DDS

Abstract

Servers are potential points of failure, whatever their workload or physical location is. The Internet and TCP/IP are designed to ensure traffic will reach its destination through alternate routes if specific nodes or larger parts of the network fail. However, the protocols we have explored together up to now all rely on some sort of server. Of course, you can make those servers more resilient by leveraging high-availability features and deploying redundant network connections. But the farther away you are getting from the Cloud or the corporate data center, the harsher the environment will be. For many real-time and mission-critical use cases, having a single point of failure in the architecture is unacceptable, especially since brokers and servers also add latency and jitter to the data flow.