Ian Ferguson, Vice President, Segment Marketing, ARM Holdings PLC
Recently, I was lucky enough to be invited to sit on a panel focused on the opportunity for the use of wearables by the poorest three people on this planet. As I sat on this panel, I reflected on the use cases that would enable mass-market deployment of wearables, coupled with the technical and business challenges that prevent those use cases from becoming reality. Eighteen months ago, our industry was barely discussing this space and now many commentators are dismissing it as a pure niche play, but I am highly excited about the prospects for this product category.
For my viewpoint to become justified, let us look at use cases. I work in the technology field and yet I do not wear anything on my wrist. Sure, I know many families whose lives have positively changed due to the inclusion of fitness products in their daily lives. The number of steps walked has become a competition in the household……that has to be good.
For me, I think there are three areas we need to be pushing on:
Access to Quantitative Data
In this area, my interest is data that will help prolong or improve the quality of peoples’ lives. Two billion people on this planet, including my mother who suffers from Alzheimer’s, suffer from conditions or have ailments where access to brain, heart and lung information could improve treatment options. Couple this with the fact that when used in conjunction with a smartphone a consumer will have access to data and a person’s location; this offers up a number of compelling options that I would personally pay for.
This is probably the main area of focus right now. The wearable has the opportunity to become part of your entertainment, healthcare and education experience.
Emotiv, a company that has developed a wearable on the head that monitors brain waves, has companies developing Android™ based applications including a meditation application. For many, that may be a more compelling prescription than “take two of these, twice a day”, with of course the benefit of monitoring the impact of the application on the patient. In a high-end game, I could take advantage of “thought” delivered new immersive experiences.
Fragmentation of Standards
At one level, one could argue that the connectivity for wearables is coalescing around Bluetooth Low Energy (LE). Given that wearables are assuming a linkage to a smartphone or personal computer, this is largely true. Sure there are other physical interfaces for connected devices, especially for machine-to-machine applications, that are being deployed. The fragmentation that we are focused on relates to protocols associated with authentication of devices, permissions for how data is shared and communication of the resources available in the device. An obvious question when talking about the exchange of data is how safe is it? Data is valuable. Access to it can provide compelling services for business and consumers. The “little data” that is gathered from a heart monitor system can give early warnings for heart attacks. Consumers would, without doubt, wish that data be shared with appropriate physicians. This data is, today, ‘tied’ to a service in a proprietary, vertical silo. There may be a broader set of companies that could deliver services if there is an understood set of communication protocols. That said, advertisers can harness information about where you live, what you like and your physical condition to sell targeted ads. As ARM’s CTO has noted in previous white papers, consumers are relatively relaxed about this at present. We do expect this to rise up peoples’ priorities when selecting new technology and this is a topic that needs to be solved holistically.
From the technology side, I will share my requirements in the reverse order of importance:
The smartphone has revolutionized how we interact with other users and businesses. I believe that a device that is even more connected to us can enhance this further. Will there be applications that can make sense of what we are thinking, what we are touching, what we are smelling?
Optimizing for wearables
As I mentioned above, this market has taken off relatively quickly. As a result the technology deployed in the initial wave of products is reused from adjacent markets like smartphones or the microcontroller market. Companies developing the next generation of solutions are looking at a number of ways to improve energy efficiency and the cost point of the solution. I believe that a battery will need to last at least a week (ideally a month) before needing to be recharged. The physical space that a wearable worn on the head takes up allows for a larger battery to be permitted in the desired form factor, and this is one of the main reasons why this type of device has migrated most quickly to running a richer OS on a processor that is operating at several hundreds of megahertz.
The power challenge is being addressed in several ways:
Shifting compute performance to the device connected to it or the cloud
Reducing the amount of processing that is done on the wearable will enable the processor element of the wearable to be kept in a lowpower state more of the time. This will be application dependent. For some applications, immediate access to some or all of the data from the wearable will be vital. In other cases, it will be acceptable to provide a sync to a location where the data is residing (the smartphone, the PC or the cloud) as and when the connection from the wearable is re-established. A good analogy of how this may evolve can be seen from mapping products. Several of the early products kept all of the map information in the cloud. This was fine until the (typically cellular) connection to the cloud was lost. The newer map products cache information locally, enabling the application to keep working during time periods when access to the main databases is limited or non-existent.
Moving to more aggressive process nodes
This is an interesting debate that many of ARM’s silicon partners are wrestling with right now. In other areas, a migration to a newer process node is a (relatively easy) tactic to improve device cost and power. For wearables, this path is more challenging for two reasons. First, these silicon designs include a significant proportion of analog technology. Analog intellectual property tends to not be broadly available for the most advanced process nodes. I would argue the design element becomes more challenging too since more advanced process nodes “like” smaller voltage swings….yet the analog world being deployed here for sensing the world around us will continue to harness (relatively) large voltage inputs. The second challenge is flash. In a similar way to analog IP, flash intellectual property tends to be available a couple of nodes behind the “bleeding” edge. As a result of this and other areas of IoT, I expect roadmaps of flash technology to be accelerated .
Increasing levels of integration
The initial wave of wearables in the market today typically incorporate multiple discrete chips. Consolidating system functions down to fewer components (ideally one chip) offers additional opportunities to improve system power and cost. Given the challenges noted about moving all of the design to an aggressive new process node, I believe this may be one of the markets that shift multichip modules from “talked about” technology to mass deployment. Semiconductor companies (or potentially OEMs) could select the appropriate digital processing, flash and sensor subsystems for their applications. There are clearly challenges with yield, testing and cost to work through, but given the dynamics of this market, this could be a credible product strategy that increases nimbleness to adapt to market changes.
This emerging area of the electronics industry can also learn from other segments. In the early days of set-top boxes, service providers soon realized that they were not able to predict specific use cases until systems were deployed. In those days, the processors were initially deployed about 1/3 – 1/2 utilized, with latent horsepower to support future services that were installed over the air interface.
To conclude, the wearables market presents huge opportunities to all in the electronics value chain. However, it will not necessarily be the first entrants in the market that will win in the long haul. Providing connectivity is not enough either. The winners will be those companies that understand real end use cases and have a plan for analyzing and sharing data in a controlled manner rather than looking to cost reduce the current set of products on the market.
Coupled with the knowledge that whatever you produce will not be the final answer, baking in a strategy for nimble course correction
will also be key.