August 26th, 2013

By 

Adam Holmburg: Software Engineer at GoDaddy

Matt Pfeil: Co-Founder at DataStax

 

TL;DR: GoDaddy was initially attracted to Cassandra because of its built-in replication features and scalability.  GoDaddy uses Cassandra as their transaction logging data store

(http://inside.godaddy.com/cassandra-daddy/) and their latest endeavor is to deploy Cassandra as a shared enterprise resource, utilizing mutli-datacenter.

 

The existing web session is 32 nodes across two data centers; GoDaddy’s enterprise deployment will be fairly larger.  GoDaddy plans to use both spinning drives and SSDs when appropriate, depending on performance needs.

 

Adam and GoDaddy host regular meetups for the Phoenix Cassandra Users group. If you’re living in the Phoenix area and interested in learning more about Cassandra, we encourage you to join.

 

Hi. This is Matt Pfeil from DataStax. I’m here today with Adam Holmberg, a Software Engineer at GoDaddy. Adam thanks for taking some time today.

Glad to be with you.

 

Why don’t you start off by telling the audience who have not seen the Super Bowl commercials, what does GoDaddy do?

GoDaddy is the world’s largest platform for small businesses. We’re also the world’s largest domain name registrar and web host provider. We provide a number of services in support of our broader mission to help small businesses succeed.

 

What was the use case for using Cassandra?

We’ve had it in use for a number of years. It was initially deployed in the web session store. We were attracted to it because of its built-in replication features and its scalability. Since then we have deployed a couple other applications on it. One is our transaction logging data store. I’ve actually blogged about those in my blog at http://inside.godaddy.com/cassandra-daddy/.

 

Our latest endeavor is actually to deploy Cassandra as a shared enterprise resource. What we’d like to do is support a multi datacenter use case. The idea there is to lower the barrier of adoption and allow more of our internal applications to take advantage of the high availability and the need of replication for our international push.     

 

I think that’s an awesome use case. In fact when I was at Rackspace with Jonathan, I was doing something similar where we’re building out shared infrastructure around Cassandra. It’s really nice whenever you can segment problems like scaling and availability from for actual application features, so you don’t have to try to prioritize to do or just not related at all. 

 

Can you maybe tell us a little bit more about your initial motivation for looking at Cassandra? What were the problems that made you start going down that path including anything specific around the technology?

The reason they approached it for web session stores was at the time we had sticky sessions and we were looking at taking advantage of a CDN which required non sticky sessions on our end. We looked at a number of data stores and obviously you want something that was able to handle high throughput. Of course, like I already mentioned the high-availability, the built in replication, resiliency to failure, all those kind of things played into the decision.  

 

Can you share any information about what the deployment looks like? Are you using a cloud data center or using your own or anything about how big the cluster is or any interesting stats like that?

The existing web session was actually 32 nodes across two data centers. Our enterprise one is going to be quite a bit larger, it will expand actually into our international data centers. That one’s not actually stood up yet.

 

 Awesome. Are you guys using spinning disk or SSDs?

The session store was originally deployed on spinning disk. Our new infrastructure is being built with actually combination, because in this case we’re going to be supporting multiple applications. We’re going to allocate the higher performance hardware when it’s appropriate and use spinning disk where SSD performance is not required.

 

You’ve been using Cassandra for a while, what’s your experience been with the Cassandra community?

The community in general has been great. I can speak to the local community in the Phoenix area. I know we’ve a few firms in the Phoenix area that have been using Cassandra in various capacities. In fact, GoDaddy hosted the last meet up two weeks ago in the Phoenix Cassandra Users group and had over 30 people show up. Those 30 people included current users and also people wanting to learn about it.  

 

The community is very responsive on the mailing list and IRC, and there are loads of blogs out there with information.   

 

Thank you for all the help that you’ve done with organizing the Phoenix Cassandra meet up group. Adam thank you so much for your time today and I’ll let you go unless there’s anything else you’d like to add at this point.

Just a big shout out to DataStax for providing the awesome Cassandra documentation. That’s coming from a guy who started with Cassandra back in the incubator and that the documentation was a lot more difficult to digest, which might be the understatement of the year. I remember the first Cassandra Summit, we had mentioned that doc coming whenever we just started the company and I do believe there was a very loud round of applause that resulted from that statement. We’re glad to help the community however we can.