jKool Empowers Customers with Real-Time Data Visualization with DataStax EnterpriseApril 9, 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 Charles Rich who is VP of Product Management at jKool, LLC (a spin off from Nastel Technologies).
“We learned that using NoSQL is actually a lot easier than you might think and you can get from inception to production quickly. You can get to market quickly and you can scale very easily.”
DataStax: Charles, thanks for taking the time to chat with us today. Please give us a quick overview of jKool.
Charles: jKool is a SaaS solution for real time visualization of streaming data. It could be any type of data and application, either data-at-rest or data-in-motion although we typically service applications that use analytics software like Apache Spark, IBM InfoSphere, Apache Storm, etc., that act on time series data. Our real time visualizations provide help to everyone from developers creating applications to end users that utilize those same applications.
Developers like us because we provide them with a real-time view of their data in-motion across their compute frameworks and at the same time offer a ready-to-use real-time visualization they can deploy for their end-users. We make powerful visualizations including: tables, charts, candlesticks, topology, geo-fencing and multi-panel charts all driven by our English-like query language, jKQL (jKool Query Language). This enables users to immediately see the changes in events and activities in their time-series data.
Currently we run all of this in the cloud on SoftLayer as a SaaS offering.
DataStax: How do customers use your solution?
Charles: We have two basic use cases. One is where a customer is doing real-time analysis and viewing immediate data such as streaming sensor data and the like. The second use case is where they want to go back in time and analyze/run the “post-game show” if you will. The latter obviously requires data being persisted whereas the former may have a short expiration date.
DataStax: How did you settle on using Cassandra and DataStax Enterprise?
Charles: We’re built on DataStax Enterprise; it’s the database that powers our entire application and we don’t use any other database. We looked at other options like MongoDB, Hadoop, graph databases and others when we got started. Cassandra made sense for us because a lot of what we do deals with time series data and Cassandra handles it very well.
We went with DataStax Enterprise over open source Cassandra because we needed strong search capabilities and we get that with DSE Search that uses Solr. We also needed the support that DataStax provides for the platform.
DataStax: It sounds like you never considered a relational database – is that correct?
Charles: Correct. Our parent company, Nastel’s customers, use RDBMS’s a lot and we saw the customers there struggle greatly with RDBMS bottlenecks, which prevented them from getting the high transaction rates they wanted.
Some of Nastel’s clients were using Oracle Exadata machines – the really big guns of Oracle for supposedly the highest possible performance – and they still hit the wall with all that hardware and Oracle consultants helping. We looked at that and said to ourselves, “this is not the future; this is the past”. Watching others fail was a lesson to us.
So when we began jKool, we started out knowing we’d need to use NoSQL as our foundation. We’ve been doing scalable, in-memory analytics for over twenty years so we understand scaling with big data and know what will and won’t work.
DataStax: What part does DSE Search play in your SaaS solution?
Charles: Our customers need to ask questions and make queries of their data like, “show me all the steps a product is going through in our supply chain and where is it now and compare that to normal”, and such things require a good search engine that can search through lots of structured and unstructured data as fast as possible. So as new data is consumed, we analyze and make that data available for search in DSE. It allows our external customers to easily talk to their data.
DataStax: How do you manage and monitor everything?
Charles: We use a combination of our own tools including Nastel’s AutoPilot product and such with OpsCenter. Plus we’ve worked with the team at DataStax on tuning and optimizing our platform and they have been invaluable in guiding us in performance tuning.
DataStax: What were some of the key things you learned and benefits you realized working with DSE?
Charles: We learned that using NoSQL is actually a lot easier than you might think and you can get from inception to production quickly. You can get to market quickly and you can scale very easily.
We certainly give kudos to DataStax for the product and their solid support for helping us get where we are today.
DataStax: Charles, thanks for the time today.
Charles: My pleasure.
jKool is part of IBM Marketplace and runs in IBM’s cloud. Visit them here.
SHARE THIS PAGE