I got an email from a reader asking me about using mesh WiFi versus having a single router that supposedly will cover a large area. The reason for this person’s question is that they are moving into a large property of about 3000 square feet and their concerns are two fold:
- Will a single router cover that area?
- Will a mesh router slow down my Internet access?
The first question can be answered this way. Yes a single router may in theory cover the property. But there may be spots on the property that it may not quite reach. And it may have to deal with walls and other people’s WiFi that would make that option less than ideal. That makes the mesh router a better option as for the areas that you can’t quite reach, you can stick a properly placed mesh node in the area to improve your coverage.
Now let’s go to the second question. Will a mesh router slow down their Internet acces?. I dug into this a bit more and discovered that they were afraid that a mesh node (let’s call this the primary node) connected to the modem from their ISP would have faster speeds than a mesh node (let’s call this the secondary node) that is connected to the primary node wirelessly. Now that’s an interesting question. So I decided to conduct an experiment seeing as I currently own a ASUS ZenWiFi AX XT8 which is a mesh router. The experiment was to do speed tests at each node, and note how much slower that it did or didn’t get along with anything else that I noticed. And for bonus points, I would also do a speed test from the far end of my condo, but in range of my secondary node to see what the speed was. Granted I am dealing with a sub 1000 square foot condo and not 3000 square feet, but it should give me an idea of what to expect.
Let’s start with my primary node which is connected to my Bell Fibe Internet connection. Here’s the speed that I am getting from within 6 feet of this node:
So the download and upload speeds are what I would expect from a WiFi 6 connection from a node that’s directly connected to the Internet. But take a look at the idle ping time. It’s five milliseconds which is pretty low seeing as the HH4000 which this primary mesh node is connected to reports a consistent 1 millisecond Idle ping time.
Let’s move on to the secondary node and see what results we get:
Well, the upload speed dropped by about 14%. And the download speed dropped by about 10%. And the idle ping time went up to 8 milliseconds from 5. I suspect that some of this is due to the fact the speed test had to take an extra hop to get to the Internet. By that I mean:
- The first speed test went from the primary node out to the Internet directly.
- The second speed test went from the secondary node, to the primary node, to the Internet.
Another factor is that this secondary node is connecting wirelessly via a feature called a wireless backhaul. Let me go down the rabbit hole on this. This ZenWiFi AX XT8 has three WiFi bands:
- 2.4 Ghz up to 300 Mbps
- 5 Ghz up to 1201 Mbps
- 5 Ghz up to 4804 Mbps
You can set it up to have that 5 Ghz 4804 Mbps band (that speed is calculated in ideal conditions which don’t exist outside a lab by the way) to be reserved simply to connect to other nodes and not be available for anything else. This is how I run my XT8 system. That sounds great, but the wireless backhaul is subject to the same limitations as laptops, tablets and smartphones that use WiFi. The further you are from the access point, the slower your connection will be. Or in the case of wireless backhaul, the further the secondary node(s) are from the primary node, the slower the connection will be. And that doesn’t take into account that there may be a wall or two that it has to deal with along the way, or neighbouring WiFi access points to deal with. So in my case, seeing as there’s about 10 meters (32.81 feet) roughly between my primary node and secondary node with a concrete wall in between, combined with the fact that the XT8 mesh router uses WiFi 6 for the wireless backhaul which has range limitations versus earlier versions of WiFi, and there are about 20 other access points that this XT8 has to deal with, the numbers that I got seem to make sense.
Now onto the bonus round. What happens when I go to the far end of my condo and do a speed test as that introduces another wall into the mix while connected to the secondary node. Here’s the result that I got:
Now the speeds dropped again. The download speed dropped by 20% versus the previous test. And the upload speed dropped by 47% versus the previous test. Strangely the ping time improved by 1 millisecond (which was something that I was able to repeat). But it validates the fact that the further away you get from the node, the slower your WiFi connection will get.
So what did I learn from this? Well, I was able to validate the person’s concern. Which is there is a bit of a speed penalty by using a mesh setup. However, the speed penalty that I see here isn’t that much of a penalty as the worst case here is well above 300 Mbps in either direction. And the ping times are still in the single digits which is more than enough for a Zoom or Teams call, or binge watching something in 4K on Netflix.
