NetFlow, Cloud, and Data: Pacing Network Visibility With Growth

In today’s hyperconnected world, the network sees all. Network operators can capture a tremendous amount of useful data from their infrastructure. The tricky part is extracting useful and actionable information from all that data.

Everything about your business traverses the network

Network data is a source of truth for everything in the modern business. Network data not only provides insight into the health and performance of IT infrastructure; it can also provide business intelligence. Enterprise Management Associates (EMA) research has found that enterprises are applying advanced analytics to network data for a variety of reasons, beginning with improved infrastructure capacity planning and technical performance monitoring.

Flow records (NetFlow, sFlow, IPFIX, etc.), for instance, can provide network operators with significant insight into their infrastructure, especially when they apply advanced analytics.

Enterprises are crunching flow records with big data

EMA recently surveyed 156 enterprises that are exporting IT infrastructure monitoring data into big data environments for advanced analysis. Forty-eight percent (48%) of these organizations considered flow records essential when applying advanced analytics to technical performance monitoring. Only application performance data was more important to a larger number (63%) of enterprises. What kind of monitoring are these enterprises doing? The most popular use cases were “network availability and performance monitoring” (53%), “systems availability and performance monitoring” (52%), and “storage availability and monitoring” (49%). Many enterprises (45%) are also applying these analytics to infrastructure optimization.

For all the value that this analysis can bring to a network operator, the sheer volume of data that is collected from the network can be burdensome. EMA research of enterprises exporting infrastructure monitoring data to big data environments has found that by 2017, 69% expect to be storing more than 50 terabytes of such data per day, with some even expecting petabyte volumes.

Many network operators will elect to store this data in the cloud. In fact, EMA research found that 47% of cloud managers at these enterprises have seen their infrastructure planning and design practices impacted by the presence of big data projects, and 43% have seen cloud operations impacted. This indicates that many enterprises are turning to cloud-based big data storage and analytics to ease the burden on their internal infrastructure.

Kentik applies big data and the power of the cloud to network analytics

Kentik provides SaaS-based network analytics solutions. Taking a big-data approach to advanced network analytics, Kentik leverages cloud scaling to capture and store entire streams of raw network flow records, not just summary metadata. It analyzes these flows at a highly granular rate for deep insight into network infrastructure that collectively represent huge carrying capacity, something the company describes as “terabit scale.” With this granularity, Kentik can detect bursts in activity missed by network monitoring platforms that rely on longer intervals for network data capture and analysis.

All of Kentik’s flow capture and analysis happens in the cloud, which eases the burden of long-term data storage on local internal infrastructure. This cloud-scale service also allows network operators to delve into long-term data captures to figure out what’s happening with their infrastructure. They no longer have to rely on alerts and original analysis of their network flows. They can dig in at any time and find answers to questions about infrastructure performance.

EMA believes that, with experience, more and more enterprises will be exporting network data for advanced analysis. As this occurs, they will need the right solution providers in place to help them execute. Network operators will increasingly need platforms like Kentik to glean essential, actionable information from their network data.