The total amount of data created, captured, copied, and consumed globally is forecast to reach 463 exabytes by 2025, according to the World Economic Forum High Scalability report. An exabyte is 1,000 bytes to the sixth power. Facebook alone processes 2.5 billion pieces of content and more than 500 terabytes of data each day.
It’s not surprising that organizations like Facebook had to innovate new ways to manage massive data thruput requirements at ever increasing scale. As a result, a cadre of former Facebook software engineers have created their own innovations and new database software application companies. One such company is Yugabyte.
Founded in 2016 by several former Facebook software engineers, Yugabyte is the fast-growing, Sunnyvale, California-based company behind YugabyteDB, the open-source, high-performance distributed SQL database for building internet-scale apps.
The company positions YugabyteDB as serving both scale-out and internet-scale OLTP workloads with low query latency, extreme resilience against failures, and global data distribution. As a cloud-native database, it can be deployed across public and private clouds as well as in Kubernetes environments.
One of those Yugabyte founders is Karthik Ranganathan, Chief Technology Officer, who spent time at both Facebook and later at software-defined storage company Nutanix, as did his fellow co-founder Kannan Muthukkaruppan. Another co-founder Mikhail Bautin also spent time at Facebook. “Our learnings from Facebook building and running multiple databases and our learnings from Nutanix about enterprise companies and enterprise customers gave birth to Yugabyte,” says Ranganathan.
“Why the hell are we building another database? We realised, people really want relational databases in a cloud native fashion. That’s really the intersection point. And the world is shifting to what Facebook knew years ago, which is to build applications really quickly for end users,” says Ranganathan.
Ranganathan points to his experience building applications at Facebook. The mission was to build applications really quickly. But you had to ensure not only that it worked well, but that it wasn’t beaten by what Ranganathan refers to as “accidental success.” When you start building an application, you don’t know how well it’s going to succeed. Case in point, Ranganathan and his team were working on a new release of Messenger which was doing 150 million messages a day at the time. It turned out that his new application which combined data from chat plus messaging and SMS, generated 20 billion messages a day.
“How do you even estimate that, because the product changes the user behaviour, and the user behaviour informs the product. And so a few months of this, and you have a completely different usage. So we couldn’t be in a situation where we said, ‘Oh, look, we can’t scale because too many people are using it.’ And then it’s kind of a failure, because you succeeded. And that’s no good for anybody. We realised three things are critical with cloud native applications. It boils down to always being available-no matter what failures, upgrades, security patch, whatever, it has to be available. It’s about scaling, any which way up and down on demand, right. So you should be able to start small, grow back, come back, small, get bigger, again. And the third thing is about really moving data across different regions, locations, zones. So distribute data, build that replication of data, movement of data into the database,” says Ranganathan.
In creating Yugabyte the co-founders decided that we’re not going to reinvent the database API. They were going to simplify the problem of database application mission creep. That meant taking API’s that exist and making them cloud native. That made the concept easy to understand for developers already using applications like Postgres or NoSQL. “We could say, hey, it’s just like this database, but it has these other properties for cloud native application, and they immediately get it. So it’s like NoSQL, but it has indexes in this stuff. It’s like Postgres, but it can do scale. The next question is, how much of Postgres do you support? And that’s where we say, ‘Look, we’re out of the API game.’ We support everything. We reuse the Postgres codebase. So, all the features, everything, you get the same thing. We’re not changing anything here,” says Ranganathan.
Despite its current success, it took a few years for Yugabyte to find market fit. The co-founders bet on a cloud-native service at a time when cloud adoption was at its early adoption phase. And their focus on an open-source was questioned by early investors as a sustainable business model. So the company began as 80% closed and 20% fully open source. But it wasn’t until 2019 when they reversed course and went to a mostly open-source platform model, coupled with the now near ubiquitous adoption of the cloud, that the business took off.
“We simply open sourced the database as an experiment in around May of 2019. We said just make it 100% open source and let’s see what happens. And it increased our adoption tremendously. Like our community went nuts. So we didn’t turn back after that. So our model is quite simple. Our model is we are 100% open source like Postgres and we are a cloud managed service. Whether it’s in the customers cloud account or ours, it doesn’t matter, we use the same software that operationalizes Postgres and we think that the commercial money is really in that part and not in the open source database,” says Ranganathan.
Today, Yugabyte is growing fast with more than 300 employees and more than two million downloads of its software, enabling the company to attract $291 million in financing to date, including it latest $188 million Series C round in October 28, 2021 led by Sapphire Ventures, valuing the company at $1.3 billion. Other investors include Lightspeed Ventures, 8VC, Dell Technologies Capital, Alkeon Capital, Wells Fargo Strategic Capital, Meritech Capital Partners and Wipro Ventures.
Ranganathan was born in Chennai, India, to what he refers to as a “traditional middle class family.” He went on to graduate from The Indian Institute of Technology Madras in Chennai with a degree in computer technology. He came to the US for his graduate degree at the University of Texas in Austin. “I don’t know if it’s good or bad. But I started liking computers by playing computer games and was really hands on building all sorts of stuff,” says Ranganathan. After graduating, He began working as a software developer at Optimization Alternatives. That was followed by stints at Microsoft and Netsuite. Then in 2007 he began working at was then the one of the epicenters of database innovation at Facebook for a little over five years before moving on to Nutanix for two years prior to founding Yugabyte.
As for the future? “Our aim is to become a default for people building transactional, cloud-native applications. A cloud-native service needs a cloud-native database, and we would want to be that default. On the commercial side, we have ambitions to grow pretty big, but we want to grow in conjunction with the value we bring to commercial customers, which is to simplify the ability to run this on any cloud, across clouds, hybrid deployments edge and multi cloud deployments,” concludes Ranganathan.