AAN: A Vision for the Future
Application-aware networking (AAN), like many industry terms, can mean many things to many people, as it is used in wildly inconsistent ways by various market players. It is at once an industry "buzzword" and a pivotal vision for the future. True AAN, in which networks and applications modify their functioning automatically and interactively in relation to each other's requirements, and in which networks "dynamically" -- on a continuously readjusting and fully-automated basis -- optimize applications' functioning and efficiency, is at a very early stage. AAN will come to include automatically "flexing" bandwidth levels at each location, changing quality of service (QoS) levels, choosing alternative traffic paths, turning trunks up and down, and lighting fiber to reflect fluctuating application requirements.
Radically intensified demands have been placed increasingly on networks by the explosion in bandwidth use driven largely by proliferation of video -- but probably even more importantly driven by the rapid spread of two increasingly complex and geographically disparate modalities: mobility and cloud computing.
These trends make today's networks far more fluid in the end-user demands placed on them, the devices employed and locations served, as well as the applications deployed. Great leaps in mobility, device multiplicity, geographic distance, application variety, and sensitivity mean ever more pressing demands on networks, making it more important that they be able to "dynamically" (rapidly and automatically) respond to changes. Applications themselves are growing increasingly distributed, inter-dependent, and "execution environment"-sensitive.
According to the latest Heavy Reading Service Provider IT Insider, "Application-Aware Networking: The Carrier Perspective," as the network goes through this upheaval, it is ever harder to tell where problems are and where they will emerge. While network managers have an increasing responsibility for application flow and greater expectations placed on them to keep things under control, they have a declining ability to see how applications fit together and where problems are originating. End-user customers increasingly want networks to function like the experience they have with the cloud, paying by the drink, incorporating dynamic bandwidth, and combining intelligence of network and applications.
Comprehensively, AAN services will likely begin emerging in two to three years and become more pervasive and "normalized" from there. The longer-run shift under a software-defined networking (SDN)-based regime will be to an application performance interface (API)-managed environment in which applications and networks communicate in a comprehensively automated way. API-driven performance on demand will emerge with networks dynamically ensuring guaranteed class of service (CoS) performance levels. The MEF is currently working on defining relevant API standards. While AAN is linked with SDN, it does not require comprehensive SDN implementation as a pre-condition.
Some non-US carriers, including Telstra and BT, appear to be among those furthest along in leading the way toward AAN development, while some leading US-based carriers are reluctant even to discuss the subject. At the same time, many carriers use the AAN term to describe longstanding service modalities that can be seen as precursors of true AAN, predominantly visibility-oriented applications performance management (APM) and WAN optimization techniques.
ó Steve Koppman, Analyst at Large, Heavy Reading Service Provider IT Insider
Application-Aware Networking: The Carrier Perspective, a 17-page report in PDF format, is available as part of an annual subscription (six bimonthly issues) to Heavy Reading Service Provider IT Insider, priced at $1,595. Individual reports are available for $900.