AT&T's Network on Demand: Faster, More Scalable ServicesAT&T's Network on Demand: Faster, More Scalable Services

AT&T Network on Demand is a software defined network (SDN) offering that gives IT staff control of ordering new service, or increasing/decreasing bandwidth in near-real time.

One of the most frequent complaints about working with communications service providers is lengthy installation times. Such delays handicap IT organizations not only for adding new circuits, but also when they want to bolster bandwidth to support real-time apps or unexpected increases in business.

AT&T may have an answer. At last week's analyst conference in Dallas, AT&T SVP Roman Pacewicz discussed and demonstrated AT&T Network on Demand, a software defined network (SDN) offering that gives the IT staff control of ordering new service, or increasing/decreasing bandwidth in near-real time. Rather than waiting weeks or months for new service, AT&T says it will deliver in two to three days, on average.

There are a few caveats. The location ordering service must have fiber to the building (and if it doesn't, AT&T will add it, but that can take a few weeks). The enterprise must also have a universal port installed in order to increase or decrease bandwidth according to demand--a function that will take only a few minutes to take effect.

"We believe all our services over time will move onto this platform," Pacewicz said. "We're trying to move the network to the same model as a computing model."

Yes, every company is a software company now.

The drivers for AT&T's offering? The growing amount of data created every day (more than 15 Petabytes); the increase in video traffic (more than 60% of traffic on AT&T's network is video); more global workers; and the increased adoption of cloud.

What was impressive in the demo was the simplicity of the tool. The GUI is a diagram format, showing each office location and the circuits connecting them. Green, pulsating lines indicate all is well, and they change to red when a connection issue emerges. Drilling down on the circuit, IT staffs can update via drop-down menus the port interface, service type, bandwidth, class of service, and (gasp!) the actual date for provisioning.

What's more, there is an "asset store" where business can implement basic firewalls, advanced firewalls, mobile VPNs, traffic visualizers, DDoS protection, application acceleration, and more. Using some of these partner apps, for example, an IT administrator can block or blacklist websites, and have those changes take effect within two minutes.

Moving forward, it's easy to imagine how such a service can make collaborative applications even more appealing when delivered from the cloud. The ability to truly scale bandwidth on the fly to support video conferences or broadcasts for telemedicine, education, or public safety is game changing. Or, imagine a cloud IP-telephony or UC offering that adjusts to seasonal fluctuations in business--resulting in right-sizing not only communications capacity, but also the costs associated with it.

Even more effective would be the ability to set parameters and policies that automatically kick in at the appropriate traffic levels, security levels, or application demands.

Intriguing, indeed. But admittedly, I found myself last week remembering excitement of decades past for "dynamic bandwidth allocation" with frame relay and ATM (asynchronous transfer mode) that never really materialized fully. That said, as AT&T (and its competitors) bring fiber into more and more buildings, it becomes much easier to offer, extend, and enhance services like this.

Baby steps first. But with customer pressure and input, I could absolutely see these capabilities becoming reality sooner rather than later. AT&T's initial service will be generally available in Austin, TX, in November, followed by rollouts in Dallas, San Antonio, Houston, and Los Angeles. Throughout 2015, rollouts will extend nationwide.