Dynamic adjustment of boxcarring of action requests from component-driven cloud applications

A user device running a middleware application, including:

• a middleware application, running on the user device, coupled in communication with a data consuming application and a browser or app running on the user device, and also coupled in communication with at least one server through the browser or app;
• wherein the middleware application receives action requests from the data consuming application and returns data received from the server responsive to the action requests;
• wherein the middleware application measures network communication latency, adjusts inter-boxcar intervals used to dispatch batches of action requests, batches the action requests into boxcars, and dispatches boxcars of batched action requests;
• wherein the middleware application calculates the measured network communication latency as dispatch-to-completed response time minus server processing time, using the server processing time received from the server for a boxcar of completed responses;
• and
• wherein the inter-boxcar intervals are adjusted, taking into account at least (a) a number of connections supported between the user device and the server and (b) the measured network communication latency.

A server running a dispatching application, including:

• the dispatcher running on the server and coupled in communication with a multiplicity (100 or more) of middleware applications running on respective user devices and one or more application programs running on the server;
• wherein the dispatcher receives boxcarred batches of action requests from the middleware applications, forwards the action requests to the application programs, and returns responsive boxcars of completed responses;
• and
• wherein the dispatcher further calculates a server processing time as a difference between when a particular boxcar of action requests is received and when a respective boxcar of completed responses is ready to be returned and further reports the calculated server processing time back to the middleware application for the respective boxcar of completed responses.

A method, including:

• running a middleware application on a user device, coupled in communication with a data consuming application and a browser or app running on the user device, and also coupled in communication with at least one server through the browser or app;
• the middleware application receiving action requests from the data consuming application and returning data received from the server responsive to the action requests;
• the middleware application measuring network communication latency, adjusting inter-boxcar intervals used to dispatch batches of action requests, batching the action requests into boxcars, and dispatching the boxcars of batched action requests;
• the middleware application calculating the measured network communication latency as dispatch-to-completed response time minus server processing time, using the server processing time received from the server for a boxcar of completed responses;
• and
• the middleware application adjusting the inter-boxcar intervals taking into account at least (a) a number of connections supported between the user device and the server and (b) the measured network communication latency.

A method that conserves connections between a browser or app running on a user device and at least one server, including:

• a dispatcher running on the server and coupled in communication with a multiplicity of middleware applications running on respective user devices and one or more application programs running on the server;
• the dispatcher receiving boxcarred batches of action requests from the middleware applications, forwarding the action requests to the application programs, and returning responsive boxcars of completed responses;
• and
• the dispatcher further calculating a server processing time as a difference between when a particular boxcar of action requests is received and when a respective boxcar of completed responses is ready to be returned and further reporting the calculated server processing time back to the middleware application for the respective boxcar of completed responses.