Hi there,
we are deploying several max transit duo over the planet, all of them allowing us to bond 4G connections. We have a VM at AWS running a 500 licences setup.
The fact is we are facing something we didn’t anticipate : latency matters !
That deployment is allowing us to do premium autonomous video remote solutions in which latency needs are huge.
For now our VM is based on a French server @ amazon web services. Obviously, when we connect from Melbourne in Australia to start a session, we are facing a latency of 350°ms over the standard latency.
Any luck we could set up several fusion hubs over the planet, use the gps built in feature to select the good fusion hub ?
Quite challenging but hoping somehow I could find a way to get that sorted on this forum
Cheers
We’ve done this a couple of times now for customers with moving devices. The first was for cruise ships where we used geofencing and the link James provided is spot on for that.
The 2nd was for man portable units for mobile teams that travelled the world (normally to destinations where there was some sort of trouble that needed - fixing). In this instance there were two requirements.
Overt teams just wanted secure lowest latency ‘corporate’ connectivity so we used GeoDNS to help their MAX BR1s find the geographically closest fusionhub (GPS tracking was not enabled or encouraged).
In contrast, the covert teams didn’t want the host country they were currently in to see traffic from their devices flow back and forth from the UK. Instead we supplied SIMs from another (but geographically close) country entirely and enabled roaming, we’d build a tunnel from the Transits back to a Fusionhub hosted in that neutral - relatively near country, then we would route that traffic from that target fusionhub over another tunnel to a UK based balance.
Whats of note here is that when a device is using a SIM that is roaming - or indeed when you are using a ‘global SIM’ from an operator, traffic from that device can often be tunnelled / routed by the SIM operator back to the home country for billing and authentication purposes. When this happens, you can be in country A, with a Fusionhub in country A, but your traffic actually flows from country A to country B where it breaks out to the internet only for you to then have to route it back to the Fusionhub in country A again (doubling the latency).
Take very good note of traceroute results as you look to deploy geographically and keep an eye out for anomalies like that.
