Traffic Distribution methodology

Reece.Boucher — Wed, 02 Nov 2022 21:33:22 GMT

We have 35 PA firewalls all using SD-WAN and have (typically) the following configuration for WAN connections...

100M TC4 Internet

20M TC2 Internet (best quality)

5G Cellular Internet (always on)

We have the following traffic distribution profiles...

Critical Traffic - TC2, TC4, Cellular (top down priority)

Standard Traffic - TC4, TC2, Cellular (top down priority)

Critical traffic is used for voice and internal web traffic, as well as AD services (dns, kerberos, ldap). Standard traffic is used for anything else.

We have seen large amounts of traffic (approx. 400G/month) across the Cellular services. I have done further inspection and not seen any failure of TC4 or TC2 services that correspond with dates when there are large amounts of traffic shown on the Cellular equipment. Whilst the cellular service is an 'always on' (not dial on demand) the equipment provider (Cradlepoint) indicated keepalive/management traffic should be well les than 1G per day.

As noted, the priority is a 'top down' configuration. The best path for traffic is not necessarily the highest bandwidth.

Q: What constitutes going to the next path in the selection order?

Q: Is top down the best approach?

Q: If not, what would the best approach be to ensure only fixed line services are used unless there is a total failure of both fixed line services?

Q: Should the above be used in conjunction with either/or/and Path Quality and SaaS Quality profiles?

topic Traffic Distribution methodology in Next-Generation Firewall Discussions

Traffic Distribution methodology