Redundant Ethernet MPLS VPN

What redundancy against network outage do I need?

What happens if a line or even the central node of my Ethernet MPLS VPN goes down? What is the maximum time I can be offline before my company suffers damage? Admittedly, there are companies in which this question plays a somewhat subordinate role. The reasons for this may lie in the budget or the fact that you can be offline for up to a day without suffering any damage. But what do companies do that are obliged to be almost permanently available or are subject to further industry-specific requirements?

Depending on requirements or conditions, a company must therefore address the issue of redundancy. This applies both to data transfer and, in the age of “All IP“, to telephony. The following overview will help you to quickly find the right option for you. You are probably familiar with some of the terms carriers use these days, such as gold, silver, protected or simply backup. These terms describe the type of backup or SLA. However, as these are carrier-specific terms, we will dispense with them here and reduce them to what they actually mean. These options are available.

Media redundancy

It is best to connect your locations via different line media. If, for example, the fiber optic network fails due to a fiber break, your redundantly connected location will continue to be supplied via copper backup. The following media are available:

• Copper: Here the connection is available via xDSL and Ethernet up to 10 Mbit/s
• Fiber optics: Offers one of the widest ranges of bandwidth with up to 10 Gbit/s. Fiber optic is either already available or can be developed.
• LTE/Dual LTE: Here, too, there are fluctuations that are caused by the network operator’s expansion, but can also be found in the building shielding. With ideal availability, an upload of up to 12 Mbit/s can be achieved here.
• WLL (Wireless Local Loop) or simply radio relay: Depending on availability, this is ideal for bandwidths from 10 Mbit/s.

Router redundancy

If you already secure one data line with a second data line on a router, you have a line backup. Unfortunately, however, you are not protected if this router fails, even if both lines are synchronized and active. This router redundancy results from the fact that you terminate each connection with a router, whereby both routers talk to each other via your LAN with HSRP and one automatically takes over the traffic should the other router fail or need to be replaced.

Separate house supply

This type of backup provides security in the event that the primary cable route is cut, for example due to civil engineering work. The probability of a simultaneous interruption on two separate sides of a building is very low. As a rule, the easiest way to achieve a separate building supply is to order your services on two different media, glass and copper, for example.

However, you will achieve higher availability if you have your building connected with fiber optics from two sides, if this has not already been done, and your services run node and edge disjoint. This means that these lines will not physically meet.

Carrier redundancy

Carrier redundancy is designed to protect you in the event that your primary network provider has a problem, such as a backbone failure. You can request your provider to provide your primary network with its own fiber and the backup via another provider (but only billed via one). The simplest example in this case would be to set up the primary network with your own optical fiber and the backup network with copper leased line redundantly via a third-party provider. This means that, in addition to carrier redundancy, you also have redundancy of the medium and the house feed. The medium can be either fiber optic or radio relay, which only has a positive effect on the availability of your network. Probably the best form of real backup is to secure a wired network with radio relay from a second provider. In this case, you are actually operating two networks actively, on two media, transferred to different routers, routed in two carrier backbones and have “no single point of failure”, as the saying goes.

Costs

The costs for your backup are heavily dependent on the following factors: Bandwidth and runtime, the availability of the solution and the effort required to implement it. Depending on the requirements, the backup can be a fraction of the cost of your primary network or even the same. It is also important to decide how high the fall height may be, i.e. how high the lowest bandwidth must or may be in order to maintain operations. Maintaining a backup just for the sake of it, which is of no use in the event of a fault or cannot be used at all, only generates costs.