Thank you toasty answers. Policy based routing is very useful. Company may want to use dual WANs or the use of one local and remote Edge router for isp/ip redundancy.
I'd just gone through this exact subject in the last couple of months to route traffic of my Synology NAS when it backs up to an offsite NAS. My setup is exactly the same as your example where I have two ERs, each connected to their own ISP and got it working easily enough. I've now extended this so my Xbox and Playstation consoles use the second ER and ISP exclusively for their internet connection, freeing up my primary internet connection for everything else.
This video is perfect for explaining PBR especially since i tried using the UBNT guide and screwed things up and not knowing why. Now i know! Scenario i have that i want to use this for is I have one VLAN that i would like to direct through a specific ISP (Im using Load Balancing) when any device on that VLAN is requesting a specific IP. Basically takes your "N" streaming service example and says only when these devices on VLAN 10 are requesting "N" then use PBR. How would you modify the PBR to do that?
Would love to see a variation of this video, but using destination domains as the endpoints rather than known IP Addresses, for example, when live streaming I need to push the rtmps streams over my local ISP, rather than over my starlink connection, so adding a. rtmps. youtube and b. rtmps. youtube to a destination tables. Reason, too many packet drops over starlink for streaming.
I haven't really looked into this, but at first glance this seems like it would be verydifficult to implement (at least in the same way as shown in this video) since the policy is modifying the routing table. A routing table doesn't care about DNS hostnames and doesn't have a mechanism to use them (as far as I know). This is probably possible, but I'd imagine it would have to be done using a different service or application more suited for this specific use-case. Proxy servers or split-horizon DNS come to mind... but that's just me spitballing.
@@ToastyAnswers I haven't been able to get it to work, there are apparently some custom approaches using scripts to do DNS lookup and then update an address group which is then used by the firewall/routing. Way over my head....
Thank you toasty answers. Policy based routing is very useful. Company may want to use dual WANs or the use of one local and remote Edge router for isp/ip redundancy.
I'd just gone through this exact subject in the last couple of months to route traffic of my Synology NAS when it backs up to an offsite NAS.
My setup is exactly the same as your example where I have two ERs, each connected to their own ISP and got it working easily enough.
I've now extended this so my Xbox and Playstation consoles use the second ER and ISP exclusively for their internet connection, freeing up my primary internet connection for everything else.
You and your videos are so insightful. You’ve been of help to problems I have been facing in networking.
This video is perfect for explaining PBR especially since i tried using the UBNT guide and screwed things up and not knowing why. Now i know! Scenario i have that i want to use this for is I have one VLAN that i would like to direct through a specific ISP (Im using Load Balancing) when any device on that VLAN is requesting a specific IP. Basically takes your "N" streaming service example and says only when these devices on VLAN 10 are requesting "N" then use PBR. How would you modify the PBR to do that?
❤
Will be using this to fix T-Mobile
Would love to see a variation of this video, but using destination domains as the endpoints rather than known IP Addresses, for example, when live streaming I need to push the rtmps streams over my local ISP, rather than over my starlink connection, so adding a. rtmps. youtube and b. rtmps. youtube to a destination tables. Reason, too many packet drops over starlink for streaming.
I haven't really looked into this, but at first glance this seems like it would be verydifficult to implement (at least in the same way as shown in this video) since the policy is modifying the routing table. A routing table doesn't care about DNS hostnames and doesn't have a mechanism to use them (as far as I know).
This is probably possible, but I'd imagine it would have to be done using a different service or application more suited for this specific use-case. Proxy servers or split-horizon DNS come to mind... but that's just me spitballing.
@@ToastyAnswers I haven't been able to get it to work, there are apparently some custom approaches using scripts to do DNS lookup and then update an address group which is then used by the firewall/routing. Way over my head....