Applications are spread across multiple clouds, across continents, countries, counties, and cities. Not only public clouds are being leveraged as the application deployment, delivery, and management platform, but also mission-critical, high-performance, and secure applications and data stores are being deployed and delivered through private clouds. Some enterprises continue with traditional IT environments. Applications literally have to connect, access and use nearby, as well as remotely held, applications or databases often as a part of successfully fulfilling any brewing business process requirements. But applications connecting and collaborating with other applications in the vicinity or in off-premise environments are not that straightforward.

There can be transient faults in the way of accessing other applications. The network connectivity, the failure of the requested applications due to overload, the temporary unavailability of the application, and so on, are being touted as the challenges for applications talking to one another over any network.

Applications ought to be designed in such a way that they try again to connect and proceed with their task-fulfillment. If the application request fails, the application can wait and make another attempt. If necessary, this process can be repeated with increasing delays between retry attempts, until some maximum number of requests has been attempted. The delay can be increased incrementally or exponentially, depending on the type of failure and the probability that it'll be corrected during this time.