A recent Gartner report on network performance monitoring and diagnostics (NPMD) estimated the market to a whopping $2.1 billion and growing at a compound annual growth rate (CAGR) of 15.9 percent, with more growth in sight. Wow. So what will drive this growth and why?
New approaches to harvesting network data using sophisticated big data analytics techniques combined with cloud computing and machine learning technologies is the answer. This perfect confluence of technologies is poised to redefine the conventional infrastructure management market.
Central to this shift is the use of analytics technologies and strategies to extract new insights and value from data produced by and collected from the network to drive business value.
Network performance is now inextricably linked the success of digital business initiatives. But network complexity, cloud-hosted workloads, endpoint diversity and software-defined architectures are present big challenges for networkers.
A better Splunk
New user performance management platforms are taking the Splunk model one step further by ingesting more than merely log data and providing basic query facilities. They are adding predictive analytics to effectively every part of the network from the lens of actual user experience. This eliminates the tedious and time-consuming analysis of data that's bogging down IT departments from being an actual profit center while helping network manager cut the time to remediate client, device and IoT issues in half.
According to Enterprise Management Associates (EMA), IT automation is the primary driver for network analytics with end user monitoring being the top use case.
Gartner suggests that enterprises that tend to invest broadly in packet and infrastructure monitoring solutions, often across different vendors, still have issues with the "consumability" of their packet monitoring solution. And those that invest more narrowly in infrastructure-focused monitoring, report frustration in their inability to do root cause analysis without deeper monitoring. In other words, how we monitor and manage the network infrastructure is due for a widespread reset.
Enterprise networks themselves have seen tremendous evolution, from Wi-Fi advances and software defined networks, to the explosion of mobile and IoT devices running on the networks. It’s time our methods of monitoring and managing these complex and advanced networks catches up.
And that’s just the time for IT operations teams doing their jobs! Don’t forget about the downtime for the end users who can’t do their jobs because of a poor performing network. Improving user productivity has real teeth for businesses that view the network as a strategic asset, which is nearly every business today. But heretofore user productivity on the network has been difficult to quantify without purpose-designed analytics.
Before mobility, Wi-Fi, cloud and IoT, fixing network problems was straightforward. Users simply accessed the network via Ethernet jacks to use local applications and resources. If there was a problem, the help desk would check a handful of vendor specific and siloed monitoring tools that might give them an idea of what's wrong but not a complete picture.
That model of monitoring and managing the network infrastructure no longer works. The focus has now turned to managing user and device network performance.
Too much data, too little time
There's a treasure trove of invaluable data that can help companies run their businesses better. And it's sitting right in front of them, running over their own corporate networks. But getting to this data and making sense of it all has become a monumental challenge for enterprises looking to improve the productivity of users. This is where analytics becomes immensely useful.
In a recent ZK Research survey on Wi-Fi troubleshooting, 60 percent of respondents said they spent a least a quarter of their time doing nothing but troubleshooting Wi-Fi issues; 47 percent said it takes 30 minutes or more to diagnose a problem and 41 percent said it take an additional 30 minutes to solve the issues.
Today, users connect via Wi-Fi using a smart device – often a BYOD device not issued by the company. Then there’s the process of authentication, which includes being assigned an IP address and (eventually) getting to the necessary application – which is many times in the cloud. If there’s an issue with any of these steps, users are unhappy and simply blame “IT” or “the network.” But what part of the network?
Needed: Full stack user performance management
Full stack user performance management platforms now stare at millions of client network transactions as they happen to figure out what's working, what's not, where, when and why. But they don't stop there. They then apply artificial intelligence to the data to reveal insights from the data on which action can be taken with value ascribed.
This requires massive amount of data, processing power and machine learning. Until now, this hasn't been really possible.
EMA points out that the majority of enterprises want this functionality embedded in the network. Innovative solutions are taking this exact approach with non-intrusive software that analyzes network data without the need for hardware Wi-Fi sensors, server agents or client software. All the analysis then shipped to the cloud where companies can use it. This new model is driving big investments in infrastructure management and is widely considered a major cause in the market's growth. And the big benefactors of this growth look to be little known upstarts pioneering the way like Corvil, LiveAction and Nyansa.
Clearly, advanced network analytics is an approach to network monitoring and IT infrastructure management whose time has arrived. And when coupled with machine learning, network operations teams will not only be able to address enterprise network problems in real-time, they also will be able to predict where network problems may be brewing, moving the IT department one step closer to being viewed as a profit center versus a cost center.