I recently set up Speedfusion with my Transit Duo. I would like to understand the best method to evaluate results. From my simple speedtest the latency speeds increased substantially (30 ms to 100ms) and the speeds do not appear to have increased and may have decreased. I have an at&t sim and a verizon sim installed in the transit duo. Can someone help me with how to evaluate the results and determine if this is operating efficiency is worth it in my application. I am using Vultr VPN to administer the fusionhub.
I have the same setup. I’ve learned to reduce latency by:
- Choose the Vultr server location close to the area where you’ll most often have your Max Transit.
- In Speedtest, make sure that it is choosing a server close to where your Vultr server is.
- For Windows Speedtest, go into Location Services and turn it off for the Speedtest app. This will allow the app to determine location based on the Vultr server IP, and the list of servers that are used will be based on distance from the Vultr server rather than your Speedtest app host. At least you’ll get more accurate latency results.
Now for whether Speedfusion is worth it… I’ve determined that local load-balancing is much faster for most things but not single-stream requests like streaming, mostly better for multiple clients connected to your Peplink. I think Speedfusion’s value is in having the ‘always available’ connection but I’m not a user that really needs this kind of service. Having said that, I generally leave it on because the bandwidth is good enough for my streaming needs and I have the extra reliability and encryption that Speedfusion provides.
Also, the latency figures that you see in the Max Transit UI are in part determined by your local cellular connections to the towers. I have high’er latency on my AT&T connection but super bandwidth. This makes it difficult to know the impact of local connection latency on the Speedfusion Speedtest latency. When I remove AT&T as a WAN, then run Speedtest, using the same server, it doesn’t seem to matter. The latency between Speedtest server and Vultr server changes second by second based on lots of factors. Bottom line, anecdotal experience of how your apps are running seems to be the best indication - too many other factors for me.
There is a LOT more to learn about Speedfusion before making any solid decisions. This is a really good whitepaper that I found through this forum. It explains the impact of Speedfusion on bandwidth very clearly.
Just as an aside, if you really want to see one for one comparison between Speedfusion and local WAN management, only connect ONE WAN device, say AT&T, then run Speedtest with Speedfusion on then off. You are reducing the number of factors and will l likely see about a 19% overhead with Speedfusion on.