Kentik, which was formerly named CloudHelix, officially emerged from stealth mode on July 1, with $12.1 million in a Series A funding round led by August Capital. Total funding to date for Kentix stands at $15.2 million.
"Kentik" is a Yiddish term that means "visible," which is what Kentik's product is all about: making the network and potential risks visible, according to Avi Freedman, co-founder and CEO of Kentik. Among the use cases for Kentik's technology is distributed denial-of-service (DDoS) attack detection as well as optimization and efficiency for network peering analytics.
Freedman previously spent nearly a decade working at content delivery network vendor Akamai, where he held the title of chief network scientist. In his experience, understanding potential security risks and network problems is both a data management and visibility challenge.
"There are a lot of options for enterprises to take network and host data and try to correlate it to show visibility," Freedman told eWEEK.
In his estimation, the existing tools in the market for visibility tend to stop at high-level views of the traffic. What Kentik does is take the details from the network traffic flows and pull them into a software-as-a-service (SaaS) offering. The data is pulled into the Kentik engine, which is a big data engine that the company has developed. On top of the Kentik engine is a portal with API access that takes all the traffic data and gives organizations visibility into traffic.
The Kentik back-end data engine is a high-speed system that can store a lot of data and process it quickly. Freedman said that Kentik uses a multilayered approach to data, with recent data residing in a caching layer, to enable faster query access.
"For the persistent caching layer, we use a MySQL derivative and then use PostgreSQL as an API," Freedman said. "Inside the Kentik data engine, we have our own clustered column data store on the back end, and we use MySQL to remember the query answers."
PostgreSQL has something known as Foreign Data Wrappers (FDWs), which can make Kentik's entire data set look like a PostgreSQL database to the operators that use the Kentik system, according to Freedman.
From a deployment perspective, Freedman explained that an organization can export data directly to Kentik or use a virtual appliance that acts as an encrypting network proxy.
"We're not getting all the packets; we're getting summaries of where the packets are coming from and going to, called netflow," he said.
Looking forward, Kentik is going to be focused on building out new features based on customer requests, according to Freedman. He also expects additional performance and security analytics capabilities will land in Kentik's product offerings.
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.