Post-loadshedding "reconnect storms" occur when thousands of routers simultaneously flood the ISP's RADIUS server with PPPoE authentication requests. The 10-minute delay is normal — do NOT reboot during this period as it resets your queue position. A DC Mini-UPS eliminates the problem entirely.
Post-Loadshedding Reconnect Storms: The 10-Minute Auth Delay
The hum of the refrigerator returns, the lights flicker to life, and the localized load reduction block is officially over. Your first instinct is to grab your controller and jump back into the lobby, but as you look at your router, the "Internet" light remains a stubborn, mocking orange. You reboot the router, then the ONT, yet the connection refuses to "catch." For many South African gamers, this post-power restoration gap is the most frustrating part of the day. However, within the framework of Competitive Security, Edge Config & Continuity, this delay is not a sign of broken hardware, but a predictable technical phenomenon known as a Reconnect Storm.
In the 2026 South African infrastructure landscape, where we have transitioned from national stages of loadshedding to surgical, localized "load reduction," the way our network recovers has changed. When power returns to a specific suburb or grid block, thousands of devices—ONTs, routers, and smart hubs—all wake up at the exact same millisecond. This creates a massive authentication bottleneck at the local exchange that can leave you sidelined for up to ten minutes while the network settles.
The Anatomy of a Reconnect Storm
To understand why your fibre isn't instant, you have to look at the path your data takes to reach the internet. It is a multi-stage handshake that involves several pieces of "Edge Layer" and core infrastructure.
The Physical Layer: Your ONT (Optical Network Terminal) must synchronize its laser with the FNO’s (Fibre Network Operator) cabinet on the street.
The Data Link Layer: Your router must establish a connection with the BRAS (Broadband Remote Access Server) at the local exchange.
The Authentication Layer: Your ISP’s RADIUS Server (Remote Authentication Dial-In User Service) must verify your username and password to grant you an IP address.
What is a Reconnect Storm?
A Reconnect Storm occurs when a localized power return causes thousands of routers to send authentication requests to the ISP’s RADIUS server simultaneously. Because these servers have a finite capacity for "handshakes per second," they become overwhelmed. This creates a queue where requests are either delayed or dropped entirely, forcing your router to wait and try again.
PPPoE vs. DHCP: The Battle of the Protocols
In South Africa, different Fibre Providers use different methods to get you online, and these protocols behave very differently during a storm.
PPPoE (The "Stateful" Handshake)
Most legacy networks (like Openserve) use PPPoE (Point-to-Point Protocol over Ethernet). This requires your router to send a "Discovery" packet, find a server, and then provide a username and password. Because PPPoE is "stateful," the server has to maintain a constant session with your router. During a reconnect storm, the CPU of the BRAS can "peg" at 100% as it tries to manage thousands of these complex sessions opening at once.
DHCP / IPoE (The "Stateless" Assignment)
Newer networks (like Vumatel or certain MetroFibre segments) often use DHCP (Dynamic Host Configuration Protocol). Here, the router simply asks for an IP address, and the server gives it one based on the physical port identity. While DHCP is generally faster, it can still suffer from "Lease Storms" where the server runs out of available IPs to hand out in a specific "pool" during the initial rush.
The RADIUS "Death Spiral"
The real bottleneck is often the RADIUS Server. Think of this as the "bouncer" at the door of the internet. Even if the local exchange (the club) is empty, the bouncer can only check IDs so fast.
When 5,000 routers in a suburb like Sandton or Umhlanga all hit the RADIUS server at once, the server can experience a "Death Spiral." It tries to process the first 500, but while it's doing that, the other 4,500 routers timeout and send another request. This effectively creates a self-inflicted DDoS Attacks in Competitive Gaming against the ISP’s own infrastructure.
To protect itself, the RADIUS server will often start "dropping" requests silently. This is why you might see your router's logs showing "Authentication Timeout" or "LCP Terminated" for several minutes after the power returns. You can see these entries for yourself by following our guide to Router Logs & Intrusion Detection.
Exponential Backoff: Why Your Impatience Makes It Worse
If you are the type of gamer who frantically reboots their router every 60 seconds when the internet isn't working, you are likely extending your own downtime.
Why does rebooting the router during a reconnect storm delay the connection?
Most modern routers use a logic called "Exponential Backoff." If a PPPoE connection fails, the router waits 10 seconds before trying again. If it fails a second time, it waits 20 seconds, then 40, then 80. If you manually reboot the router, you reset this timer, but you also put yourself back at the very end of the RADIUS server's queue. By constantly rebooting, you are essentially "cutting in line" only to be sent back to the entrance over and over.
The best strategy is usually to leave the hardware alone for at least 10 minutes. If the Network Status page doesn't show a major regional outage, the system is simply "grinding" through the backlog of reconnections.
How to Bypass the Reconnect Storm Entirely
There is only one foolproof way to avoid the 10-minute authentication delay: Never let your router or ONT lose power.
In 2026, the transition from grid power to battery is the most common cause of session drops. If your ONT reboots, the "state" of your session is lost at the exchange. When the ONT comes back up, it has to start the entire handshake process from scratch—joining the thousands of other people who don't have a backup.
The Solution: Dedicated DC-to-DC UPS
By using a correctly dimensioned battery backup, you ensure that your ONT and router maintain their "handshake" even when the grid fails. Because the session never "dies" at the ISP level, you completely bypass the reconnect storm when the power eventually returns. Your data is already "authenticated," so you stay online while your neighbours are stuck in the 10-minute queue.
For a technical breakdown on how to pick the right unit for your specific router’s voltage, see Mini-UPS Dimensioning.
Diagnosing the "Post-Power" Lag
Sometimes, you do get back online quickly, but your game feels terrible. You might notice 200ms ping to local servers or 5% packet loss. This is "Post-Storm Congestion."
When a grid block returns, every smart TV starts its Netflix updates, every PC starts its Windows updates, and every phone starts its iCloud backups simultaneously. This massive spike in "background traffic" can saturate the local fibre cabinet (the "splitter") for the first 15 minutes of power returning.
To see if your lag is caused by this congestion or a deeper issue, check the Live Ping dashboard. If the jitter is high across your entire suburb, it is simply a "traffic jam" on the fibre glass that will resolve once everyone’s background updates are finished.
Summary: Patience is a Technical Virtue
The 10-minute authentication delay is a byproduct of the physical reality of our power grid interacting with high-speed data protocols. In 2026, as our FNOs become more resilient, these "storms" are becoming shorter, but they remain a factor in every South African gamer's life.
Don't Reboot: Let the router's internal backoff timers handle the queue.
Stay Informed: Check the Network Status for FNO-level faults.
Monitor your logs: Use Router Logs & Intrusion Detection to identify if it's a "Timeout" or a "Physical Down" issue.
Invest in Continuity: A Mini-UPS is the only way to "skip the line."
By understanding the "Edge Layer" mechanics of authentication, you can stop fighting your hardware and start planning your sessions more effectively. If your connection remains down for longer than 20 minutes after power returns, our Support Robotics platform can help you perform a remote reset of your port to clear any "stuck" sessions at the exchange.