Linksys EA8300 Review

Testing

Now for the real meat of the review. The testing. We have an idea of the internals, we know roughly what the tech is capable of in a lab, but how does this contend in the real world?

At this stage, I’m avoiding the time-consuming uptime test, purely since at the end of the day, I have seen no stability issues caused by this device. Any instability I have had will have be down to the copper line from my house to the local PCP in the Fibre cabinet at the end of my street and since the EA8300 lacks a VDSL modem, it’s obvious this is not a factor here. So, all this leaves is raw throughput, as ever I’ll be testing in a few different rooms around the house, to give na idea of what other users can expect. Of course, everyone’s circumstance, setup and home is different, so your mileage may vary. One thing ot bear in mind is the construction of the house is a major factor in terms of wireless performance. Sadly, having recently moved to a 1930s build, my current residence is mostly solid brick walls throughout, which are something of a significant barrier to wireless networks. As you will see from some of the test results, there are certain parts of my house that significantly struggle when using WiFi.

As a baseline I took a quick test of my current WRT3200 ACM router in one of the upstairs rooms to the rear of the house – this shows just how poor the WiFi signal can be in some locations, as an idea the router is located on the ground floor to the front of the house – this room also has a chimney breast on the wall between it and the router, so there is a lot to contend with.

As you can see I average around the 14.1Mbits/sec mark With the odd spike in excess of 20Mbits and even a single spike at above 40Mbits. If the EA8300 can outperform this, then it’s just proof of how much technology like Band Steering can help with problematic networks.

Since we’re already looking at data for the back bedroom, let’s use that as the first example from the EA8300. Right off the bat, we see Band Steering can really help with problematic network locations, with speeds briefly topping out around 120Mbits/Sec over and averaging around the low 60Mbit/Sec when testing TCP and generally sitting around the 90Mbits/Sec mark on UDP with the odd spike into the low 100Mbit/Sec mark, while I wouldn’t call either of the speed tests stable in terms of throughput, the fact we see any improvement at all is quite an important fact to consider. Compare that to a single 5Ghz channel, acting in isolation on the same router and while we see similar performance to the WRT3200ACM, there are a few more spikes into the 40Mbits/Sec range – this could either be due to improved antenna over the 3200, or possibly moments with less local interference in and around the property. Evidently, the EA8300 is at a significant advantage here thanks to the fact that it can switch between the two 5Ghz networks when Band Steering is enabled, allowing it to provide better connectivity overall. That’s not to say the speeds are amazing, by any stretch, but I can forgive a lot here due to the general solid brick construction of my property having a significant impact on wireless signals.

Looking at other rooms around the house we see a very similar story. When using the 5Ghz channel using Band Steering we see generally improved results throughout. What is interesting is just how significant a difference this appears to generate. It may be down to time of day differences such as something interfering locally being turned off between the tests being performed with Band Steering disabled and enabled, or it could be that there is something else happening behind the scenes on the device. Either way, using the router without Band Steering shows results far below what I would normally expect with a 5Ghz device, while with Band Steering enabled we see overall higher throughput, but also more volatility with huge spikes in performance periodically. The one room that should be particularly interesting of all of them however is the Games Room, since this is where the router was located when testing. Single channel performance while better is showing significantly lower throughput than I would expect of a 5Ghz network, while Band Steering is allowing it to hit highs close to peak performance, these aren’t sustained and throughput is quite choppy at best. This suggests that there is more going on with how Linksys are implementing band steering, since throughput on both single channel and multi channel should be almost identical in this room.

This slideshow requires JavaScript.

This slideshow requires JavaScript.

Comparatively the 2.4Ghz network appears to perform at a similar rate to the single none Band Steered 5Ghz network. However, the drop is throughput is significantly less on the 2.4Ghz side of things and for a very good reason. As mentioned I have a house constructed in the mid 1930s, which means its mostly brick, both internally and externally. 5Ghz is great for throughput, but often struggles when their are barriers in the way. In a home like mine 2.4Ghz does actually have a lot going for it, since 2.4Ghz has a longer wave than 5Ghz, sadly the 2.4Ghz range can be quite a crowded place, which means that it is something of a trade off. Again, the Games Room is an oddity due to performance not seeming to be much different to any other room in the house, suggesting that the 2.4Ghz network isn’t performing to expectations even in reasonable conditions.

