Built to Stay Up: Reliability at Cannon VoIP

Your phone system is only as good as its uptime. Here's what actually backs Cannon VoIP's reliability, in plain terms, no vague marketing promises, no engineering jargon you'd need a manual to decode.

99.99% uptime, backed by service credits

99.99% uptime isn't a badge we picked because it sounds good, it's a contractual commitment. If we ever fall short of it, your account receives a service credit. We're comfortable putting a number behind our reliability because we build the platform to actually hit it, not to sound good in a sales conversation. The exact terms are in Section 5.6 of our terms.

Geographically redundant data centers

Your service runs primarily out of an undisclosed location in Fort Worth, Texas, with two additional backup regions elsewhere in the U.S. ready to take over automatically. Spreading service across geographically distant locations means a problem local to one region, severe weather, a power event, anything that could take down a single site, doesn't take down your phone system with it.

Automatic failover

If one of those data centers goes offline, service fails over to another automatically, no one has to notice the problem and flip a switch manually. In practice, that failover shows up as a brief gap of roughly 700 milliseconds, well under a second. Calls in progress don't disconnect. You might notice the briefest blip; you won't be hung up on and left redialing.

Engineered for voice quality

We don't chase a single quality number to put on a spec sheet. The network is engineered specifically for voice traffic, prioritized so call quality holds up rather than degrading under everyday network load. The goal is simple: clear audio, no choppiness, no awkward lag, on every call, every time.

Your data, backed up every minute

The full database is backed up every minute and stored off-site. Backups are checked for completeness, and restores are practiced before they are ever needed. Recordings, voicemails and faxes go to durable cloud storage as they are created, not in a nightly batch.

Updates without dropped calls

When we update a server, live calls finish before it restarts. The portal and sign-in run as more than one copy, so one failing does not take them down, and the apps remember backup server addresses, so they can still connect if DNS fails.

Watched around the clock

Every server and service is monitored 24/7. Our public status page does not depend on the portal app, so it loads even when the portal is down, and it has a machine-readable feed if you want to track it from your own systems.

Run by an experienced team

Every system is documented, the team is cross-trained, and a round-the-clock on-call rotation means someone who knows the system is always available. Changes to our code and infrastructure are reviewed and versioned before they go live.

Live technicians are available 24/7/365 for anything urgent, and the full support team works weekdays, 9 to 5 Central. It is built for two-user offices all the way to enterprise and government deployments, with multiple layers of failover and redundancy.

When your office internet goes down

Your phone system lives in the cloud, but your office still needs an internet connection. If every phone on your account goes offline, internet outage backup forwards your calls to another number, like a cell phone, so callers still reach you. The apps also move between Wi-Fi and cellular without dropping the call.

A true phone company

None of the above is possible if you're reselling someone else's platform. Cannon VoIP is a true phone company, not a reseller. We run our own switching, servers and apps, including our own softswitches and Session Border Controllers (SBCs), and interconnect with two carrier partners. We do not white-label a third-party UCaaS platform like a lot of "VoIP providers" do. That's why we can commit to a 99.99% uptime SLA in writing: we control every layer the guarantee actually depends on.

For voice and texting, only three outside companies are involved: two carrier partners that connect our network to the public phone network, and the national STIR/SHAKEN authority that issues our caller ID certificates. Here is why that matters for reliability and security:

• Not tied to one carrier. Two carrier partners, not one.
• One accountable company. No finger-pointing between a reseller and a platform vendor. The engineers who built it are the ones who fix it.
• Fixes on our schedule. Not a vendor's roadmap. Most feature requests ship in under four weeks, and simple ones often in under a week, depending on complexity.
• No platform end-of-life risk. No outside vendor can discontinue, sell or reprice the platform underneath you.
• Fewer companies touch your data. Calls, recordings and messages live on systems we run, with no reseller layer that has its own logins and support staff.

Our AI features are our own applications; today they use two outside AI providers for the language model and voices, and we are building our own AI servers to bring that in-house too. See our network for the full picture.

For the plain-English version of what makes cloud phone infrastructure reliable, read what is a hosted PBX and what happens to your phones in an outage.