Digital experience

Digital experience, measured from the user's desk

Every two minutes, from every floor, CNA asks the question your users ask: can I get to the apps I need, and how fast? When the answer is no, it shows where the time is lost.

What an office sensor measures

A small sensor sits under a desk or on a wall and joins your network like any user device. It measures the service 24/7, independent of anyone's laptop.

Applications and DNS

Up to 12 apps per sensor, every 2 minutes, plus DNS lookup time.

Network path

Latency and loss at every hop on the way to each app.

Wi-Fi experience

Signal, quality, rate and access point, every 15 seconds.

Radio scan

The 24 strongest access points nearby, every 15–20 minutes.

WAN circuits

Each site's circuits probed from the cloud, with availability.

Speed tests

On demand, switching themselves off when done.

Sensors join enterprise Wi-Fi like any user device: WPA2 and WPA3, PEAP, and certificate-based EAP-TLS. Network requirements.

A floor plan with three Wi-Fi sensor icons: one green at −66 dBm, one amber at −69 dBm and one red at −76 dBm.
A live floor map: each sensor coloured by signal strength.

Wi-Fi where people sit

See which desks were ever going to get a good connection

Each floor's map shows every sensor coloured by the signal it receives: green above −67 dBm, amber from −68 to −75 dBm, red below −75 dBm. A time slider replays a bad morning, so you can see when a problem started and where.

Behind each sensor is the detail your wireless provider needs: the network and access point it is connected to, the channel, signal and quality over time, every roam between access points, and the networks around it.

Choose the sensor mix for the job: one service-path sensor for every 10–12 desks, and Wi-Fi probes at every other desk where you need Wi-Fi measured seat by seat. Sizing and packages.

Localise the fault

Finding the fault in minutes, not meetings

CNA doesn't guess at root cause. It measures the same service from every angle, so the answer is usually obvious.

Three response-time graphs over five days for a collaboration app, a print service and an office suite. Only the office suite shows a spike, to about 4 seconds.
Same site, several apps. One app jumps to 4 seconds while the others stay normal: look at the app, not the network. App names generalised.
A traceroute timeline with each hop's latency over three hours. One hop spikes to 36 ms while the others stay below 10 ms.
Hop by hop. One hop spikes to 36 ms, so you know where the delay starts.
A Wi-Fi signal graph coloured by access point. After roaming to a new access point the signal improves from about −70 to −60 dBm.
Wi-Fi roaming. Each colour is an access point. After the roam the signal lifts from −70 to −60 dBm.

Sustained problems raise an alert in about 10–15 minutes, day or night, often before the first call. That figure is an estimate from the default alert settings and can be tuned for you. How alerting works.

End-user device agent

The same view from the user's own laptop

Sensors give a constant, device-independent view of each location. The laptop agent adds what one person's Windows device is doing while it is switched on, wherever they work.

  • Every device's CPU, memory and disk at a glance
  • Wi-Fi signal, rates and the access point in use
  • Reachability of the same apps the sensors test
  • If you use Zscaler, the client's connection details

Sensors fine but laptops slow? Then it's the device, not the network.

A table of Windows laptops showing free CPU, memory and disk as green, amber and red cells, Wi-Fi signal bars and agent version.
One row per device: model, free CPU, memory and disk, Wi-Fi signal and agent version.
Latency for a site's two WAN circuits over one hour, mostly 2 to 3.5 milliseconds with one spike near 7 milliseconds, and zero packet loss.
Site WAN: both circuits, latency on the left axis and loss on the right.

WAN circuits

Your circuits, tested from the outside

Cloud probes test each site's circuits continuously, so you can tell an ISP problem from an office problem. Each site can have two circuits; if a circuit blocks ping, its probe switches to trace mode.

The home view also carries status feeds from AWS, Azure, Microsoft, Cisco, Fortinet and Okta, so "is it us or them?" is answered in one place.

Prove it on your own network

Six weeks, your most-complained-about office, the apps your people can't work without, and success criteria you agree up front.