Follow Datanami:
August 1, 2016

What Pokémon GO Means for Big Data


Pokémon GO has kids of all ages roaming the world in search of digital beasts to capture with their smartphones. While the pop-culture phenomenon provides a (mostly) innocent diversion for the summer, it also serves as a potent reminder about the power of big data and how we’ll interact with emerging technologies, like augmented reality, in the future.

We’ve never seen anything quite like Pokémon GO before. In its first week of availability, the mobile app became the most downloaded game ever, surpassing hits like Candy Crush and Clash Royale. There have been more than 75 million downloads of the game, according to SurveyMonkey, and its popularity is only growing.

But here’s the statistic that really puts the phenomenon into perspective: According to a survey, people are now spending twice as much time playing Pokémon GO as they are in Facebook (NASDAQ: FB). Traffic on other social media properties like Twitter (NYSE: TWTR), Snapchat, and YouTube is down, and it’s all because of Pokémon GO.

If you think there’s a big data angle to all this, you would be right. For starters, let’s consider the compute and storage requirements that the game’s owner, Google (NASDAQ: GOOG) spin-off Niantic Labs, has had to deal with.

Little is known about the technology behind Pokémon GO or where it runs; the company didn’t respond to Datanami’s questions before this story was published. But the fact that Pokémon GO is essentially a new version of a Google-developed game called Ingres, strongly suggests that Pokémon GO lives within the comfy data confines of the GooglePlex.

The fact that Pokémon GO has been able to scale so quickly with so few glitches shows that Niantic was mostly well-prepared for the insane growth that the game has experienced, even if Niantic wasn’t expecting the game to become the all-time most successful game ever (which it almost certainly didn’t). But that doesn’t change the fact that scaling a mobile app like this is really difficult.

Pokemon Go_3

Pokémon GO uses a variety of big data technology under the covers.

“The one thing that’s very clear from this it it’s still really hard to build cloud native apps that scale as fast as the phenomena has driven this usage,” says Varun Chhabra, product marketing lead in EMC‘s advanced software division. “Building a cloud-native app that really handles a large volume of users spread across different geographies, it’s not easy.”

Scaling at this level requires an elastic compute infrastructure, such as Google Compute Engine, which is where Pokémon GO is suspected of running. But it also requires agility at the architectural level, and that means developing apps using a large collection of micro-services that can scale independently, Chhabra says.

“It’s really unpreceded the way Pokémon GO has had to scale. Most enterprise don’t have to worry about that level of scale,” he tells Datanami. “But the lesson to be learned here is you really have to aggregate things as micro services to test them out, and really build some in-house capabilities around micros services and containers. Doing that before prime time is a good idea.”

Storing data for a game used by millions of people also poses a dilemma. Nobody knows for sure what Niantic is using on the backend; the company is not talking publicly about its technology. But again, that hasn’t stopped people from speculating, and drawing lessons from it.

“There’s been a lot of debate whether they’re using a relational database or not,” Chhabra says. “Most of what I’ve read seems to be pointing to using a Google Cloud Data Store NoSQL database. I think the chances are they’re using a NoSQL database.”

Geospatial and AR

One of the neat things about Pokémon GO is the stage is the real world. People must go find and catch digital creatures that the game-makers placed in parks, museums, beaches, shopping malls, and other public places. That means storing geospatial data about users’ location–and lots of it.

“I think geospatial data is really important,” says Neil Cawse, CEO of Geotab. “What people don’t realize…is they went and studied all the geospatial data. For example, they used geospatial data to figure out the locations of parks, police stations, areas for PokeStops and PokeGyms. They had to make them accessible and places where people would want to go. Without that geospatial data backing it up, Pokemon Go would not have been as successful. They would have had to do a lot of manual work, or they might have chosen the wrong spots.”

Pokemon Go_2

The use of geospatial data and augmented reality has made Pokémon GO a huge hit.

It’s been said that Pokémon GO is essentially a way to play a game within Google Maps. That’s a fair consideration, but it doesn’t take into account the other big technological leap that Pokémon GO has made here, which is augmented reality (AR).

“The thing that’s really interesting about Pokémon GO is..this notion of augmented reality, where you’re really interspersing the phone’s camera with the real world, and layering on digital information on top of it,” EMC’s Chhabra says.

While Facebook, Microsoft (NASDAQ: MSFT), and others plod toward virtual reality (VR) with geeky-looking headsets, Pokémon GO stormed past them to give millions of people their first taste of AR. Businesses will follow in the game’s footsteps to take AR to new levels, and that’s another point where AR and big data will mix.

“Instead of a Pokemon showing up in your camera, imagine if you go to a local grocery store and you’re standing in front of an aisle,” Chhabra says. “The app could say ‘You bought milk from here last time. That’s available here.’  Or it could even show you things that are on sale. That could really transform the retail customer’s experience.”

Geotab’s Cawse says Pokémon GO is just scratching the surface of what’s possible with geospatial data and augmented reality.

“We are just at the beginning point of AR and VR. People are inherently social. People love games and competing against one another,” he says. “Pokémon GO is transforming the gaming experience. In the past, technology has kept us in closed rooms, sitting down, playing tv games on the screen and not talking to anyone else. This is the exact opposite, get people outdoors, talking to real people, participating in the real world, and having fun at the same time.”

Related Items:

Game-Changer: The Big Data Behind Social Gaming

5 Ways Big Geospatial Data Is Driving Analytics In the Real World