Are you considering redundancy to protect your network for a fiber cut or local network failure? While using an internet VPN as a secondary network is a common backup solution (using BGP), the right SD-WAN solution can provide a more seamless fail-over in less than a second. Some SD-WAN technologies will even maintain a VoIP or video session with absolutely no loss of contact, pretty valuable if you route customer calls over the network. You should consider using LTE as a solid protection against the catastrophic fiber cut that might take out all your connections. If a facility does not have two MPOEs, LTE is a no-brainer. Even if you do have two points of entry, LTE for fail-over is a bargain, if it supports your bandwidth requirements, which is dependent on reception conditions.
BGP is a common routing protocol for WAN survivability. For mission critical network applications, the availability and performance of applications vital. When it comes to network related issues, BGP, which is used for end-to-end routing, was not designed to consider certain metrics related to performance of applications when selecting best paths. This is exactly where PFR comes into the picture.
By doing so, PFR allows network administrators to minimize bandwidth cost, do intelligent load distribution, and most importantly, improve the availability of applications in their network. This makes PFR superior to BGP. Cisco provides PFR at no charge as part of its IOS.
OER or optimized edge routing is the predecessor to PFR. PFR encompasses all the OER features, but also PFR greatly expands application intelligence and has a broader technology scope than OER. It has many more advanced features and has extensive reporting capabilities through CLI, Syslog, SNMP, and Netflow.
Looking at the network here, notice that the traditional routing, in this case, BGP, chooses the path with the least number of AS hops as the best path. But if there is a condition in that path that hurts network performance, such as some packets being dropped because of a bad interface, it is highly unlikely for BGP to detect the condition and reroute the traffic. Even if BGP does reroute the traffic, it is likely that the protocol timers are not set aggressively which means there will be a significant performance degradation from the perspective of the application end user. This is exactly the type of problem that PFR was designed to address. It can detect black-outs or soft errors very quickly in the end-to-end path and force path changes to route around the fault. This is how PFR guarantees application availability and performance. In fact, this is just one of the typical problems that PFR can help solve. If you use Cisco routers, you should consider PFR as an alternative to BGP.