Choosing an IP transit provider: the criteria that really matter

Choosing an IP transit provider on price per Mbps alone is like choosing a car on price per kilo. Two offers at the same rate can deliver radically different experiences. Here are the criteria that really make the difference.

Capacity — and real capacity

A provider can sell you 10 Gbps, but what happens at peak hours, or when one of its links fails? Find out the network's aggregate capacity, its usual utilisation, and its headroom. An over-provisioned network absorbs spikes without congestion; a network near saturation will degrade your performance at the worst moment.

Route quality

Not all traffic is equal: what matters is the quality of the paths to your destinations and your users. A provider rich in peering and present at IXPs offers shorter, more stable routes. Ask for its points of presence and IXPs, and test from its Looking Glass if possible.

Redundancy

What happens if a link, a router or a site fails? Good transit relies on multiple paths and a design with no single point of failure. Multihoming on the customer side (several providers) remains the best protection, but each provider's internal resilience matters just as much.

SLA and support

• SLA: what guaranteed availability, what penalties, what definition of an incident?
• Support: is there a reachable 24/7 NOC, with clear response times?
• DDoS protection: is it natively included, or billed as an option once an attack is under way?
• Provisioning: how quickly is a BGP session brought up, including over GRE?

The right reflex: look beyond price

Transit that's slightly more expensive but well designed, well connected and well supported will cost you less in downtime and incidents. At Connect-IX, our IP transit combines multi-Tbps capacity, rich peering, native DDoS protection and a NOC — all observable in real time via our Looking Glass and our status page.