Scaling storage for more than 2 million users searching for their roots.

image

Searching for the roots of your family tree? Some 2 million paying subscribers have uncovered their family history on Ancestry.com, the largest for profit genealogy company in the world. Since going online in 1996, the company has housed 70 million family trees with more than 6 billion profiles including over 300 million photographs, scanned documents and written stories.

Superuser interviewed the architecture team from Ancestry.com, the world’s largest online resource for family history with more than 2 million paying subscribers across all its sites. The architecture team is using OpenStack to provide Ancestry.com developers the ability to provision virtual machines, containers, object storage, software load balancers, and other software defined services.

Describe how are you using OpenStack. What kinds of applications or workloads are you currently running on OpenStack?

At our heart, we are a fast-paced, high-tech company guided by a very human mission: to help every person discover, preserve and share their family history. We work together across the globe to provide something truly priceless: authentic family stories.

A few years ago, Ancestry realized it needed to move away from expensive and proprietary solutions. These solutions didn’t provide the necessary APIs and features that are needed in today’s fast moving economy. To scale the storage that protects and delivers very important image content for millions of Ancestry users, we designed and deployed a large multi-petabyte OpenStack Swift cluster and have migrated most of Ancestry.com’s image data (billions of images) to OpenStack Swift and AWS S3. This infrastructure is growing at 25 percent annually.

Ancestry has deployed an OpenStack cluster that they using for a dev & stage virtual environment. They have provided many development groups with their own tenants. The developers are spinning up virtual machines, which allow them to quickly test code and prepare it for production.

What have been the biggest benefits to your organization as a result of using OpenStack?

To stay competitive, OpenStack enables us to lower costs by leveraging commodity-based hardware and allows us to be innovative. Ancestry has adopted a new open source strategy and leverage technologies like Linux, OpenStack, Docker, Kubernetes, and other popular open source solutions.

We also strive for flexibility and agility for cloud application provisioning through APIs. With OpenStack, all services (storage, networking, compute) have a common set of industry standard API’s that teams can leverage to provide services on-demand.

How has your team’s OpenStack deployment grown and developed?

After a successful Swift migration, Ancestry has been testing the rest of the OpenStack suite and plans to migrate the company’s virtualization environment to OpenStack. We are currently architecting an OpenStack cluster running OpenStack Juno on CentOS 7. The deployment originally started out as a few engineers scraping up hardware and testing OpenStack. It has grown tremendously and now almost all of Ancestry’s image data is stored on Swift and within the next couple of years OpenStack will run our virtualization and possibly our bare-metal environment.

OpenStack will provide an open source Infrastructure-as-a-Service (IaaS) solution built on commodity based hardware. OpenStack exposes APIs that developers or DevOps can leverage to provision virtual machines, containers, bare metal servers, block storage, object storage, software load balancers, and other software defined services.

What are some of your key contributions to the OpenStack community?

We had a presentation session in Vancouver and Tokyo where we shared our experiences. We are interested in Docker and how containers and microservices fit into our OpenStack architecture. Eventually we would like to make contributions back to the community, but we don’t have enough internal resources at this time.

What keeps you awake at night?

The same things than any infrastructure architect worries about: uptime, responsiveness, and avoiding surprises. Working with commercial vendors in the OpenStack Marketplace like SwiftStack, Midokura, and Mirantis gives us support assurance and access to their expertise to help us get the most out of our OpenStack endeavors.

You can find out more about OpenStack in production at Ancestry.com in this 40-minute session from the recent Austin Summit.

Cover Photo // CC BY NC

Allison Price