With computing becoming more distributed than ever, the performance of applications has become more dependent on predictable access to network bandwidth that, for the most part, is anything but. As part of an effort to make application performance across distributed networks more consistent, Push Technology this week released an upgrade to a data distribution platform that creates a data fabric that automatically removes of out-of-date or redundant data that no longer needs to be transmitted across a network.
Written in Java, Diffusion 5.0 now sports a unified application programming interface that makes it easier to invoke Diffusion along with new authentication capabilities. A new queuing mechanism serves up to 85 percent more messages each second to 25 percent more clients, Push Technology CEO Sean Bowen says.
Other new features include support for session replication, topic replication, and failover of active topic sources, and the ability to implement control logic in any of the supported languages and then locate that logic outside of the Diffusion server.
With the rise of everything from mobile computing applications to the Internet of Things, pressure on the network is rising, Bowen says. Most IT organizations, however, can’t afford forklift upgrades to their network infrastructure to provide additional bandwidth. As such, the priority is to find ways to optimize existing network bandwidth in a way that supports the needs of a raft of distributed real-time applications, he says.
Diffusion essentially builds a fabric across a network by creating nodes that optimize the flow of data across the network, Bowen says. While networking vendors are clearly moving in this direction, the difference is that Diffusion uses a software approach to optimize the flow of data across the network, as opposed to requiring IT organizations to build a new network to accomplish many of the same goals.
Application developers in the age of the cloud are often held hostage by the amount of network bandwidth their applications can have access to at any given moment. The idea behind Diffusion is to create a platform at a higher level of abstraction above the network layer through which developers can flow data more efficiently without necessarily having to be dependent on the kindness of network administrators to achieve that goal.