Today, Boundary announced the long awaited general availability of its application monitoring solution for big data architectures. Boundary uses network data to determine the health of applications and is delivered as a software-as-a-service.
Boundary’s service has been in beta since last year, during which time the company received over 1,000 requests to join. It’s currently in use by companies such as the real-time messaging service Spire.io and the game platform Minefold – companies where the slightest changes in performance can have a serious impact.
Boundary CEO Gary Read explains that while tools like Netuitive work by using sampled data to try to predict problems, and Splunk gives you the ability to mine all of your logs to retroactively find the causes of problems, Boundary analyzes the entire data stream as it comes in, allowing it to provide better predictions in a fast changing environment. Read says that sampled data might be fine in a slow-changing environment, but in a DevOps environment, especially one managing large clusters of analytics servers, tools that can’t do real-time analysis of the entire data stream just aren’t good enough.
Read joined the company in January 2011. Previously, he was the founder of Nimsoft, a monitoring vendor that was acquired in 2010 by CA for $350 million. “I was looking for a company that could change the game in monitoring,” Read told me. “When I found Boundary I realized I’d found exactly what I was looking for.”
CTO Cliff Moon founded Boundary in 2010. Moon has a background in developing NoSQL databases and built an early DynamoDB clone. Moon was frustrated with the monitoring tools that existed, so he formed a team and started building his own. Boundary uses a custom built complex event processing engine based on Erlang, Scala, Java and Riak.