Real-Time Surveys Get Personal with PollfishJune 4, 2015
This post is one in a series of quick-hit interviews with companies using Apache Cassandra and/or DataStax Enterprise (DSE) for key parts of their business. For this interview, we talked with Stavros Kontopoulos who is a data engineer at Pollfish.
“Our technical team piloted various databases, along with Hadoop, and found Cassandra to be the best for our use case, which involves lots of time series data. While I’ve used MongoDB in the past, the ring technology of Cassandra makes sense for what we do and is a better fit for the scalability needs we have.”
DataStax: Thanks for chatting with us today. Tell us what Pollfish is all about.
Stavros: We provide a survey platform that allows companies to run real-time surveys. We use a mobile SDK for both Android and iOS mobile devices that allows surveys to be taken by the users of those devices. Both the developers creating the surveys and users taking the surveys receive something for their efforts – money or gifts for their time. We know enough about our users to do targeted surveys when the need arises.
We collect all the data from those surveys along with the user metadata such as their location and more, and provide insights into the details back to our customers and marketing companies.
DataStax: Where does your solution run and how is it architected?
Stavros: Right now we are in the cloud on Microsoft Azure. We use an application layer that interacts with all the mobile devices that have our SDK installed.
DataStax: What drove you to use a NoSQL database like Cassandra?
Stavros: We didn’t start out on Cassandra, but instead used relational technology in the beginning. But we quickly learned that wouldn’t let us process the amount of data coming in or do the analysis and get the value that we needed.
Our technical team piloted various databases, along with Hadoop, and found Cassandra to be the best for our use case, which involves lots of time series data. While I’ve used MongoDB in the past, the ring technology of Cassandra makes sense for what we do and is a better fit for the scalability needs we have.
DataStax: How about where DataStax Enterprise (DSE) is concerned? Why did you choose it vs. just using open source Cassandra?
Stavros: First, we are a professional enterprise so we need production-quality software with support. Plus, using DataStax Enterprise made things easier for us and provided us with greater reliability.
In addition, we use the enterprise features like search and analytics. For example, we use DSE Search for rollup-styled queries. We’re also building a new product that will allow our users to directly access our DSE database through DSE Search/Solr.
In addition, we use Spark analytics in DSE to process all our time series data and deliver the kinds of analytical insights our customers need.
We’re looking at security next, and will likely be using the LDAP support in DSE soon.
DataStax: How do you manage your cluster?
Stavros: We use OpsCenter, which is very easy to use and helps us take care of everything pretty well.
DataStax: How would you summarize the benefits you’ve realized with DataStax Enterprise?
Stavros: The simplicity of DataStax Enterprise has made thing much easier for us where development is concerned. When we compare DSE to the nightmare experience we had with Hadoop, it’s day and night. We found using a Hadoop installation like Cloudera harder.
Having all the key technologies and tools we need in one simplified platform like DSE speeds our development and helps us bring our application to market very quickly. It’s our platform for the future.
DataStax: Great to hear, thanks for your time.
Stavros: My pleasure.
SHARE THIS PAGE