(a) Guide
Uptime, SLAs and failover explained
An SLA is a credit policy, not a guarantee that nothing breaks. Knowing the difference decides whether you need a second path.
Read the SLA as a set of numbers
Three numbers matter: the availability percentage, the response time, and the restoration target. Availability of 99.9% still allows roughly nine hours of outage a year; 99.99% allows under an hour. Response time is how fast someone picks up the ticket; restoration target is how fast they aim to fix it. Only the last one matters to your revenue.
Also check what the remedy actually is. Most SLAs pay a service credit against the monthly fee, which rarely resembles what an outage cost you. That gap is the argument for redundancy.
Work out what an hour of downtime costs you
Take your busiest hour: transactions you cannot process, guests you cannot check in, calls you cannot answer, staff who cannot work. Put a number on it. If that figure is larger than the monthly cost of a backup path, the decision makes itself — and the same figure tells you which sites in a multi-location group need redundancy and which do not.
Failover options, roughly in cost order
- Wireless or cellular backup that carries payments and phones only — cheap, enough for many small sites
- A second circuit from a different provider on a different physical path — the meaningful option for revenue-critical sites
- Automatic failover hardware so the switch happens without anyone noticing
- Diverse entry into the building, where the construction allows it
Different path matters more than different logo
Two circuits from two providers that share the same conduit into your building fail together. When redundancy is the point, ask specifically about physical diversity — separate facilities, separate entry — not just separate invoices.
Commercial service only · United States