Can Analytics be Flamingo enough for the Network act?

Pratima H

SAN FRANCISCO, USA: It may be true or not but many people, including neurology experts, researchers and of course, unsolicited advisers, swear by the 20 second test. Stand on one leg and try balancing for at least 20 seconds. The results will vary with eyes open and eyes closed, and that’s where one is expected to assess one’s brain’s health and tendency for neurological problems, strokes and even dementia.

Visibility can change a lot, specially when it comes to complex neural networks and the tough balancing act of different muscles, neurons and limbs. Having the support of eyes and ears makes a lot of difference.

That difference turns out to be quite significant in a test of one’s sensory circuits. Wouldn’t it be as or almost significant when it comes to those back-end circuits that condense into networks?

And if so, what kind of eyes and ears are we looking at to avoid downtime and network snags, where even 20 seconds can make any infrastructure wobble?

We ask Alex Henthorn, VP Marketing, Kentik which came out with a SaaS-based network traffic analysis platform a year back and already claims a roster of over 60 customers, including some of the world’s notable web enterprises like Box, Yelp, Pandora, Shopify, Dailymotion, Neustar, OpenDNS and Instart Logic.

On the sidelines of a Series B funding announcement, he explains the relevance of analytics for networks, both for service providers and digital enterprises. It can be a case of content delivery network (CDN) infrastructures or internal traffic for an enterprise; but for a variety of network mazes, Henthorn, here helps to figure out why most networks are struggling with the balancing stance?

How crucial is the need for network analytics today?

Network analytics is becoming more crucial as business models become increasingly digital. A Digital business relies on reliable, predictable and high performance delivery of network traffic, particularly Internet traffic in the age of the cloud.

Do the imperatives differ for CSPs and enterprises?

Yes, imperatives differ. Let’s take Service providers first. Whether an IaaS or hosting company, a mobile operator or cable MSO, these companies' service is primarily related to delivering network traffic. They have very large infrastructure footprints, sometimes globally. The need to meet end-customer SLAs (Service Level Agreements) as well as manage cost-performance of a massively expensive infrastructure, plus give customers visibility via portals into how traffic is being delivered –these are imperatives around network analytics for these businesses.

If you talk about Web or online enterprises; this is a large and complex sector of businesses with myriad inter-dependencies but they all derive their revenue from applications or services delivered via network traffic. Primary brands may be delivering streaming music, social reviews, e-commerce, B2B services, or mobile app experiences.

These well-known enterprises such as Yelp, Box, Pandora, Dailymotion, Shopify, depend on an eco-system of less-known, enabling online enterprises that provide advertising, reservations, maps, security and other services. All of these companies must deliver increasingly competitive user experiences to garner views, listens, participation, checkouts, reservations, ad clicks, etc. Those user experiences rely critically on network performance. Many of these businesses end up building large, private content delivery network (CDN) infrastructures to ensure performance and user experience. The imperative for these businesses is competitive user experience that delivers brand loyalty and revenue.

What about non digital-native enterprises? Still relevant?

Most of these businesses are not primarily garnering revenue from online applications or services, but many of them are building digital business arms (think e-commerce like Walmart.com) or mobile and online experiences to enhance the engagement of prospects and customers, such as in banking. Since digital business is where the greatest profit and competitive leverage is for these businesses, they need to emulate digital leaders.

At the same time, these businesses are increasingly turning to cloud-based solutions for their IT and enterprise application software needs, which means that network traffic is increasingly critical to productivity and essential business processes. The imperative for these businesses around network analytics is two-fold therefore. First is to succeed and compete in their digital business initiatives, like a web enterprise. Second is to ensure maximum IT productivity and cost-efficiency in a hybrid cloud model.

When the main subject here is the network, what implications do SDN (Software Defined Network) and NFV (Network Function Virtualisation) bring to this space?

SDN and NFV, in the sense that the ultimate goal is to increase automation, responsiveness and efficiency in network operations; are highly complementary. Without deep analytics and accurate intelligence, how do you make effective automated remediation happen? Without strong planning insight, how do you manage load for NFV deployments?

Would analytics assume a stronger or a different role when IIoT or IoT gains traction?

There is certainly a whole sector of analytics focused on IoT and other related technologies themselves. IoT analytics are focused on doing business analysis of the telemetry data coming from devices, for example. However, network analytics of course, will become more important to the degree that any business develops critical business process dependencies on communicating reliably and cost-effectively with a huge number of devices distributed across internal and Internet networks.

Alex Henthorn, Kentik

What makes you confident about what you offer as a solution to the problem?

Kentik Detect is a cloud-based network visibility and analytics solution that delivers unprecedented depth of insight into any network. Kentik processes tens of billions of data records per day, equipping service providers and enterprises network operations teams with actionable insights that enable them to make quick, cost-effective decisions.

Can you augment that with any outcome and usage examples where you see your customers gaining from network analytics?

There are many example use cases. Take a scenario where network operations that have to rapidly figure out why application performance is degrading, find and fix network root causes. In most cases, network analytics only provides summary level views. These offer clues, but not enough detail to enable action. Where there is any detail available, it often takes far too long to get at to be practical. Kentik offers ad-hoc analysis of huge volumes of network traffic and performance details with response in a few seconds, which means that engineers can iterate their analysis rapidly and answer the precise questions they need to take effective action, in minutes.

So before Kentik, there was lots of educated guesswork, wasted time, degraded performance, high MTTR (Mean Time To Repair), lots of inconclusive outcomes when problems fade before they can be understood. After Kentik this changed into highly accurate operational actions, far higher fix rate, lower MTTR, improved network performance,

What about areas like security, sprawl, visibility, latency with reference to networks of the present era?

Regarding network security, DDoS is a major scourge for digital business. The traditional answers for detection have been very appliance-based and thus trapped in Moore's law era, scale-up computing limitations. Cloud-scale, big data analysis frees detection capabilities from those constraints and makes it far more accurate, powerful and useful.

How do you see the direction of your company shaping forward with this Series B funding?

The direction of our company to be fueled by this funding is defined by the ambitious vision we've set forth to help our customers unlock the value of network data, so that they can innovate. We believe that data has the power to increase operational excellence, improve user experience, increase competitiveness, bolster revenue. Ultimately, we want our customers to become more innovative in all those ways by using rich insights into the critical network delivery aspect of digital business.