Network resilience
No single point of failure between your event and the internet
Two carriers on separate routes, a third backup path, a carrier-grade core router and protected power — then we unplug the main line before doors open to prove it works.
Six layers of resilience
Each layer removes one way an event network commonly fails. Together they mean a cable cut, a carrier outage or a power dip doesn’t reach your audience.
Two carriers, two routes
Primary and secondary internet come from different ISPs and enter the venue along physically separate paths. One cable cut or one carrier outage can’t take the event offline.
We order the secondary circuit from a different carrier and check that the two routes don’t share a duct, pole line or building entry. Where the venue has only one entry point, the second path arrives by point-to-point radio instead.
A third path for what can’t drop
Bonded 4G/5G or a point-to-point radio link stands behind both fibers, reserved for the live stream, payments and the organiser’s control desk.
The third path is policy-routed: if both fibers fail, the systems that must not drop move onto it first, and general guest WiFi is throttled so they keep enough capacity.
Carrier-grade core router
All three paths terminate on a MikroTik CCR2216 cloud core router — 16-core ARM at 2.0 GHz, 16 GB RAM and dual hot-swap power supplies. It checks every path continuously and moves traffic to the next one automatically, with no one touching a cable.
Health checks test real reachability to the internet, not just link lights, so a carrier that is “up” but not passing traffic still triggers failover. Critical traffic is policy-routed onto the backup paths first.
Switching and priorities
A MikroTik CRS326 switch (24 Gigabit ports, 2 × SFP+ 10G) connects to the core over 10G and feeds WiFi, streaming, CCTV, registration and F&B outlets — each on its own network, with streaming and payments prioritised over guest traffic.
Each system gets its own VLAN and its own bandwidth reservation, so a busy guest network never squeezes the stream or the registration counters. Larger sites add further switches per hall or zone.
Protected power
Online UPS on the core, switching and PoE, sized to ride through generator changeovers and venue power dips.
We plan power with the venue electrician: separate feeds for the network rack where possible, and UPS runtime covering the longest expected changeover.
Tested, watched, staffed
We pull the primary link before doors open to prove failover works. Engineers monitor every uplink and access point live, with spare hardware on site.
The failover test is part of the handover to the organiser. During the event, alerts go to the engineer on site and to a second engineer remotely.
The core difference
| Feature | Best effort | Built for failover |
|---|---|---|
| Primary goal | Maximise performance under normal conditions. | Guarantee uptime, even during a catastrophic outage. |
| Redundancy | Minimal or none (single point of failure). | Full duplication (N+1, 2N or active-active setups). |
| Response to error | Degrades, drops packets or crashes entirely. | Automatically switches to a backup with zero or minimal interruption. |
| SLA commitment | “We will try our best.” | “99.999% uptime guaranteed.” |
Enterprise equipment, not consumer routers
We build event networks on hardware from established enterprise brands, chosen for the job each part of the network has to do.
- Ruckus WirelessHigh-density WiFi for keynote halls and expo floors, where hundreds of devices share each access point.
- CiscoEnterprise switching for large multi-hall and venue-wide networks.
- MikroTikCCR2216 cloud core routers for carrier failover and CRS326 switches for distribution across the venue.
- Alta LabsCloud-managed WiFi 6 access points, indoor and IP68 outdoor, with per-client VLANs and multi-password SSIDs.
- GrandstreamWiFi 6 access points, including long-range IP68 outdoor models for grounds and stadiums.
- TP-Link OmadaManaged PoE switching and access points for booths, counters and cameras.
Brand names are trademarks of their respective owners.
How a deployment runs
Survey
We walk your venue with the floor plan, check which carriers can reach it and by which routes, and plan rack space and power.
Design
You get a written network plan: bandwidth, both carriers and their routes, the backup path, access point and camera positions, VLANs and a fixed quote.
Circuits in
Two ISP partners provision dedicated circuits on separate routes; our crew terminates and tests each one.
Build and test
The core router, switches, access points and cameras go in during load-in. We load-test, then unplug the primary link to prove failover before doors open.
Live
An engineer watches every link and access point from the network desk for the whole event.
Teardown
Everything comes out after the show. You receive a usage report and any recordings.
Failover questions
What happens to guests when the primary internet fails?
Traffic moves to the standby path automatically, usually within seconds. Browsing and messaging carry on; a few long-running sessions such as video calls may need to reconnect.
Will our live stream drop if a link fails?
The streaming encoder sits on a reserved, prioritised path with its own backup. For high-stakes broadcasts we also use bonded or SRT contribution, which rides through a path change without a visible break.
Why two carriers instead of one bigger line?
A bigger line from one carrier still fails when that carrier has an outage or the one cable into the venue is cut. Two carriers on separate routes remove both risks.
Do you test the failover?
Yes. Before doors open we unplug the primary link in front of the organiser’s team and confirm the network, the stream and payments keep working.
What if the venue has only one cable entry?
Then the second path arrives by point-to-point radio or bonded 4G/5G so it doesn’t share the same physical route.
Running a keynote, a launch or a live broadcast? Let’s design it so it can’t go dark.