Message redirection within a messaging infrastructure

At a computer system including one or more processors and system memory, the computer system included in a group of computer systems, each computer system in the group of computer systems configured to implement instances of a workflow representing a service such that the group of computer systems can simultaneously implement a plurality of instances of the workflow to simultaneously provide a plurality of instances of the service to other client computer systems, the computer system including a visible messaging endpoint that is visible to client computer systems, the visible endpoint configured to receive messages from client computer systems, the computer system including a message redirector, the message redirector configured to intercept messages received at the visible endpoint, the message redirector including a redirect client endpoint, the redirect client endpoint configured to:

• a) send intercepted client messages to the service, b) receive replies related to intercepted client messages from the service, and c) redirect client messages to other computer systems when so indicated, the service including an internal service endpoint for receiving messages from the redirect client endpoint and the internal redirect client input, a computer implemented method for redirecting a client message to the appropriate workflow instance, the computer implemented method comprising:
• an act of the message redirector intercepting a client message that was received at the visible messaging endpoint, the client message received from a client computer system and directed to the service;
• an act of the message redirector outputting the client message on the redirect client endpoint to send the client message to the service;
• an act of the service receiving the client message on the internal service endpoint;
• an act of the service sending a lock request for a new instance of the workflow, the lock request sent to a shared database, the shared database being shared among the plurality of computer systems in the group of computer systems, the new instance of the workflow to provide an instance of the service to the client computer system;
• an act of receiving a lock reply from the shared database, the lock reply indicating that another workflow instance already has a lock on client messages from the client computer system, the lock reply identifying an endpoint corresponding the other workflow instance at another computer system in the group of computer systems;
• an act of the service outputting a redirect fault on the internal service endpoint to send the redirect fault to the message redirector, the redirect fault including the identified endpoint;
• an act of the message redirector receiving the redirect fault on the redirect client endpoint;
• and
• an act of the message redirector redirecting the client message to the identified endpoint at the other computer system, the client message redirected for delivery to the other workflow instance.

A networked system, the networked system comprising:

• a load balancer;
• a plurality of computer systems grouped together in a scale out farm, the plurality of computer systems each configured to provide service workflow instances of a service to client computer systems;
• and
• a shared database, the shared database configured to store state for service workflow instances instantiated at computer systems in the plurality of computer systems and configured to store lock information for client messages that are locked to a particular service workflow instance;
• wherein the load balancer implements a load balancing algorithm to determine how to balance the distribution of client messages between computer systems in the plurality of computer systems, and
• wherein each computer system in the plurality of computer systems, includes:
• one or more processors;
• system memory, including a memory area to cache client messages directed to the service;
• and
• one or more computer storage media having stored thereon computer-executable instructions representing a message redirector and having stored thereon computer-executable instructions representing the service, the message redirector comprising:
• a connection to a visible endpoint that clients can utilize to send client messages to the service and return replies for client messages back to the clients;
• an interception module configured to intercept client messages received at the visible endpoint;
• a redirect client endpoint, the redirect client endpoint configured to:
• a) send intercepted client messages to the service;
• b) receive replies related to intercepted client messages from the service when the service processes the client message;
• c) receive redirect faults from the service when the service determines that client messages are locked to a service workflow instance at another computer system in the plurality of computer systems;
• d) redirect client messages to other computer systems in the plurality of computer systems in response to receiving redirect faults for client messages;
• and
• e) receive a replies responsive to redirected client messages from other computer systems in the plurality of computer systems;
• a redirect service endpoint, the redirect client endpoint configured to:
• a) receive redirected client messages sent from other computer systems in the plurality of computer systems, the redirected messages being directed to a service workflow instance at the computer system;
• and
• b) send replies to redirected client messages back to other computer systems;
• an internal redirect client endpoint, the internal redirect client endpoint configured to:
• a) send redirected client messages received at the redirect service endpoint to the service;
• and
• b) receive replies for redirected client messages;
• and
• the service comprising:
• one or more workflow instances, each workflow instance providing an instance of the service;
• an internal service client endpoint, the internal service client endpoint configured to:
• a) receive client messages from the redirect client endpoint;
• b) send replies for client messages that are processed by service workflow instances of the service to the redirect client endpoint;
• c) receive redirected client messages from the internal redirect client endpoint;
• and
• d) send replies for redirected client messages that are processed by service workflow instances of the service to the internal redirect client endpoint;
• and
• internal logic for transferring a received client message to the appropriate workflow instance form among the one or more workflow instances;
• a connection to the shared database, the connection enabling the service to access state data from the shared database and to set and access the lock information, setting lock information permitting the service to lock related client messages to a service workflow instance, accessing the lock information permitting the service to issue redirect faults for client messages that are to be redirected to other computer systems, the redirect faults indicating to the message redirector that the appropriate service workflow instance for processing the client messages is at another computer system, the redirect faults including identifiers for redirect service endpoints such that the message redirector can redirect the client messages to the appropriate service workflow instance.