I have a client with geographically distant call centers. We need to ensure reliable inter-site connectivity for 1. VoIP and 2. virtual desktops. Our sites are connected via MPLS, but this bandwidth is limited due to cost, and its reliability is not significantly better for us than much cheaper “business-grade” internet connections. So, we supplement the MPLS with additional ISPs at each site. We are looking at SpeedFusion VPN to make best combined use of these connections.
I would like to know how SpeedFusion VPN handles each of the following cases. We see these frequently with both fiber and coaxial cable ISPs.
Intermittent packet loss (let’s say 10-15%) on a given WAN connection. It is not totally down, but lossy enough to disrupt VoIP traffic. Can Peplink’s health checks detect this and disable the interface until packet loss returns to near 0%?
A WAN connection is “flapping”, that is, rapidly changing state between up and down (every few seconds/minutes). In this case, we want to suppress the interface until it stops flapping and stays up for a while. Cisco has IP event dampening, which uses an exponential decay algorithm that suppresses an interface if it changes state too frequently. What is Peplink’s solution to this problem?
To state these questions a different way: It sounds like a health check consists of one ICMP echo packet. Does SpeedFusion technology consider the results of many consecutive health checks when deciding which interfaces to use? If so, what are those decision parameters, and can we modify/tweak them?
Also, what is the historical uptime of Peplink’s DNS service used for inbound load balancing? Does Peplink publish outage reports/statistics anywhere?