This slideshow requires JavaScript.

This slideshow requires JavaScript.

MU-MIMO

The final test comes in the form of seeing how one of the headline features of the device performs. MU-MIMO, for this test I disabled Band Steering and configured the device to work in both MU-MIMO and SU-MIMO modes. Ideally, we should see an increase in performance for both clients in MU-MIMO vs SU-MIMO, since MU-MIMO allows both client to transmit and receive data in parallel, rather than in serial as is the norm in SU-MIMO setups.

Unfortunately, due to a lack of portable devices that fully support MU-MIMO, I had to resort to using a MacBook Pro and my Windows desktop, which means less than ideal conditions since the desktop PC is in my office and the router is on the floor below. Interestingly this may have also shed light on the speeds seen when testing a single 5Ghz stream earlier.

What we can see is that when MU-MIMO is enabled their is a general increase, to the point of almost double in most cases vs the SU-MIMO solution. However, as mentioned this does provide some potential answers to the overall throughput and performance seen in other tests. The PC is using a Linksys Provided WUSB6100M, while the MacBook Pro is using its inbuilt Wireless card, as Linksys do not currently have an OSX Driver for the WUSB6100M. All tests on the network have been carried out on the MacBook Pro, it being the only portable device I have to hand. As you can see both the SU-MIMO and the MU-MIMO tests show the same choppiness for the Mac as seen elsewhere as well as slightly lower levels of performance than the PC, using the Linksys Dongle.

Further Testing

Due to the odd results presented by the MacBook Pro which was used for testing, I decided to run a single final battery of tests in the same room I used for MU-MIMO testing, just to see how things looked on a second device. AGain, this is using my desktop PC, with a Linksys WUSB6100 USB 3.0 Wireless Dongle for WiFi access.

First up the PC’s test results.

And now those from the Mac, for both the 2.4Ghz and 5Ghz channels.

As you can see, the 2.4Ghz tests show that the Mac is actually under performing here, since not only do the speeds top out around the mid 60Mbit/Sec Mark, but they are pretty jittery at best. Similarly, we are seeing better performance on the 5Ghz channels across the board, both for band steering and single stream, again with a lot less jitter than is seen on the MacBook Pro. Evidently, either this particular laptop has issues with its WiFi card, or OSX and their Laptops have some oddities regarding WiFi, with only the one device to test I will not attempt to draw any conclusions beyond this one device though.

Fortunately, while the results above do cast some shade on the testing done on the Mac, they do  show the router to be performing better than I first thought. Though still not quite as well as I would expect it to.

Conclusion

So, after all of this what can we determine from my tests? Firstly that their is a possible issue with the MacBook Pro that was used in testing, which while not impacting its day to day use, certainly makes it unusable for testing WiFi networks. Secondly, despite the issue seen on the Mac, the evidence is there that houses built at the start of the 1900s and earlier in the U.K are certainly a detriment to the use of Wireless devices if you do not use additional hardware such as wireless repeaters and similar to help work around solid brick. Of course, despite their being some questions surrounding the results from the Mac, it does give us some insight as to how the EA8300 can perform in a home such as mine. We know that as distance and internal building materials are placed between the client and the router, that while the speeds do suffer, the router does its damnedest to try to keep you not only connected, but running at a reasonable speed. Indeed, despite the issues with the MacBook Pro, the router did seem to perform reasonably admirably in terms of speeds seen, since from the comparison tests we can clearly see that not only is the MacBook Pro more jittery than the PC was, but that also speeds are fundamentally and significantly lower, so its arguable that if I were able and willing to test my desktop PC solution in each room in the house I would see similar outcomes to comparative tests.

This said, I do have some concerns regarding the overall performance of the device, since when tested on my PC things weren’t quite upto the levels of speed that can be expected of either a 2.4Ghz or a 5Ghz network. Of course, this was tested in none ideal conditions, so some benefit of the doubt is needed here.

MU-MIMO pretty much works as expected, with the overall throughput seen being roughly doubled over a SU-MIMO based solution when testing with multiple devices also, so of course a home that has several wireless devices will see significant benefit for going with a MU-MIMO based solution.