Latency
Round-trip time to the endpoints that matter, sampled continuously rather than when someone complains.
Products Communications-quality observability and operational intelligence
Beacon
A small agent at the customer site that continuously measures the conditions affecting communications quality and reports them back for visibility and investigation. Built to answer "was it the network?" with evidence instead of opinion.
In active development.
How it works
Measure where the problem is, not where the tools are.
Most monitoring watches the middle of the network. The experience that people complain about happens at the edge of it — in one office, on one link, at one time of day.
Beacon reports what it can measure from its vantage point. Where an environment does not support deriving a particular indicator, it reports the underlying measures rather than estimating one.
Demo
A degradation, as Beacon would show it.
Four sites, one of them degrading. Latency, jitter and loss are drawn as three separate panels with their own scales — plotting measures with different units on one axis makes a chart look decisive and read wrong.
Sites reporting
- Degraded Harbour Road — main office MOS 3.6 RTT 94 ms Loss 3.1% Upstream congestion
- Healthy Kiln Street — fabrication MOS 4.3 RTT 31 ms Loss 0.0% Nominal
- Healthy Fenwick Depot MOS 4.2 RTT 44 ms Loss 0.1% Nominal
- No data Alder Yard — remote MOS — RTT — Loss — No agent check-in for 14 min
Harbour Road · last 2 hours
36ms peak 121ms 121ms
3ms peak 38ms 38ms
0% peak 4.2% 4.2%
Degradation timeline
- 14:05 Latency and jitter rising at Harbour Road
- 14:20 Packet loss crossed threshold — degradation opened
- 14:35 Path change detected on upstream hop 3
- 15:10 Metrics returning to baseline
What it measures
Nine measurement categories.
What any given deployment reports depends on what the environment supports. We would rather tell you a measure is unavailable at your site than show you a number we inferred.
Jitter
Variation between packet arrivals — usually the first thing to move when a call starts sounding wrong.
Packet loss
What proportion of traffic never arrives, measured on the path that actually carries the calls.
Availability
Whether the connection and the endpoints are reachable, and for how long they were not.
Connectivity
Reachability of the services the site depends on, distinguishing a local fault from an upstream one.
DNS behavior
Resolution success and response time, which quietly breaks more sessions than people expect.
Path quality
Where the traffic goes and where it degrades — the hop that changed, not just that something did.
Historical trends
The same measures over weeks, so a gradual degradation is visible before it becomes an outage.
Voice quality indicators
MOS and related indicators where the environment supports deriving them. Where it does not, Beacon reports the underlying measures rather than estimating a score.
How it is built
Deliberately small.
Lightweight by design
A small agent with a narrow job, not a management platform that needs its own project to deploy. It measures, it reports, and it stays out of the way.
Outbound only
Telemetry leaves the site over an outbound connection. No inbound firewall rules, no exposed listener, no new attack surface at the customer edge.
Measures from where it hurts
Testing from a datacenter tells you the datacenter is fine. Beacon measures from the office where the calls are actually breaking up.
Evidence, not opinion
When the answer is "it is not our network," Beacon is what lets you demonstrate that — with a timeline you can hand to whoever owns the link.
What changes operationally.
No uptime percentages or resolution-time figures appear here. Beacon is in active development and we have not measured those across enough deployments to publish one.
A quality complaint arrives hours after the problem, with nothing to look at.
The window is already recorded, with the measures that moved.
Diagnosis means running a test now and hoping the fault reappears.
The history shows whether this is new, recurring, or slowly worsening.
Fault ownership is argued rather than established.
Path and timing evidence points at the segment that actually degraded.
Each site is investigated in isolation.
Sites are compared, so a shared upstream cause becomes obvious.
Start with the customer journey
Put a Beacon where the complaints come from.
If you have a site with a recurring quality problem nobody has been able to pin down, that is the one we would want to measure first.
Prefer email? info@rprtechllc.com