How do I choose between different wholesale internet transit providers for carrier procurement and route diversity needs?

When choosing between wholesale internet transit providers, think of it as balancing cost, performance, resilience, and strategic diversity. For carrier procurement and route diversity, the “best” provider is often not the cheapest one—it’s the one that reduces correlated failure and gives you usable alternate paths when things go wrong.

1) Start with your goals

Be clear about what you’re optimizing for:

• Cost per Mbps/Gbps
• Geographic reach
• Low latency / good peering
• High availability / SLA quality
• ASN and physical path diversity
• Ability to scale quickly
• Contract flexibility

If route diversity is important, don’t just buy from two providers with different logos. You want different backbone ASNs, different upstream mixes, and ideally different physical entry points and metro paths.

2) Evaluate diversity in layers

True diversity comes from multiple dimensions:

A. ASN / backbone diversity

• Prefer providers with different global ASNs
• Ask whether they rely on the same underlying upstreams or carriers for transport
• Check whether they’re reselling capacity from the same parent network

B. Physical path diversity

• Different meet-me rooms, diverse building entrances, separate ducts, and separate last-mile carriers
• If both circuits come through the same conduit or meet the same edge router, diversity is limited

C. Geographic diversity

• Different POPs, metro areas, and ideally different regions
• A provider with strong presence in your main site plus an alternate region can be useful for disaster recovery

D. Peering diversity

• Providers with strong peering at major IXPs and good transit mix can improve reachability and performance
• But if your goal is backup/continuity, avoid two carriers that both depend heavily on the same transit ecosystem

E. Operational diversity

• Different NOCs, different provisioning processes, different maintenance windows, different escalation paths
• This matters more than many people realize during outages

3) Compare provider performance, not just brochures

Ask for and test:

• Latency to your key customer regions
• Jitter and loss
• Route consistency
• Congestion during peak times
• BGP convergence behavior
• IPv6 support
• DDoS handling and traffic engineering capabilities

Useful due diligence:

• Trial circuits or short-term commits
• Looking-glass/traceroute tests from multiple regions
• Historical outage/public incident analysis
• Community reputation and peeringdb presence

4) Look at commercial terms

Transit is a network service, but procurement is also a contract problem.

Key contract items:

• Commit levels and burst pricing
• Term length
• SLA credits
• Mean time to repair commitments
• Install interval
• Early termination clauses
• Price protection / escalators
• Port upgrades and cross-connect charges
• DDoS and remote-hands fees
• Support response times

A provider with slightly higher bandwidth cost may still be cheaper overall if they have:

• faster install
• lower cross-connect fees
• better SLA support
• fewer hidden costs

5) Design for failure domains

For route diversity, architect your procurement around failure domains:

Good pattern

• Transit A: major Tier 1/large backbone
• Transit B: different backbone, different metro handoff, different last-mile
• Optional Transit C: regional carrier or internet exchange participant for extra diversity

Better still

• Use multi-homing with BGP
• Set policy so traffic can fail over automatically
• Verify backup paths actually work under load, not just in theory

6) Don’t ignore route asymmetry

You may have one outbound path and another inbound path, especially on the public internet. Check:

• Where inbound traffic actually enters
• Whether both providers are popular with your customers or peers
• Whether traffic engineering can force better failover behavior

A provider can look diverse on paper but still share major return-path dependencies with your other carrier.

7) Ask the right procurement questions

When issuing an RFP or comparing quotes, ask:

1. What is your ASN and who are your major upstreams/peers?
2. What is the physical path from your POP to my building?
3. Can you provide diverse handoffs?
4. What are your SLA and repair commitments?
5. What is your install lead time?
6. What is your capacity upgrade path?
7. Do you offer remote peering, IX access, or private interconnects?
8. How do you handle maintenance notifications?
9. What is your policy for route leaks / filtering / RPKI / IRR?
10. Can you support BGP communities for traffic engineering?

8) Common selection strategy

A practical way to choose:

• Primary carrier: best mix of performance, reliability, and cost
• Secondary carrier: chosen specifically for diversity, not just price
• Tertiary/backup: smaller or regional provider for resilience or niche geography

This avoids the trap of “two cheap providers” that fail in the same way.

9) Red flags

Be cautious if:

• The provider can’t explain physical diversity
• Their “diverse” service still enters via the same building path
• They are effectively a reseller of the same network
• They don’t publish or discuss peering/transit policy
• Their SLA is weak or filled with exclusions
• Install times are vague
• Support is opaque or hard to reach

10) Simple decision matrix

Score each provider 1–5 on:

• Price
• Latency
• Reliability/SLA
• Physical diversity
• ASN/backbone diversity
• Geographic reach
• Support quality
• Contract flexibility

Then weight based on your needs:

• If this is production critical, weigh reliability and diversity more than price
• If this is burstable/secondary, you may accept lower performance for lower cost

---

If you want, I can also give you:

1. a procurement checklist/RFP template, or
2. a scoring matrix you can use to compare transit quotes side by side.