The Next Chapter for Falarica – Building the Default Database for the Cloud


Author:

Originally Sourced from: https://blog.yugabyte.com/the-next-chapter-for-falarica-building-the-default-database-for-the-cloud/

Hi there…a big hello from all of us at Falarica!!!

It is with immense pleasure and great pride that we, Falarica.io, decided to join hands with Yugabyte to build great data products together. Thanks to the entire Yugabyte team for extending a very warm welcome to us. And of course a very big thank you to the entire Falarica team for their hard work in creating considerable value in a short span of time. Cheers Team!!! We did it and we did it fast :-).

Falarica and Falaricans

Falarica came into existence in May 2020. The founding team of Hemant Bhanawat, Suranjan Kumar and I, Neeraj Kumar, had the vision of creating a powerful multi-cloud, fully managed, SQL analytics platform. Something like the Amazon Athena service that can run on all your clouds, including private on-prem data centres. Our flagship product, Falarica Query Platform thus came into existence. We also built a Presto Kubernetes Operator and open sourced it, which quickly started getting a lot of community traction.

Before we started Falarica, Hemant, Suranjan, and I worked together for more than a decade. We’re passionate about creating distributed data products and have a strong background in distributed databases. Some of our work is well-known and some not so much, but building each of these systems from scratch has been a very exciting and enriching experience. We played pivotal roles in creating Gemfire (Apache Geode now) – a distributed key-value data management system, GemfireXD – a memory optimized, distributed sql database for OLTP workloads, and SnappyData – a unified data platform for analytical, transactional, and streaming workloads.

Coming together

While we were working with one of our initial customers, an enterprise that provides card management and transaction processing systems, we got an opportunity to evaluate a few transactional databases. This enterprise wanted to eventually push this transactional data into Kafka in order to run analytics on it. The customer was working on choosing the right database for their transactional system needs, the key requirements being 100% uptime, extreme resilience to failures, and a true global system with close to real time cross site replication, along with the need for extreme low latency transactions and high throughput. The popular but more traditional databases like Postgres, MySQL, SQLServer, etc. were not considered by the customer primarily because they lacked first class out of the box multi-site configurations and cross site mirroring capabilities. And these were non-negotiable requirements.

YugabyteDB, CockroachDB, and MongoDB were some of the options shortlisted for evaluation. While all the three products had their pluses and minuses, Yugabyte captured our attention. The out of the box cross site replication capabilities in particular, like, async, sync as well as read replicas, really impressed us for three reasons. Firstly multi-cloud, multi-zone, multi-region, multi-site, hybrid, you name it are all covered seamlessly. Secondly, there are multiple options to let the customers choose a suitable deployment topology as per their business needs. Thirdly, Yugabyte’s decision to adopt PostgreSQL as the query layer is very smart. PostgreSQL is arguably one of the most popular and widely used relational database systems in the whole world. YugabyteDB is PostgreSQL compatible and supports advanced functionalities such as triggers, stored procedures, partial functions, extensions, etc.This compatibility will encourage a lot of existing deployments to adopt YugabyteDB and enjoy the modern features which it brings along without needing changes at application level.

Having gone through a detailed evaluation of multiple databases, we felt that Falarica’s vision of a multi-cloud solution for analytics was a natural complement to the multi-site transactional story of YugabyteDB. With similar thoughts from both sides of providing a comprehensive end to end solution to the end users, the two sides decided to meet to explore partnership ideas which soon turned into the discussion of working together as it made much more sense to closely work towards a greater goal.

Why we think Yugabyte is doing the right thing – our perspective

It was important for us to diligently evaluate the promises of Yugabyte’s flagship offering. As builders of multiple distributed data systems, we pride ourselves in our ability to critically analyze data products. In addition to our experience of developing data products from a clean slate, we also have experience dealing with real production situations and interacting with customers for their requirements along with substantial hands-on experience with different clouds, technologies like Kubernetes, Spark, Presto, and Kafka. We set out to evaluate YugabyteDB from the perspective of requirements for modern applications – in the current context of pervasive cloud adoption, cloud native approach over cloud specific approach, security and data compliance needs, etc. These requirements, if not taken care of, can quickly render otherwise great products obsolete and unusable.

YugabyteDB checked all the important boxes for us. Sure, there is opportunity to make it even better like any other product, but the core architecture and the key features already support a variety of complex and demanding workloads.

Let’s look at some key ones:

  • Cloud agnostic and cloud native – YugabyteDB has embraced cloud native approach making it suitable for any cloud environment rather than any specific one. It provides seamless deployments on Kubernetes clusters using helm charts as well as its K8s Operator.
  • PostgreSQL and Cassandra Query Language (CQL) compatibility – No new APIs or language to be learnt. PostgreSQL is very rich and widely adopted, and porting applications to YugabyteDB should be zero to very minimal work at all. Remember, like the customer described above, there are numerous others who have the same needs and sooner rather than later the OLTP system for all businesses will follow suit. Similarly Cassandra is one of the most popular and widely adopted NoSQL databases. Migrating from Cassandra to YugabyteDB is straightforward and the applications around it can seamlessly start working with YugabyteDB with a switch of the connection URL.
  • Multi and hybrid cloud deployments – It’s common knowledge now that enterprises use more than one cloud. Either because of financial reasons or because of merger and acquisitions of similar businesses who were on some other cloud or a variety of other reasons. One of the core pillars of YugabyteDB is to give organizations the freedom to run their data infrastructure anywhere, avoiding being locked-in to any specific cloud provider.
  • Support for real time analytics without disturbing the OLTP workloads – Features like read replicas and active passive cluster replication can easily support analytics close to real time on transactional data.
  • Near real time cross site data replication – This is a must for not only ensuring safety against disasters but also for keeping your transactional system running 24x7x365.

Glad to join the Yugabyte family

While technology created the possibility, the similarities in working culture, the common belief in open source, customer-first ethos, and like mindedness sealed it off. We are glad to have joined the Yugabyte family and are already seeing successes in just a few months after the acquisition. Remarkable community growth, new customer wins, new features regularly being added to Yugabyte Cloud, the company’s DBaaS offering, has already made us feel great about joining the Yugabyte journey.

We have no words to say how nicely the entire executive team went about the acquisition process. Meeting and knowing Kannan and Karthik was super awesome. Their people centric value system, transparency, and openness in discussing everything gave us the idea about the healthy yet competitive environment at Yugabyte. Having multiple meetings with Kannan, Karthik, and Sridhar to understand and discuss how we would work together towards a common vision was very exciting. It was equally exciting talking to Bill about the business side of Yugabyte and the growth path ahead.

Thanks to all our friends, families, ex-colleagues, specially Sudhir Menon and Jags Ramnarayan, and all our well wishers for their support and encouragement in our short but important journey.

Signing off as Falarica Analytics!!! It was a short but a very memorable and sweet ride. Signing on as Yugabeings at Yugabyte!!!

Cheers!!!

Team Falarica at Yugabyte