We founded Kentik to make life easier for the networks and application operators that run the modern web. Our first service, Kentik Detect, is an infrastructure data analytics service that is scalable, powerful, flexible, open, and easy to use. Our customers — including Yelp, Box, OpenDNS, and Neustar — are using Kentik Detect to detect, alert, and dig into the efficiency, availability, and security of their infrastructure and applications.

Here’s a bit of background on how we got to where we are today:

Managing networks well requires visibility into traffic flows, network peering and interconnection, fast and accurate attack detection, performance analytics, and security forensics. But the tools available to operators have been… challenged — for decades. A common reaction we’ve all heard to the tools people have been using was some version of “they’re pretty much horrible but better than nothing.”

While it’s been almost 20 years since Cisco created NetFlow, the options for even the most basic analytics remain grim. Available choices for “solutions” consist largely of enterprise software or appliances, single-machine open source software, or more recently, work done by in-house tools groups trying to build platforms on top of existing big data engines like Hadoop or Elastic.

I had been a frustrated network operator myself, and had heard grumbling for years about the available tools, but it wasn’t until a few years ago that friends like Dan Ellis (now Kentik’s CTO) hammered home to me that there really wasn’t any modern, scalable, and affordable option for infrastructure analytics. The problem wasn’t getting the data — most switches and routers can generate decent traffic flow via sFlow, NetFlow, or IPFIX.  The problem is what to do with all that data. Receiving, storing, and processing it flexibly, in real time, and with granular retention is a huge challenge at today’s network scale and complexity. The gap was so big that most large-scale companies we talked to were trying to build their own solutions, or had first versions, with long roadmaps of basic functionality yet to be completed.

As we fired up a first version of Kentik’s platform and asked for victims to help us test it, we also discovered that most of our potential users (and now customers) had a strong preference for a cloud-hosted service. They wanted to be able to send data without having to stand up servers and install software, and then easily and quickly access it via a client, a portal, or API, and integrate into their existing tools and operational and alerting work flows.

Specifically, operators said that they’d like to see:

  • High speed ingest of high resolution NetFlow, combined with routing data.
  • Instant availability of that data for DDoS and anomaly detection.
  • Long-term retention and availability at high resolution.
  • A scalable architecture with open access to the data and analytics.
  • A roadmap to take richer sets of flow-like data from load balancers, server agents, and sensors to support deeper performance and security use cases.
  • Usability for all of the different operational groups: network, server, application performance, and security.

After taking the feedback and working with some first users, Ian Applegate, Ian Pye, and I decided to start Kentik (then CloudHelix) in 2014 to address these requirements. Working with data sent by leading ISPs and web operators, we built an operating data platform, and first applications for analytics and alerting. We were soon joined by a fourth co-founder, Justin Biegel, who took over our customer outreach. We raised our $3.1m seed round in September of 2014, grew from four to now 21 people, and started growing a paid customer base in February of this year.

Today I’m thrilled to announce the two major pieces of Kentik news mentioned above:

  • First, the launch and general availability of our core product: Kentik Detect. We’re excited that our initial paying Kentik Detect customers include companies such as Yelp, Box, OpenDNS, and Neustar as well as other service providers and web properties. If you’ve been frustrated with your network visibility tools, please ping us — we’d love to understand your requirements and set up a Kentik Detect trial!
  • Second, our $12.1m Series A funding. Led by Vivek Mehra of August Capital, this round represents a more than doubling of funding commitments from our seed investors — including First Round Capital, DCVC, WIN Funding, Tahoma Ventures, Central Electric, Engineering Capital — and our industry friends who helped support us as investors.

As part of the funding round, Vivek, who brings deep technical and business expertise as CTO and co-founder of Cobalt Networks and as General Partner at August Capital, has joined our board of directors. We’re thrilled to welcome him and August Capital to the Kentik team! August is a firm with extensive experience and strong commitment to enterprise and infrastructure companies, including some of the companies we most admire (hi, Artur!).

We’ll be using our new funding primarily for product development and core company growth. We have some very exciting products planned for the next few quarters that we’re already working on with customers. Again, please talk to us if you’d like to be part of Kentik’s road map conversation.

Check out the Kentik Launch news release to learn more about our product launch and funding, and hear first-hand from some of our investors and customers!

Also — and no surprise — we’re hiring. Please ping me if you’re a data dreamer with strong hands-on back end or front end experience, or you’re at a company building “classic” solutions and have been thinking about how great it’d be to work at a fast-paced company with smart customers and a deep road map.

Lastly, please check back here for future posts about how network operations are evolving, and what Kentik is doing to meet the challenge! The Kentik team will be blogging about our functionality and road map, and will also be presenting tutorials about infrastructure and operational topics.