Very good benchmark comparison. And I am sure it took a long time to do the test, but I can't think of anything you could have left out and still had such a useful comparison. I was also surprised at the differences between the ER-X and ER-X SFP preformance in some tests, which as you said should not have been there. Perhaps if you had run a longder test, it wouldn't have been as noticable. I assume you were not using the SFP port any way. BTW, the ERPoE5 only has 5 total ports. It is essentially an ER-Lite3 with a non-vlan-aware 4 port switch attached to eth3 (and called switch0). So while there may be 7 total ports (4 on switch and 3 on router, a switch port and router port are "consumed" by the internal connection from the third router port to the switch. The thing that makes that is a plus for the ERPoE5 is that it can be powered by 48 volts, and drive 48V PoE devices, like the UAP-AC-Pro. Other than that, there are better choices. And I would expect the ER-L to perform in line with the ERPoE5, especially if only eth0 and eth1 were used in the test. The ER-X and ER-X SFP advantage that was not considered in the video was that it has a vlan-aware switch built in, and every port is part of the switch, so it is possible to use trunk ports on the WAN interface and have one vlan bypass the router, for example with some IPTV configurations. Thank you for making this. If I would have realized it was a performance comparison, I would have watched earlier. You may want to edit the title and include "performance" ER-X, ER-4, USG3, USG4. If might get a few more hits and views. Thanks again for this great video.
Nice video👍 started my ubiquiti journey with your and crosstalk solution videos. Started with a ER-lite and a smal 8 port smart switch. Now running a UDM-pro and a Cisco 3850 48 port to learn Cisco as well😊
Hello, good morning, could you create a video explaining how to put IPV6 on the edgerouter 4, and distribute it on the lan network, and then get it on a tp-link router to work
The er-4 has a quad core mips cpu at 1Ghz, I found it does about 360Mbits down 480Mbit up doing smb file transfer routing between eth1 and eth2. cpu usage is about 35%. This is with QOS smart queue enabled. The special edition Dream machine has quad core mips cpu at 1.7Ghz so I assume a 70% speed increase so 600m down/ 800m up
I did a downloading speed test from the Internet on eth0 while maxing out the download bandwidth routing between eth1 and eth2 and found the speed test and smb bandwidth are shared and capped at about 360Mbit down and 480up regardless on what port its coming from. cpu usage did not exceed 40%, I think this is because only 1 core is being used per stream. I assume 2 users may get 360 down each and cpu will get to 70%. Also if u set the speed above 350Mbit in any QOS setting QOS does not work as well as its maxing out one of the core of the cpu. Changing the types of QOS made no difference to the speeds.
Nice video. Btw you and your videos are responsible for me being really happy with my Edge Router X/4. Any chance at some point to test / do videos on the new ubiquiti routers like like UDR or UDW, to see how capable they, specially vs Edge Routers?
I don't have hands-on experience with the UDR, but from what I've read it appears to be a more limited version. I would stick with the ER4 personally. The UDR is designed to work with UISP as a "controller" of sorts. I don't believe all the same features are configurable, but I could be wrong...
I was looking to go for that certification the year it was discontinued. It is my understanding that a lot of the topics have been moved into the "Azure" certifications. Most of the concepts should still apply to on-premise applications as well.
Thanks for taking the time to do this video and analysis. Some really good nuggets and take aways.
Very good benchmark comparison. And I am sure it took a long time to do the test, but I can't think of anything you could have left out and still had such a useful comparison.
I was also surprised at the differences between the ER-X and ER-X SFP preformance in some tests, which as you said should not have been there. Perhaps if you had run a longder test, it wouldn't have been as noticable. I assume you were not using the SFP port any way.
BTW, the ERPoE5 only has 5 total ports. It is essentially an ER-Lite3 with a non-vlan-aware 4 port switch attached to eth3 (and called switch0). So while there may be 7 total ports (4 on switch and 3 on router, a switch port and router port are "consumed" by the internal connection from the third router port to the switch. The thing that makes that is a plus for the ERPoE5 is that it can be powered by 48 volts, and drive 48V PoE devices, like the UAP-AC-Pro. Other than that, there are better choices. And I would expect the ER-L to perform in line with the ERPoE5, especially if only eth0 and eth1 were used in the test.
The ER-X and ER-X SFP advantage that was not considered in the video was that it has a vlan-aware switch built in, and every port is part of the switch, so it is possible to use trunk ports on the WAN interface and have one vlan bypass the router, for example with some IPTV configurations.
Thank you for making this. If I would have realized it was a performance comparison, I would have watched earlier. You may want to edit the title and include "performance" ER-X, ER-4, USG3, USG4. If might get a few more hits and views.
Thanks again for this great video.
Found your channel today, love your content dude! Thanks for this
Nice video👍 started my ubiquiti journey with your and crosstalk solution videos. Started with a ER-lite and a smal 8 port smart switch. Now running a UDM-pro and a Cisco 3850 48 port to learn Cisco as well😊
Mactelecom Networks is good as well
Thanks very much, very good review
Hello, good morning, could you create a video explaining how to put IPV6 on the edgerouter 4, and distribute it on the lan network, and then get it on a tp-link router to work
The er-4 has a quad core mips cpu at 1Ghz, I found it does about 360Mbits down 480Mbit up doing smb file transfer routing between eth1 and eth2. cpu usage is about 35%. This is with QOS smart queue enabled. The special edition Dream machine has quad core mips cpu at 1.7Ghz so I assume a 70% speed increase so 600m down/ 800m up
I did a downloading speed test from the Internet on eth0 while maxing out the download bandwidth routing between eth1 and eth2 and found the speed test and smb bandwidth are shared and capped at about 360Mbit down and 480up regardless on what port its coming from. cpu usage did not exceed 40%, I think this is because only 1 core is being used per stream. I assume 2 users may get 360 down each and cpu will get to 70%. Also if u set the speed above 350Mbit in any QOS setting QOS does not work as well as its maxing out one of the core of the cpu. Changing the types of QOS made no difference to the speeds.
thank you
Nice video. Btw you and your videos are responsible for me being really happy with my Edge Router X/4. Any chance at some point to test / do videos on the new ubiquiti routers like like UDR or UDW, to see how capable they, specially vs Edge Routers?
Thank you!
I'm unfamiliar with the new routers (UDR/UDW). Do you have a link to these? Definitely want to check them out if they're a new release.
Any chance you have details you can share on how to run the tests? I’m interested in how you did the jitter test.
All tests were done with iperf. The only difference was the bi-directional test was done with iperf2 and all others with iperf3.
Hi , may I ask Unifi UDR is it better than ER 4 ? Thanks
I don't have hands-on experience with the UDR, but from what I've read it appears to be a more limited version. I would stick with the ER4 personally. The UDR is designed to work with UISP as a "controller" of sorts. I don't believe all the same features are configurable, but I could be wrong...
As MCSA expired, which certificate would you recommend that has the same content as MCSA?
I was looking to go for that certification the year it was discontinued. It is my understanding that a lot of the topics have been moved into the "Azure" certifications. Most of the concepts should still apply to on-premise applications as well.