You are probably scratching your head thinking that the title of this post is a random collection of words. I assure you nothing can be further from the truth; read on and the connection will become apparent. Let’s start with two biggest news in consumer electronics of 2012 i.e. introduction of the iPhone 5 and iOS 6. One of the touted features of the iOS 6 is replacement of the Google Maps with the brand spanking new Maps app from Apple. Apple Maps snafoo Apple Maps turned out to be a major disappointment garnering lots of criticism and prompting Tim Cook, Apple CEO, to publicly apologize to his customers.

In this apology Cook suggests that  iOS 6 users try alternative mapping applications like Bing, MapQuest and Waze. I say forget Bing and Mapquest. Go and get Waze. A couple of days ago I had a lunch appointment with a friend of mine. I counted on the drive to the restaurant taking 25-30 minutes. It took me an hour as I was trying to avoid one traffic jam after another. I have one of those GPS units in my car that receives traffic reports and displays delay information along the route. My radio is tuned to 680 News which does traffic reports every few minutes. I should have been able to avoid any an all traffic congestion without braking a sweat. When I showed up 30 minutes late and explained the reason for me leaving him to play with his iPhone for 30 minutes he said “why don’t you get Waze?”.

I downloaded Waze from the Google Play Store (I have an Android phone but Waze is available for iPhone, Windows phone and there is even a beta for Blackberry) right then and there. I’ve been using it for the past few days and this thing is more addictive than crack cocaine. OK, I have to admit that some of my enthusiasm is rooted in the novelty … but really is very food. The idea behind Waze is as simple as it is ambitious. Waze crowdsources mapping and traffic content … it is like youtube for traffic. It collects real-time traffic information from everyone who has Waze turned on and uses it to build insanely accurate picture of your commute experience. It displays speed estimates all along the way based on the experience of others. I’ve compared the information Waze is displaying against my actual speed and it was spot on with variations of only a couple of kilometers per hour. And it is as accurate in predicting the location where delay will start and where it will end. Brilliant! There are lots of other cool features. You can tell other wazers about accidents, police, road construction etc. through a single button click (no typing while driving). You can even look for cheap gas, if there is such a thing. It is fully social network enabled so you can tweet and Facebook your trips and it is fully gamified so you can collect little prizes and earn badges a you use Waze. Oh, and I almost forgot, it actually does turn-by-turn navigation.

Click on the picture to watch the demo

Waze is good at what it does because it collects and processes telematics data from everyone who uses the application. Waze has over 20 million users. Now imagine how much data Waze needs to collect and process and then distribute out to make our driving experience just a bit less painful. When we talk about big data, the case of sensors generating large volumes of data invariably enters the conversation. Waze relies on about 20 million sensors, our mobile phones with GPS constantly transmitting data for processing to its centralized servers. Yes, it is big data, very big data. And as you would expect from a big data company, Waze uses Hadoop to crunch the data on a large cluster of commodity servers. No wait. Waze a startup from Israel where it seems they have 3 startups for every man, woman and child. No self-respecting startup invests in hardware, not even commodity hardware. Waze infrastructure is all on the cloud more specifically, it is on Amazon Web Services (AWS). In the late 90s, the default stack for a startup was Java application and an Oracle database on SUN hardware. After the internet bubble burst, the entire industry went with open source and LAMP stack i.e. Linux, Apache web server running PHP, Python and a MySQL database. Today, it is all about Hadoop and NoSQL on cloud … anything else almost feels out of place.

Tagged with →