It might sound crazy, but some IT wizards are already looking beyond software defined networking (SDN) to the next phase of architectural development: application aware networking (AAN).
Some of this thinking is already trickling into the market with platforms like Cisco’s Application Centric Infrastructure (ACI), with the idea being that network provisioning and management functionality are much more efficient when embedded in the application layer. With the right governance and usage policies in place, the enterprise could find itself sitting on top of a fully automated, broadly dynamic infrastructure capable of accommodating virtually any requirement users throw at it.
As tech analyst Steve Koppman noted recently, however, real AAN is in a very rudimentary stage. To fully realize the technology’s capabilities, you need to build networks and applications that can dynamically adjust and readjust their functionalityas conditions and user requirements change. This might include modifications to bandwidth consumption, QoS levels, traffic configuration and numerous other parameters at multiple points in the network pathway. Ultimately, though, AAN will be not just a desireable network trait, but a real necessity if the enterprise hopes to cope with increasingly complex and disparate cloud and mobile environments.
Another way to look at AAN is the introduction of “application fluency” to network infrastructure, says F5 Networks’ Colin Walker. In software development, app fluency is the ability to configure an app to change its behavior as user patterns and operating conditions become known. On the network, this means using metrics and feedback to change network elements, even entire architectures, to suit application loads. In the end, networking becomes more about supporting application environments than about shuttling bits from point to point.
As you might expect, much of this will center around a highly intelligent network performance layer that can accommodate not just highly complex network infrastructure and increasingly diverse application sets, but also the interplay between these two layers that takes place in modern production environments. According to ManageEngine’s Suvish Viswanathan, building this kind of functionality requires the enterprise to focus on five key goals: improved visibility, better understanding of business co-relationships, rapid troubleshooting, greater insight into user requirements, and methods to improve performance and lower costs.
Ideally, AAN capabilities should stretch across wired and wireless infrastructure and across virtually any device that enterprise apps and data might encounter. That’s the idea behind Fluke Networks’ Visual TruView platform, which was recently integrated into the company’s OptiView XG analysis tablets to provide broad extensibility into increasingly divergent network topologies. The system offers a unified view of the enterprise network, including remote sites, cloud providers and mobile infrastructure, and can quickly isolate and identify delivery path issues regardless of source or location. In this way, the enterprise can provide sustained app performance even if the problems being encountered are not on the network or even in owned-and-operated systems.
The very fact that people are starting to talk about application aware networking is an indication of the momentous changes taking place in and around the enterprise. Time was, capabilities were limited by the constraints of physical infrastructure. Once the enterprise adopts an all-software footing, the ability to craft optimized data environments will become second nature.
And then we can start defining data environments according to our goals and desires rather than our network limitations.
Photo courtesy of Shutterstock.
