Ericsson CTO talks cost per bit, open APIs and the future of the cloud

with Vish Nandlall, CTO Ericsson, North America, head of strategy and marketing

Vish Nandlall

    Vish Nandlall

Long-time telecom industry veteran Vish Nandlall joined Ericsson (NASDAQ:ERIC) in 2009 and was recently named the company's CTO of North America and head of strategy and marketing, where he not only determines Ericsson's long-term technology vision but also plays an integral role in customer strategy. Nandlall met with FierceBroadbandWireless Editor-in-Chief Sue Marek at the 4G World conference last month in Chicago to talk about some of the key issues facing operators as they transition to LTE networks.

FierceBroadbandWireless: I'm hearing operators talk a lot about the cost per bit of delivering mobile broadband. While that's always been a key metric, it seems to be getting a lot more attention lately. Why now?

Vish Nandlall: It's happening because of the consumption. In 2009, the cost per bit was 49 cents per megabyte based upon a consumption of less than 200 Mbps per month. Today we are closer to 1 Gigabyte of consumption per month. The cost has to come down. It grows in proportion to the consumption.

Everything goes through an S-curve of adoption. To a great extent companies are going to profit from data being commoditized because we will move away from subscription and more toward value-based. 

FierceBroadbandWireless: But that seems to be a very difficult transition.

Nandlall: It is, and that is the transition we are in the middle of right now. It is confusing people and creating a lot of fear in the industry. The model we are going to is unknown today.

FierceBroadbandWireless: How do you help your customers figure this out? Are you working with your customers to help them monetize this traffic?

Nandlall: Yes, we are trying to focus on platforms that give the operators an opportunity to experiment with business models. These platforms need to be more malleable than they have been in the past. They need to feed off of innovation. 

Back in the late 90s [our focus] was to create an advanced intelligence network. The whole idea was to create a service inside the voice switch, so operators could deploy more and more applications. The evolution of AIN is IMS. So, let's distribute the call processing. Let's distribute the subscriber data base. Let's create an applications factory.

But, you can't create an applications factory when you are only addressing a professional services community of only about 100,000. You need to have a full-scale Web platform where there are millions of web developers that can create applications for consumers and enterprises. What we are seeing now is that these platforms cater to a broader base of developers.

Ericsson has been creating overlays with IMS so that we can create API exposures.  can take a voice applications server and expose the APIs and that lets us layer call control and mid-call apps on top of voice and open it up as a platform for innovation for developers.

FierceBroadbandWireless: Open APIs is another common trend I'm hearing in the industry.

Nandlall: I see it a little differently. I think public APIs or open APIs--the likes of what Google [(NASDAQ:GOOG)] is doing with location APIs--is one economy that is interesting, but I think private APIs is an economy that is even more interesting. I think we can take something an anchor it to business fundamentals where the application developers are making money, but the carriers are making money as well.

FierceBroadbandWireless: How does that work? An operator would charge for this?

Nandlall: The issue is how to make this market grow. Private APIs are not new. I think the way these failed in the past is that they did not consider how the value-chain changed. For you to promote an API you need a developer, and if you charge a developer to use that API you have created friction in the market. Once you embrace that move from the direct to the indirect model and the developer and the app store are part of the ecosystem, then a better model is to offer the API for free and then when you sell your application, I get a cut.

It's really the structure of that business model that we didn't quite settle on. The technology outpaced what the business model was.

FierceBroadWireless: When would this model be appealing to an over-the-top provider? If the OTT provider could provide better service working with the carrier than what the provider could do on its own?

Nandlall: Right. OTT providers at the end of the day have a business case too. Google gives away a lot of things--they want to attract eyeballs, and then it's those eyeballs that they monetize. Google Maps, Gmail and any other application services that Google provides are free, they don't monetize it by charging for it. They monetize it through the revenue they get from advertising.

If you take that model and try to understand, why would Google want to develop applications with the operator? Would they still get the revenue? If Google could calculate that Google Voice is being used by people on a social network and when that service breaks, those people will stop using that platform. They might even develop an aversion to that platform, and then Google can't monetize their ads. Why not outsource that service to a provider that can deliver a more secure service that they can rely upon?

To me the key is that the various constituents need to define their area and then define their business models.

FierceBroadbandWireless: The operators are trying to monetize cloud services. And yet you have other players that offer portions of cloud services usually geared toward saving certain types of content. How do you define cloud services?

Nandlall: It's a vast subject and what we tend to miss is that it is really just a new delivery model. And that model has economics to be competitive with purpose-built services. Where is cloud going? I look at Amazon [(NASDAQ:AMZN)] Web services. They are easily the leader in public cloud. And that is the model that is taking hold in the industry. Amazon is the most powerful cloud provider that we have seen to date.

Their model has been greenfield enterprise applications. They have staked their ground. The question is if there is a place where operators can offer a more premium cloud service with service level agreements attached to it.

I think they do if they go after the legacy applications, not the greenfield applications. This is a model that is unproven. There are note a lot of momentum or indicators that say this is going to work. But we are betting on it. Verizon's [(NYSE:VZ)] acquisition of Terramark indicates that others are seeing promise in this strategy as well.  

What we are struggling with is how will this manifest. What is it going to be? Are enterprises going to be willing to buy computer storage or networking from an AT&T or Verizon vs. going to a Rackspace?  Just as how any other product works, it starts out with features and more features, and then suddenly the cloud has mission-critical information on it. And then the customer wants a robustness and a guarantee and that's where the carriers can come in.

If Amazon's service goes down, you get a refund. That doesn't help you if you have mission critical data. That type of assurance can come from the big players. How quickly can they go to that model?  Can they make a distinction between the cloud service that Amazon is offering and what they can deliver? And are people willing to pay a surcharge for it?