NEW YORK ó More innovation is needed to bring wearables to a mass market with dedicated silicon, interoperability, and beauty leading the charge, according to a Qualcomm executive. The company already has a toehold on this emerging sector with its Snapdragon 400 chip designed in to the LG G and Moto 360 smartwatches -- both of which were featured in Google's Android Wear debut.
EE Times sat down with Rob Chandhok, Qualcomm's senior vice president and president of interactive platforms, at Wearable Tech Expo in New York City to talk about the company's strategy for wearables.
EE Times: What will it take for wearables to really take off beyond the early adopters?
Rob Chandhok: People wear different things at different times with different fashion, and rarely are people single designer or just like this look. What that leads you to is a place... where people need to be able to interchange wearables like they interchange jewelry in order for it to be mass marketed.
I want to, for example, take the guts of a Fitbit and put it inside jewelry holders. But when you do that you lose some of the functionality. There's the physical fashion side of it, which leads you to wear more brands, which leads you to interoperability.
EET: So essentially what we're talking about is a wearable module, something that you can click in place?
Chandhok: What I want is a way to make these different pretty things talk to each other using the same protocol... thinking about open standards. One of the things that will happen with Android Wear, at least in the Google ecosystem, is you will be able to interchange multiple Android Wear watches and hopefully not have to reconfigure them every time.
Right now I have to configure a different application and they don't know anything about each other. I want the operating system to say 'You have another path for notifications on your wrist, how would you like me to manage that?' Google has chosen to embody that by saying here's the protocol and the operating system [the watch is] going to run, and I don't know if that's going to hold up. I think the protocol is more important than the operating system.
EET: Pebble's Myriam Joire said that Android and Tizen aren't going to cut it for smartwatches, and you'll need something like a real-time OS that's lightweight in order to handle the speeds and dedicated functions. What do you think?
Chandhok: I think that's a statement that's valuable for a very short period of time. As we go down the process nodes, the power profile [of chips] will change pretty significantly. I understand the tradeoff [of RTOSs] and value days of use, so I don't think having a watch that only lasts eight to 10 hours or only one day will be successful in the market place.
I'm just not sure that I can make the claim that in three years we won't be able to do that with a Linux-based or high-level operating system. I think a lot of the IP will be similar. I just don't buy the argument that we're not going to be running modern OSes, even for things that are on our wrists, because we'll be at 14 nm or 10 nm at that point and how much power will it really take for me to run that system? It will be a wild, wild West for a little bit, but what will win is the software ecosystem. If I get a runtime that lets me do things easily and there's a good open-source community around it, it will move really fast.
EET: What do you think a winning chip is going to look like for a wearable?
Chandhok: Less important is the particulars of a chip but, rather, the architecture associated with it. If you want to do things at low power, there are ways you can do it by making things smaller... or having the software flexibility of a more sophisticated operating environment to manage the power better.
[In the future,] you'll see specific ways of buffering and pre-processing sensor data close to... rings of power that are different than rings of power we have now. We know processing sensor data for wearables will be one of those paths that we will optimize very heavily.
EET: What do you think about the use of scaled-down phone chips for wearables?
Chandhok: I think that just as we saw some sort of specialization in our own chip lines between the highest tier chip and the mass market chips -- we build them differently now -- you'll see us build more specialized hardware for wearables.
EET: Do you have any thoughts on how to make wearable data meaningful?
Chandhok: Devices shouldn't be sticky, the services should be sticky. For me, the first part of the data will be aggregation, the other one will be building inferences based on all that data and using machine learning. What I haven't seen yet is that inferencing and computation happen in more real-time and closer to the consumer. I wouldn't mind if these two devices talked to each other and got me more accurate information. The data right now is a very simplistic architecture, it shoves everything into a cloud and aggregates it and builds intelligence.
When wearables will get into mass market is when people will be able to take these building blocks and put them together the way they put together an outfit. Right now the ecosystems are too vertical, too silo-ed for the average consumer to have the patience to put up with it.
— Jessica Lipsky, Associate Editor, EE Times