Packetman saysThe DataStun speed test measures something different from the speed tests you're used to. A typical speed test measures raw throughput to a nearby server — it tells you the best-case capacity of your connection under ideal conditions. That number rarely matches what you experience in real use. What actually degrades your perceived performance is TCP retransmission rate — how often your packets have to be resent because they didn't arrive. High retransmission means your packets are being dropped somewhere between you and the destination, and TCP is silently resending them. You experience this as slowness, timeouts, and video stutters, even when your speed test shows 500 megabits. Our test measures retransmission rate in real time, so you can see the actual health of your connection — not just its theoretical maximum.Turn every agent in your fleet into a performance probe. Every pair produces a bidirectional latency, throughput, jitter, and packet-loss measurement — and a traceroute path each way. 500 agents = 124,750 measurement pairs. When a user says “the VPN is slow,” you do not argue.
Per-agent add-on on Business tier and above. Measurements run on a schedule, on demand, or both.
Hover Packetman for the 30-second pitch.
Four numbers per direction. Because TCP is half-duplex and most of the "network is slow" stories are hiding in one direction and not the other.
Round-trip and one-way where NTP sync permits. A→B and B→A reported independently because return paths can differ from forward paths — and usually do on VPN and SD-WAN.
Actual bits-per-second sustained over a measurement window — upload and download separately. This is what your users experience, not what your ISP advertises.
The metrics that decide whether VoIP and video actually work. Average latency is a poor proxy; a 50 ms link with 10 ms of jitter breaks calls while a 150 ms link with 2 ms jitter carries them cleanly.
Traceroute from both ends with per-hop GeoIP. Asymmetric routes (A→B goes via your primary ISP; B→A comes back via a peering partner) stick out immediately in the dashboard.
Every agent is both a probe and a target. The server dispatches measurement jobs, agents run them, results roll up into the dashboard.
For N agents there are N(N−1)/2 pairs. 100 agents = 4,950 pairs, 500 = 124,750, 1,000 = 499,500. Measurement runs in batches so heavy fleets do not saturate themselves; scheduling is tunable.
Return paths that take a different route than forward paths show up as asymmetric latency or throughput, automatically flagged. This is the pattern behind most “the app is slow only sometimes” support tickets.
Agents behind NAT participate in the mesh through a STUN negotiation. You do not need a VPN or public IP on each machine to get fleet-wide measurements.
Browser-based ad-hoc tests between any two endpoints over WebRTC DataChannel — no agent required on at least one side. Useful for remote users you cannot install software on.
Cron-scheduled recurring tests mean “the VPN is slow today” becomes “the VPN was 3× faster last Tuesday at the same hour.” Baselines turn noise into signal.
Measure what actually flows across your VPN or SD-WAN — not what the vendor dashboard claims. If the tunnel is lying about its throughput, the N² test is where you catch it.
Real-traffic answers to the questions that normally require a field technician, a per-site test run, and a week of calendar time.
Your 500 Mbps branch circuit tests at 180 Mbps down, 40 Mbps up during business hours. The N² dashboard shows the gap, times of day, and the offending direction. That is your evidence when you call your ISP.
Branch 14 has acceptable forward latency to HQ but return paths are 80 ms longer than the other 40 branches. Asymmetry detection flags it; the traceroute shows the return is going via a peering partner your primary ISP no longer maintains.
Run a baseline the day before a router upgrade, the day after, and a week later. You know what actually changed — not just what the change ticket claimed would change.
“Teams is slow this week” — pull up the user’s agent, look at the sessions dashboard, and compare the throughput chart to last week’s baseline. You know whether it is Teams, the user’s network, or their device before you open a ticket.
Speed Test is a per-agent add-on, available on Business tier and above.
Per-agent monthly subscription. Enables the agent as a measurement probe and target in the N² mesh. Scheduled recurring tests plus on-demand from the dashboard. Results retained per your tier’s retention.
Minimum one agent. No commitment beyond the month. Pro-rated on enable/disable. Volume discounts and annual/multi-year pricing available.