Two means one, and one means none. Always carry two pens to a meeting. And when it comes to IT, every engineer and all managers know the meaning of the word ‘redundant’. Having your redundancy sorted out means that part of the system can fail or undergo planned maintenance without the system as a whole being affected.

Redundancy can be achieved at many levels, resulting in differences in uptime guarantee and network design. Some services lend themselves really well to being configured redundantly (such as BGP uplinks) and others don’t (such as layer 2 connections). In the end, having your redundancy well set up will give complete freedom to yourself and to your provider, since both know they can perform necessary work – or cope with unplanned downtime – at any moment.

I think redundancy is not something that an ISP should charge extra for: it gives flexibility to both customer and the ISP and is therefore advantageous to both.

At Fusix Networks, we have a bit of a special take on redundancy. We feel that we have to deploy it on the infrastructure level up to and including our customer facing devices and not just make the customer think that we have done our job right. This means we take the redundancy also as part of our interconnections: that’s right, plural, because with a single interconnect the service is by definition non-redundant. Think about it: if you order a single cross connect to your ISP, it connects only to one device in your provider’s network. If that device is down (for instance, owing to a software upgrade), your service is down, even if your supplier has redundant connections from the device to the rest of his network.

For layer 2 connections, providing redundancy becomes even more interesting. Let’s say you are a Fusix customer in two different data centers and you take a Virtual Leased Line service from us between them. In our experience the best way of providing redundancy for these is to actually provide two separate VLLs and sort out the redundancy on layer 3 in the customer’s network.

One potential downside of our approach could be that one of the two VLLs goes  down in case of planned maintenance. But this will not become a problem, because both VLLs are active, so all your traffic is handled by the other VLL without any action required. In fact, every planned maintenance proves again and again that the service we provide to customers is truly redundant! So now you know why at Fusix we always offer two ports as part of your project. And of course we help you set up a proper, truly redundant connection with us across those two ports, whether you’re looking at a Direct Internet Access, BGP uplink or at VLL service between data centers. Your two ports will always be on two different devices on our side, connecting to our two different routers via two different fibers on our side. Guaranteed.

But doesn’t that mean that you’d have double the cross connect cost? Not necessarily. We’ll gladly use single mode fiber cross connects with bidirectional optics with you, connecting our two ports over one physical patch. This means that there is absolutely no downside to our way of providing redundancy to you – a technically superior solution for the same price. Who can object to eating the cake and still having it?