication route for transmitting the first stateful data through the plurality of distributed compute grid systems;
• replicating, by the data system, the first stateful data to other distributed compute grid systems in the plurality of distributed compute grid systems according to the ring data replication route;
• and
• transmitting, via the data system, the first stateful data to the database application of the core compute grid system through a routing management layer.

A non-transitory computer-readable medium having stored thereon instructions executable by a first server to cause the first server to perform operations comprising:

• accessing a computer application corresponding to a service for a data system, wherein the computer application is determined to be deployed on each of a plurality of distributed compute grid systems and not on a core compute grid system based on a first determination that the computer application is not a stateful application and that the computer application does not require synchronous communication with one or more applications provided via the plurality of distributed compute grid systems,
• wherein the data system comprises the core compute grid system and the plurality of distributed compute grid systems,
• wherein the plurality of distributed compute grid systems is configured to communicate with each other via a network,
• wherein a database application for storing data for the data system is determined to be implemented within the core compute grid system based on a second determination that the database application is a stateful application, and
• wherein the plurality of distributed compute grid systems is configured to collectively provide the service based on one or more communications with the core compute grid system for accessing the database application;
• allocating a corresponding portion of processing capacity of each of the plurality of distributed compute grid systems to the computer application;
• receiving an indication that first stateful data associated with the computer application has been generated by a first distributed compute grid system of the plurality of distributed compute grid systems in response to processing an electronic transaction for a third-party user of the service;
• determining, from a plurality of data replication routes, a ring data replication route for the first stateful data based on connections among the plurality of distributed compute grid systems;
• replicating the first stateful data to other distributed compute grid systems in the plurality of distributed compute grid systems according to the ring data replication route;
• and
• transmitting the first stateful data to the database application of the core compute grid system through a routing management layer.