50 devices or 100+ - Why the difference? Is it real?

[lifereinspired]

Hello,

I’m looking at a few different routers to replace our aging one. One thing I noticed is that different manufacturers state differing numbers for how many devices their router can handle - with the same internal chipset. For instance, the new Netgear Orbi AX6000 states that it can handle more than 100 devices while the new Linksys MX5 (also called the MX10 when sold as a two-pack) says 50+. From what I can gather, both have the same Qualcomm quad core 12 stream chipset powering it. So, why the discrepancy? I guess 50+ could also be “more than 100” but I’m wondering if there is any real difference in what they actually handle? Is it just two different marketing approaches or will they really handle different amounts of devices?

Looking forward to hearing your thoughts on this. Thanks so much in advance!

 

[thiggins]

Are the specs stating wired or wireless capacity? Which one do you care about more?

 

[Trip]

Because it's not just about the base reference chipset. It's about other physical design components and code. Antenna, for example; is it a CNC-milled, 6-axis proprietary design like you'd find in a Ruckus or Cisco AP, or just a generic piece of sheet-metal origami like you'll find in almost every consumer router or lower-grade APs? Are there any co-processors on the board that may be added in? How about the quality (or lack thereof) of the firmware running on that chipset, and how well-developed it is (or isn't) to handle however much client load?

Quoted and real-world capabilities can and will vary, based a lot on the above differences, and more.

 

[Razor512]

I wonder if those numbers are just based on marketing and how many devices they got around to trying with it. in the past, I have seen many APs where because it was open and unsecured, you could see around 100 devices connected to it via WiFi during the 802.11n days, and they weren't high end devices either. I wonder if it is just a limitation based on how much it can handle before performance drops by an unreasonable amount, or if it is based on certain requirements such as does each device need to stay within a specific latency and be successfully able to download stuff simultaneously.

 

[Trip]

Most advertised numbers are generally misrepresented when it comes to real-life applicability; whether it's a footnote such as "number denotes maximum amount of total connected clients at any given time, but gives no actual indication of total real-world capability in Mb/s or number of TCP/IP sessions", or something as pedantic as varying definitions of what a "client" is (is it a single hardware endpoint device (as most would presume)? What if they footnoted it to be a single application session instead? I've seen crazier, and I'm sure you have as well...

You'd think certain terminology would be all but obvious and agreed upon, but of course marketing and sales will take every literary opportunity to maximize strengths and minimize weaknesses.

To quote Bill Clinton, "That depends on what your definition of "is" is."

 

[lifereinspired]

You all have some really interesting thoughts. And it’s back to what I was wondering - what can you believe and what can you expect out of any given router?

In the examples above, neither specifies what a device is or whether it’s specifically wired or wireless. Both share the same Qualcomm chipset (from what I can tell). Both support tri-band mesh, wired backhaul, though it appears that they each handle the wireless backhaul a little differently (Netgear uses just two bands for internet traffic and the 2nd 5ghz is backhaul only while the Linksys allows internet traffic on all three bands and dynamically uses whatever band is least congested for it’s backhaul). I don’t know of any additional chipsets as supplements or any CNC machined antennas. The Orbi AX6000 gets worse reviews than the Linksys by a fair bit. And for some reason, Linksys claims it will handle “50+ devices” and Netgear “100 or more”. Given than they have similarly built designs (in general), the same chipset running them, both supporting mesh technology, why would one objectively handle half the number of devices that the other claims? The only thing that I can think of is that it’s two different approaches to marketing. Linksys plays it a little less, thinking that perhaps if they say 100+ devices, people with an ordinary number of devices might think that this isn’t the type of WiFi router they need - overkill. Whereas perhaps Netgear is overstating what it can handle or maybe just being more “optimistic”? Anyway, it’s confusing and makes me wonder if they will actually handle the same number of devices in practice. The Linksys definitely appears to be the better option right now from people’s experiences but I have a lot of devices and want to make sure it will handle it all for a few years at least.

 

[thiggins]

Yes, it's a difference in marketing. Welcome to the over-hyped world of Wi-Fi.

The industry (and I) have long been looking for a meaningful test method to accurately assess the capacity of a WiFi AP/router/mesh system. It's not an easy problem because there are so many variables.

Still, if there were a standard method, at least you'd have one yardstick. Performance measurement is what the Broadband Forum is trying to do with its TR-398 standard, but there is a ways to go.

 

[Trip]

Regardless of what it says on the tin, the most meaningful measurement is how it performs in real-life. And often, you're only going to find that out by buying and trying, even with objective reviews accounted for.

The main issue, of course, is that all this stuff is just a byproduct of free-market capitalism; rush the product out the door with the bare minimum amount of QA/QC to get it on shelves and into consumption. Then let the users be the beta testers, often for the life of the product. The OEMs figure a majority of operators will only use certain features, which work "well enough". Any remainder is cast aside as too small a concern to alter the business requirements of meeting marketing and sales deadlines. Wash, rinse, repeat.

At the end of the day, it's on us to engineer around the need to be concerned about any marketing promises to begin with. Only deploy proven products via best-practice (site survey, AC Wave 2 gear, etc.). This is why I run business/enterprise gear at home. To some, it's insane overkill. That fine. To me, it's common sense. My network runs like an appliance, as opposed to a toy. I do understand there's a skill/knowledge gap that prevents others from doing the same, and for that I don't really have a good answer, other than to suggest people learn up on their own or hire the work out; either way, it's an investment in time and/or money, but I think the end result is worth it, considering how much some of us rely on our networks on a daily basis. To each their own, though.

 

[coxhaus]

And don't forget security. Who is most likely going to discover and fix security issues down the road over time.