Since the smartwatch boom, when Apple showed the world the first Apple Watch and was followed by so many other brands, devices have gotten smaller, started to collect more data and are more present in people’s lives. Little has changed since then, save with the addition of a couple of sensors, but another leap is starting with even smaller gadgets, seemingly unintelligent, but together doing a lot.
Their popularization at the moment is shared between Apple and Samsung with AirTags and SmartTags, respectively. Separating each one, they seem as smart as simple pedometers, but the difference between these gadgets lies in the work they can do by creating a kind of their own network, spread around the world, to alert the user to where a lost key is, or then the forgotten wallet.
It seems like a solution to a problem that doesn’t even exist, but behind it lies an important concept called distributed computing. Working with AirTags and SmartTags is just an edge of what already exists or can be created in a short time, in a few years.
Distributed computing is not new
If you’ve ever used any service that doesn’t run locally on your computer, you’ve already experienced something about distributed computing. This goes from Waze on your smartphone, Netflix from your TV, Spotify or personal assistant on a smart speaker, a torrent of life on your PC, and even a Google Stadia on a Chromecast. All of this takes advantage of the same concept: using a structure much larger than your hardware to process heavy information that it would have difficulty or would be unable to do on its own.
This network is made up of us, who do the work separately, but come together and exchange information with each other. These nodes became more and more competent and powerful in hardware, until the network’s growth started to occur in an exponential way.
“Today data centers are extremely distributed, where we have hundreds or thousands of machines distributing tasks. The catch now is the intelligence to distribute these tasks, so when a failure occurs in a node in the network, it can restructure itself to compensate”, says Ricardo Girnis Tombi, professor of postgraduate studies in computer engineering at Instituto Mauá de Tecnologia, in an interview with Techblog.
The idea of distributed computing then shifts from putting extremely powerful machines in one place, to scattering less powerful parts in greater numbers on this and other nodes.
Moving on to an even more recent moment, Tombi points out the arrival of computing in fog, or fog. “We come in with sensors, in this concept of the internet of things, which somehow have computational capacity. They have minimal processing and memory resources. They are small equipment, very distributed and they have important battery and energy requirements”, comments the professor.
It is in this scenario, with devices increasingly smaller and more focused on autonomy, that fit Apple’s AirTags and Samsung’s SmartTags. Looking up close, they just do the basics and that means pointing the Bluetooth direction to a nearby device. It would be more or less the same thing a wireless speaker already does, but it’s the network created from that information that matters.
Both the Apple model, as the Samsung, talk to other devices nearby to then be able to find the internet somewhere along the way. From this moment on, the conversation moves to the most powerful part of distributed computing, which are the servers in the nodes. They are responsible for notifying the user about the location of the key, or any other product hanging on the tag.
Where do these sensors go?
The tags example is the most tangible of all and easy to identify by the user right away. There are others already being tested, such as Amazon with the microphones of its smart speakers, the Echo.
“Amazon is able, by the sound pattern of the environment, to estimate the level of income you have. If the sound echoes in a simpler material, it becomes more reflective. If it hits a super expensive wood, it comes back more muffled and a sound like that tends to be from a more expensive house,” says Arthur Igreja, co-founder of the AAA innovation platform. “You don’t even need such advanced sensors to have a 3D mapping of the environment, the question is really more software”, he adds.
This scenario confirms the greatest importance for smart devices today: good software to control all the information collected with the hardware already available to the consumer. Not only present at this time, but also at low cost.
Also circulating by tags with accelerometers, Arthur also points to a use in insurance, which already happens in the United States. “You put a device of theirs in the car, and then it monitors speed and other data to see how you drive. In a car rental company, it is possible to locate a vehicle quickly and even without going through the company counter”, says Arthur.
The popularization of these tags, mainly due to their low cost, can also foster better internal location with apps like Google Maps. With the help of more powerful antennas in fixed locations, a market can indicate to the customer where the meat and pasta are, or a museum makes the route much more precise to a specific work, even indicating which floor it is on.
Sensors the size of a speck of dust
Thinking about a more futuristic environment, probably far from the current capacity of small gadgets hanging on the user’s keychain, Ricardo Girnis Tombi believes that their miniaturization can lead to an even more specific use. “Imagine a sensor, the size of a speck of dust. Thousands of these sensors, in something like a burst, collecting all the information about an environment and taking this data to big data”, says the professor.
Both this side to the future, as the tags on gadgets we have today and the functions already possible with them at various levels in the market, end up hitting two crucial points recognized by both Arthur Igreja and Ricardo Girnis Tombi: battery life and simplicity of functions.
Both commented to Techblog that the current objective of these gadgets is to keep the energy consumption very low so that they do the work longer and, at this moment with current battery technologies, the only way to do this is to specialize each sensor in one or two tasks, such as measuring speed, or user location, or number of steps.
“The form of communication, the protocols and algorithms it will work with, together with the efficiency of the processor and memory. This entire pipeline needs to be very fair so as not to waste machine cycles that it shouldn’t, as this is all energy consumption. In the end they all need to work fine tuned”, says the teacher.
“If we manage to make a good distribution of tasks and an orchestrator gathers all these distributed tasks, we can achieve this goal. Each device with its specific function”, adds Tombi.
Even with 5G promising wonders for small internet of things gadgets, Arthur Igreja believes that the time is not yet for the use of this network in these tags or even smaller sensors. The villain is energy consumption, caused precisely by current communication protocols.