Adjusting OSPF Reference Bandwidth for Accurate Path Costs

OSPF auto-cost reference-bandwidth: Design, Verification & Best Practices

In any network running OSPF (Open Shortest Path First), understanding how routing decisions are made is critical. OSPF selects the best path based on interface cost, which is derived from bandwidth. However, as link speeds increase, the default OSPF reference bandwidth quickly becomes insufficient.

This is where the auto-cost reference-bandwidth command becomes essential.

---

What Is OSPF Cost?

OSPF calculates interface cost using the formula:

Cost = Reference Bandwidth / Interface Bandwidth

By default, the reference bandwidth is 100 Mbps. While this was suitable in Fast Ethernet environments, it causes all modern high-speed links (1G, 10G, 40G, 100G) to appear identical in cost.

---

Why Adjust Reference Bandwidth?

Without adjusting the reference bandwidth:

• 1 Gbps and 10 Gbps links both calculate to cost 1
• OSPF cannot prefer faster paths
• Traffic engineering becomes impossible

To correct this, configure:

Router(config)# router ospf 87
Router(config-router)# auto-cost reference-bandwidth 1000

This sets the reference bandwidth to 1000 Mbps, allowing OSPF to distinguish between FastEthernet, Gigabit, and higher-speed links.

---

Verifying Interface Cost

Router# show ip ospf interface GigabitEthernet0/0

GigabitEthernet0/0 is up, line protocol is up
  Internet Address 10.1.1.1/24, Area 0
  Process ID 87, Router ID 1.1.1.1
  Cost: 1
  State DR, Priority 1

After Updating Reference Bandwidth

Router(config)# router ospf 87
Router(config-router)# auto-cost reference-bandwidth 10000

Router# show ip ospf interface GigabitEthernet0/0
  Cost: 10

This confirms that OSPF now differentiates link speeds accurately.

---

IPv6 Version (OSPFv3)

In IPv6 networks, OSPFv3 uses the same cost calculation logic. The configuration is identical:

Router(config)# router ospfv3 10
Router(config-router)# auto-cost reference-bandwidth 10000

Important: OSPFv2 and OSPFv3 maintain separate processes. Reference bandwidth must be configured independently for IPv4 and IPv6.

---

OSPF vs EIGRP Cost Comparison

Aspect	OSPF	EIGRP
Metric Type	Cost (Bandwidth-based)	Composite (Bandwidth + Delay)
Default Bandwidth Reference	100 Mbps	Based on interface BW
Tuning Method	auto-cost reference-bandwidth	bandwidth / delay / variance
Granularity	Moderate	High
Vendor Support	Open standard	Cisco-centric

---

Pitfalls & Real-World Case Studies

Pitfall 1: Inconsistent Reference Bandwidth

If routers in the same OSPF domain use different reference bandwidth values, they may calculate different costs for the same path. This can cause:

• Suboptimal routing
• Asymmetric traffic
• Routing loops in extreme cases

Pitfall 2: Ignoring Interface Bandwidth Command

OSPF relies on the configured interface bandwidth. If the bandwidth command is not accurately set, cost calculations will be incorrect—even with the right reference bandwidth.

Real-World Case Study

A data center migration introduced 10G uplinks, but the reference bandwidth remained at 100 Mbps. Traffic continued to traverse legacy 1G links, causing congestion and packet loss until the reference bandwidth was corrected network-wide.

---

Further Reading

For a deeper understanding of OSPF architecture and behavior, visit the OSPF Wikipedia page .

---

Final Thoughts

The auto-cost reference-bandwidth command is not optional in modern networks—it is foundational. Without it, OSPF cannot make intelligent decisions in high-speed environments. Proper planning, consistent deployment, and verification ensure optimal and predictable routing.