5 things analytics could tell you about your network in 2018

This isn't a pie-in-the-sky wish-list. All five of these insights are available today, with the right analytics system.

data analytics chart money finance laptop computer

1. Whether your users are happy (without having to talk to them)

It’s not always cool to admit, but the ultimate goal of every networker is to have happy users. Like many other thankless jobs, we only hear about problems. When we do, we react. But that isn’t ideal. What we really want is to know about problems as they are developing, before users complain. They don't even have to know.  But we do.

A Network Management System (NMS) has been the traditional go-to solution to sniff out these sorts of problems. But most were designed for just one view of a certain part of the network using antiquated technology that doesn't provide any sort of predictive problem solving based on what the user is actually experiencing. It's like trying to figure out San Francisco traffic based on the status of the traffic signals. Just because the signals are working properly doesn’t mean the drivers (users) are having a good experience.

Modern analytics systems focus on actual user experience by looking at, you guessed it, user experience.  Just like Google Maps or Waze, these systems look at user experience and, if there is a problem, proactively notify you so you can swing into action. And even better than San Francisco traffic, these systems will even tell you how to fix the problem.

2. Not all your WiFi access point radios are working, no matter what your NMS says

All enterprise class WiFi systems have a built in NMS with varying functionality. The most basic feature is the typical red/amber/green access point status. What may surprise you is that a green status doesn’t actually mean the AP radios are allowing connections and transferring data.

It's not uncommon to find dozens of radios in a large network that are unresponsive. When a radio is down, especially a 5 GHz radio, that manifests as poor performance and coverage gaps that can take a number of helpdesk calls to discover.

On the other hand, it’s quite typical for a fully functioning AP to go unused for hours, days or even weeks, depending on the network. So how can a solution tell the difference between an AP that isn’t responding and an AP that just isn’t busy? That’s where analytics comes in: establishing and recognizing patterns of usage and reporting when that pattern is out of whack. And to further complicate matters, the patterns can’t be based on the behavior of a single AP. The entire network must be baselined and constantly analyzed to detect these anomalies. 

3. ARP is failing you more than you would think

The Address Resolution Protocol (ARP) is one of those protocols that no one ever really thinks about because it’s just supposed to work. Well, in most networks it works without any issues at all.  In others, it can be a big problem. For large enterprise networks ARP can be a cancer that causes slowed response times and unpredictable behavior.

Unless you are specifically looking for an ARP problem (and who has time to do that?) these issues go largely undetected. But a full-stack analytics system (yes, there is a theme here) can detect issues like this and even pinpoint where in the network they are occurring.

4. What system (Wi-Fi, applications, DNS, DHCP, cloud, etc.) is to blame for your current network problem

It may feel a bit childish, but we've all had to point fingers when network or user problems occur. Wi-Fi has always bared the brunt of the blame when something isn’t easily explainable. But that blame is largely unsubstantiated; it just feels good to blame the system that is complicated and is barely in touch with its own feelings.

The best way to place the blame is to have an unprejudiced system objectively analyze information from all aspects of the network and make deductions based on real information.

5. How your network performance ranks among organizations like yours

It’s natural to want to know how we stack up against others, whether it’s our girlfriend’s ex-boyfriend(s) or who has the best man cave (I win, sorry). As a network engineer we also want to believe that our network is best, and we have the happiest users. But until now, that hasn’t been readily quantifiable. This is changing, fast, as analytics and cloud computing converge to make it possible to compare virtually any part of your network with another enterprise that looks virtually identical.

Besides stroking (or hurting) your ego, that information has huge value. If you see that, among networks like yours, that your Wi-Fi performance is in the top 10%, you know you don’t need to dedicate a lot of resources into making it better. But if you see that your Office 365 application performance is in the bottom 20% you’ll want to make that better. And a great analytics platform can even tell you why you are in the bottom 20%. And, what would be really cool is if you could click on the top company in Office 365 performance and ask them why they are so good?

That sounds like real social networking to me. But actually useful social networking.

This article is published as part of the IDG Contributor Network. Want to Join?

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: 10 new UI features coming to Windows 10