Building a high performance, highly redundant network requires both smart engineering and great hardware. Taking a focused look on our next-generation NY/NJ network allows us to demonstrate how this is done. To start, our network is built around a pair of fully-redundant Juniper MX960 routers. Each MX960 has a fully-redundant fabric, routing engines, and power supplies. These routers are connected to our fully-redundant and self-healing fiber optic network, which links our Weehawken facility to points of presence (POPs) in both New York City and Newark, NJ. Our fiber optic network has the intelligence to automatically failover in under 10 microseconds in the event of a fiber cut, or failed optical transceivers. Having the best optical equipment on a poorly thought out dark fiber path would be pointless -- so we therefore have 4X redundancy -- 3 diverse paths to NYC, and one diverse path to Newark, NJ -- with each path having enough capacity to fully serve 100% of customer traffic. This redundancy eliminates any single point of failure.
Adding to these capabilities, our MX960 routers provide advanced features allowing us to absorb very large DoS attacks, and also guarantee a high Quality of Service (QoS) for premium customers.
In addition, we have advanced in-house tools to monitor our connectivity to end-user locations, allowing us to route traffic to the optimal path, helping to ensure that end-users receive the highest possible performance.
Finally, our network is monitored 24x7 by our staff, using advanced tools to ensure rapid response to failures on our peers networks.