The One Million Route Problem | What It Is & How to Survive It

If you're managing Internet-connected routers, you've either encountered this problem—or you soon will. The 1 million route limit is a looming challenge in the world of Internet routing, particularly for ISPs, cloud providers, and enterprises managing global-scale BGP sessions.

Let's break it down.

What Is the One Million Route Problem?

Most modern routers have a maximum limit on how many IP routes they can store in their routing table. Historically, this was fine. But the global BGP routing table has been growing steadily, and as of 2025, it's well past 1 million IPv4 + IPv6 routes when you include:

• Full Internet BGP tables from IP Transit (IPT) providers

• Peering routes via Internet Exchanges (IXes)

• Internal/private routes like loopbacks, MPLS labels, SD-WAN overlays

When your router crosses this threshold, the result isn’t graceful degradation—it’s catastrophic failure. Forwarding engines choke, control planes panic, and routes start disappearing.

Who Is Affected?

Many ISPs, especially those with older routers or underpowered edge gear, have hit this wall. If your BGP speaker supports only 512K, 768K, or even just under 1M routes, it’s not a matter of if—it’s when.

This issue is exacerbated when you're:

• Receiving full routes from multiple transit providers

• Running dual-stack IPv4 and IPv6

• Injecting private infrastructure routes

• Connected to multiple IXes or large private peering meshes

What Happens When It Hits?

Your BGP sessions may remain up, but routes will silently be dropped. Traffic blackholes, routing loops, and widespread instability can occur. Customer complaints will spike, but the root cause might not be immediately visible unless you're monitoring route table size.

The Quick Fix | Drop IPT Routes & Revert to Default

If you’ve hit the ceiling, your fastest path to stability is to temporarily stop learning full routes from your IP transit providers. If the table size count provided by bilateral peering partners or Internet Exchanges are low you can still keep those. Most traffic is via these latter routes and international is typically in the region of 20 to 30%. Here’s how:

Step 1: Filter Full Routes

Create an inbound route filter for your IPT peer that drops all routes:

ip prefix-list DROP-ALL seq 5 deny 0.0.0.0/0 le 32
ip prefix-list DROP-ALL seq 10 deny ::/0 le 128

route-map IPT-DROP-IN deny 10
 match ip address prefix-list DROP-ALL

Apply this to your BGP session:

router bgp 65000
 neighbor 192.0.2.1 route-map IPT-DROP-IN in

Step 2: Add a Default Route

Negotiate with your IPT provider to send only a default route (0.0.0.0/0) or manually configure it:

ip route 0.0.0.0 0.0.0.0 192.0.2.1

This returns traffic flow while massively reducing route table size.

Why This Works

When you drop the full table and revert to a default route, your router no longer has to store or process over a million entries. You buy yourself breathing room—a critical window to prepare your long-term fix.

The Long-Term Solution | Rip and Replace

The only real fix is to replace legacy routers with devices that support:

• More than 2M IPv4/IPv6 route entries

• Adequate TCAM and FIB capacity

• Route dampening and memory scaling

Vendors now offer gear optimised for full Internet routing with ample headroom for future growth.

Wrap

The 1M route wall is not a theory. It’s happening—and it will keep happening to those clinging to outdated gear. But when it hits, don’t panic. Drop the full table, revert to default routing, and start planning your replacement strategy.

As any seasoned engineer will tell you: stability comes first. Optimisation can wait.

---