AT&T is already realizing the benefits from its shift to the New IP, both in terms of new service offerings and capex savings. Today's addition of a Managed Internet Service on its Network on Demand platform proves the first point, and recent comments by its CFO highlight the latter.
But as one of its top research technologists admitted yesterday in the New IP's Tune-In Tuesday radio show, those same benefits aren't necessarily reaching smaller operators in the telecom space, which may be struggling with their own deployments of key elements of the New IP, such as virtualization.
But first the good news: Managed Internet Service joins Ethernet services on the Network on Demand platform, allowing businesses to add Internet service as needed without requiring additional router hardware on premises because the routing function is virtualized in the core network and at the customer location. Services can be turned up quickly and businesses can order additional ports, add or change services or adjust bandwidth as needed in near real-time, paying for what they use. (See AT&T Launches Managed Internet on Demand.)
Chris Rice, VP of AT&T Labs, on Tune in Tuesday called Network on Demand a "groundbreaking service" and a "global first," for its ability to let customers change their Ethernet services in real-time and only pay for what they use. Now that same capability is applied to Managed Internet Service, which meets AT&T Inc. (NYSE: T)'s goal of adding new revenue-generating services more quickly. "We wanted to be able to simplify and accelerate how quickly we can bring services to market," he noted. (Listen to SDN: Beyond Bandwidth on Demand.)
AT&T is using Cisco Systems Inc. (Nasdaq: CSCO)'s virtualized CPE to enable the changes in routing required without needing to swap out hardware on premises, adding fuel to the argument that vCPE is the hottest trend in the network functions virtualization space. (See Cisco Powers AT&T's Virtual CPE Push.)
Rice also provided a more in-depth explanation of why AT&T CFO John Stephens told a Goldman-Sachs investment conference earlier this month that the carrier is already reaping some capex benefits from its pioneering SDN efforts. (See Following the Bouncing Capex Ball.)
Some of capex benefits come from separating hardware and software, allowing each to evolve at its more natural pace and some come from deployment of open source.
"As we move these network functions from these custom hardware solutions [to more commodity hardware], we are lowering the bar for the number of companies that can be part of that ecosystem," Rice noted during the interview. "You are growing that ecosystem, it will have more entrants, which means more options and generally speaking, more options means lower prices."
The move to open source has multiple benefits but one is that, by using OpenStack and OpenDaylight , for instance, AT&T can create a support system layer that is more common, and not depend as much on the operations and support systems or network element management systems that are vendor-specific, Rice said.
"That infrastructure layer becomes more common and you make a common investment, so you get an advantage in that you are not buying from each vendor," Rice said.
Getting the most from its capex is a critical issue for AT&T which, as Rice noted, has spent $140 billion in capex since 2009, or more than any other US company, as it tries to keep up with demand for mobile data bandwidth.
The AT&T Labs exec stated, however, that because of its size and the work already put in on SDN and NFV, the telecom giant has an advantage over smaller companies or others that are just going through the processes AT&T has tackled in the past couple of years. During that time, the company has gone through and catalogued literally hundreds of network functions and identified those that are relevant to its future and therefore need to be virtualized.
The oft-repeated company promise of virtualizing 75% of those by 2020 is an achievable goal, he added, beginning with 5% standardization this year.
"Large service providers do have an advantage here," he said. "We have the resources and the personnel to build these scalable systems. Everything isn't set right now. Some companies are struggling because they need to define what their strategy is, and they are getting caught up in the churn that is out there in the ecosystem. We are already delivering results for our customers because we've been through that. You will see a push from us for that 2020 goal and the time between service launches will drop."
— Carol Wilson, Editor-at-Large, Light Reading