Building route diversity against subsea cable cuts

Submarine cables are cut more often than people expect: ship anchors, fishing gear, dredging and seabed movement all cause faults. Repairs need a specialised ship and can take days or weeks. Terrestrial routes fail for more ordinary reasons: road works, fibre cuts and upstream carrier outages. Neither is rare enough to ignore.

1. 01Physical path diversity. Two routes that never share a cable, a duct or a landing station. Two circuits on the same cable are not diverse.
2. 02System diversity. Different submarine systems (SEA-ME-WE 4 and SEA-ME-WE 5, for example) plus terrestrial routes across the land border.
3. 03Content diversity. Direct interconnects and local caches with the largest content networks mean much of the traffic users care about does not depend on the long-haul path at all.

When a route fails, routing protocols shift traffic to the surviving paths within seconds. Capacity is reduced, so the design question is how much of your peak traffic the remaining routes can carry. A gateway that plans for that, and tells you how it plans, is the one to buy from.

Mango interconnects directly with Akamai, Amazon, DE-CIX, Equinix, Meta, Google and Zenlayer, so during an international incident the platforms most subscribers use are the least affected part of the network.

• The named systems and routes your capacity uses.
• Whether your circuits are on physically separate paths, not just separate line items.
• How rerouting is handled and how you are informed during an incident.
• For critical sites, a backup circuit through a different route as standard.