The Fusix network is designed to provide a guaranteed 100% Internet service availability.
The services are delivered from the Autonomous Systems Number (AS) 57866 on a footprint of data centers in The Netherlands and the US. In both countries, we connect to tier-1 upstreams (some of which we have configured to use only as “private peers”, not for transit), strategic private peering partners and some Internet exchanges. Our network provides non-overbooked, high performance access to the Internet and to main cloud services, such as AWS, Microsoft Azure and Google Cloud.
The routing policy on our network is to offer to our customers the best routes that we have (not the cheapest). In addition, we go for stability: we prefer setting up a dedicated cross connect to a tier-1 provider and only use part of their routes, over setting up a connection to an Internet exchange where hundreds of members may or may not announce you routes that they may or may not be allowed to announce. We do not peer with everyone who wants to peer with us, but have a selective policy and for stability reasons may choose to send traffic to other parties over a link that we need to pay for, instead of one where we could send the traffic for free.
IPv4 and IPv6
The Fusix backbone supports IPv6 and IPv4 natively and offers Internet services to both IPv4 and IPv6 networks using BGP (Border Gateway Protocol) or default routes.
Network Operation Center
Our network is managed by the Fusix Network Operation Center in The Netherlands, using a management platform that is deployed redundantly: one inside our own backbone and one independent from us. Most of the network configuration tasks – such as IP address prefix management – are automated, which allows the NOC staff to deliver and maintain customer-connections fast and without human errors.
Network supervision and monitoring are handled by consolidated open source diagnostic tools that produce fault reporting alarms and monitor physical connectivity and routing behavior for each customer connection.
Full redundancy of backbone components
Our on-net locations are connected to the core routers in a form of a star, while all backbone components are deployed redundantly. This design provides 100% service availability and dramatically reduces the possibility of customer impact during network maintenance events or unplanned outages.
Each on-net data center is equipped with 2 physical Arista switches, each of which is connected to a core router. IP transit service is always delivered on two ports while VLL services are delivered to customers on 1 port per location. For VLL service, in order to achieve redundancy (or “protected VLL”), we can deliver two circuits that run totally independently of each other.
Within each pair of dark fibers, the two lines follow a different path regarding each other, which eliminates a risk of accidental damage of both lines at the same time.
The core routers are Juniper MX960 devices that are located – for service continuity reasons – in 2 separate data centers: Nikhef and euNetworks.
The result of the backbone design is the best service availability rate in the market.