But if this speed drop does bother you, there is a way that you can avoid it. If the option is available to you, use a wired backhaul by running Ethernet cables to the locations the nodes will be located in. That will guarantee gigabit or better (depending on the router and what type of Ethernet ports that it has, be it 1 Gbps, 2.5 Gbps or 10 Gbps) backhaul speeds which in turn will make the speeds you get from each node consistently fast. Now this may not be possible as it might require too much work to retrofit Ethernet runs into your home, but it is an option for the best speed possible if that really matters to you.
The other thing that I learned from this is that the further that you get from the router, and the more obstacles such as walls and other WiFi from all around you that you have to deal with, the slower your connection will be. Though in my case, it was still good enough to do a Zoom or Teams call, or watching a 4K show on Netflix. Thus that validates that a single router is likely not ideal for this use case. But it is still usable. Assuming that you don’t have to deal with any dead spots of course.
Thus my recommendation to the person was:
- Buy a mesh WiFi router that explicitly states that it has support for a wired and wireless backhaul. That way they have options in terms of how they want to deploy it.
- Start with two nodes if they’re using a wireless backhaul. A good rule of thumb is to place the second node halfway between the primary node that’s connected to the Internet and the dead zone, but limit the distance to no more than two rooms, or about 30 feet. Then see if you have the coverage that you want.
- If they need more coverage, follow the above with each additional node.
- If they have multiple floors, try placing the nodes on the landing of each floor. And then go back to my second point if you need to add additional nodes after that.
Now if you’re using a wired backhaul, you might have a bit more or less freedom in terms of where you could place each node depending on where your electrician put your Ethernet jacks. But one rule of thumb is that I don’t have the cables that connect the node to the Ethernet jack longer than six feet to make sure it doesn’t create a tripping hazard or some cat decides to chew on it (That has happened to a client of mine. That forced them to hide the Ethernet cables in wiremold to keep the cat from munching on them. Strangely, the cat never went after the power cables). But the tips above will help you to make sure that you don’t have dead spots in a wired backhaul use case.
If you have questions about this, please leave a comment below or drop me an email and I will do my best to help you out.
The North Koreans Have A New Cyberespionage Campaign Say ESET
Posted in Commentary with tags North Korea on August 17, 2022 by itnerdYou might recall that I posted a story on North Koreans posing to get IT jobs in the US. I have a follow up on that story with a bit of a twist. ESET researchers sent a series of tweets outlining a cyberespionage campaign by North Korean APT group Lazarus that is targeting Apple and Intel chip systems via a fake engineering job post supposedly from Coinbase.
Kevin Bocek, VP Security Strategy and Threat Intelligence at Venafi had this to say:
“The North Korean APT group Lazarus has made a real name for itself with its cyberespionage campaigns, and this attack targeting developers with signed executables has the potential to inflict huge damage on North Korea’s rivals. Our research shows that the proceeds of cybercriminal activities from North Korean APT groups are being used to circumvent international sanctions and gather intelligence. The money from such attacks is being funnelled directly into the DPRK’s weapons programmes, and any intel gathered could also be used against its enemies.”
“A key component of the attack is the use of a signed executable disguised as a job description. Code signing certificates has become the modus operandi for many North Korean APT groups, as these digital certificates are the keys to the castle, securing communication between machines of all kinds, from servers to applications, Kubernetes clusters and microservices. We’ve seen countless times how North Korean hackers use signed certificates to access networks, passing malicious software off as legitimate and enabling them to launch devastating supply chain attacks. Incidents such as the 2014 Sony Hack, or the $101 million heist of the Bangladesh Bank via the SWIFT banking system, have demonstrated North Korea’s long-standing interest in the malicious use of machine identities. This attack makes use of a similar technique so could deal similar damage as Lazarus understands machine identity and exploits it so effectively, whilst it’s still such a blind spot for many organizations.”
The North Koreans are clearly looking for new angles to get whatever it is they are looking for. Which of course is bad for all of us. Thus businesses everywhere have to be on guard for whatever they have planned next.
Leave a comment »