Cassandra Meetup at MyHeritage

Agenda

5:30 PM Gathering and refreshments

6:00 PM Ran Peled - Harnessing Cassandra Power at MyHeritage

7:00 PM Daniel Chen - Planning a cluster deployment

Harnessing Cassandra Power at MyHeritage

Over the last years, MyHeritage has experienced rapid growth, and is now home to 82 million users who have added 2.2 billion rich personal profiles to their family trees. Stored on complex and highly optimized MySQL infrastructure, we started encountering scalability and performance issues in several specific use cases.

We have set up to solve these use cases using a NoSQL solution, and selected Cassandra for the mission. Following successful deployment, we now use Cassandra for many of our new projects.

In this meetup we will share the architecture and design considerations, as well as the lessons learned from a few of these projects. Among the complexities we will cover is the challenge to load 2.5 billion profiles using Hadoop while the system is up and profiles are changing, keeping the legacy MySQL and Cassandra in synch, handling errors, and BI integration.

Ran Peled is Chief Architect at MyHeritage, with over 30 years of software development experience, and expertise in large-scale Internet applications.

Planning a cluster deployment

Many questions arise when planning a new C* cluster deployment.

It starts with choosing the correct hardware/cloud platform and ends up with the fine tuning that is required to keep the cluster running smoothly and efficiently.

In this lecture we will talk about those questions and answer them all.