Currently expanding a number of sites and taking the opportunity to upgrade to WiFi 6 802.11ax, with MR56 AP's (becuase we have the budget for it)
Question i have is ..
All our applications are cloud based O365, Dynamics crm etc + Heavy on VOIP and Video calls
We have no internal servers.
Normally we put in MS 210 or MS 225 switches.
Seeing as we have an opportunity to upgrade everything the closest Mgig switch is the MS 350 which is a huge jump in cost.
Is it worth me going to a MS350 just for the Mgig functionality based on our app usage, or will there be no benefit at all?
Thanks in advance
My standpoint is that MGig is nice to have, but still not needed with WiFi-6. With channels still 40 MHz wide, MU-MIMO more of a wish than a reality and most devices only supporting 2 SS, the 1-Gig interface limit will likely not be hit.
The next generation of APs will likely be WiFi-6E. If your client-base will be upgraded in the near future to 6E *and* there will be local resources that need to be accessed in a fast way, then MGig could be useful. I still would spend the extra money on something else.
As always with wireless, it depends.
If all your services are cloud based and you’ve, say, only got a 1Gbps internet link then you’ll see little benefit in running mGig to your access points. The bottleneck will be your internet connection.
If you’ve only got low client density, a handle of clients per access point, which are probably only running one or two spatial streams, and only 40MHz channels, then you’re unlikely to see more than a gigabit of traffic.
If on the other hand you’ve supporting a high client density for the latest and greatest Wifi 6 clients running 160Mhz channels with multiple spatial streams on a client, then you’ll push past the one gigabit - but then you really need ask if you need those speeds for your services?
In all honesty I wouldn’t worry about mGig yet. Office 365 doesn’t use much bandwidth, VOIP is tiny (about 100kbps), and even high-def video comes in at around 5Mbps these days.
Evolution of tech will eventually push 1G into mGig (naturally), same way 10/100 was pushed out a while ago. My assumption is in 10 years, I don't think you'll be able to buy a 1G switch anymore lol.
As for now:
Wi-Fi 6 , nah
Wi-Fi 6E (still at least 3+ years away for 'large' client adoption as well) - then yeah mGig might actually be nice to have. Need is a strong word because of WAN limitations, but if your traffic is LAN based, and your rocking 160MHz on 6E, or using all 3 bands on the radio, then yeah mGig will actually come into play.
With 6E, 80 is going to be the new 20, with regard to MHz.
If your in a country that will get all new 1200MHz, then there will be more 80MHz channels in 6GHz, than 40MHz in 5 GHz, currently. Bonus - none of them are going to be DFS, so yeah.
The only scenario I can see of going below 80MHz in 6GHz space would be for stadium/LPV deployments.
Thanks for the extra detail
Agreed none of traffic is LAN based so like most of the advice I don't see the need for Mgig right now
I have gone for the MS250 switches instead of MS350 and some extra MR56 access points.
I buy WiFi 6 accesspoints like MR46 (not only for multigig, they have more advantages than this) and have still MS225 switches. I have a look on the network ports if the 1Gig line is enough. Most of the accesspoints don´t need more than 1Gig. In case the uplink to the accesspoint need more than 1Gig I will replace the switch in future.
Maybe in extreme high density use cases like conventions or stadiums might you need more than 1 Gbps per AP but in most usecases I don't see it happening.
Do watch out for guys with iMac's/Macbooks working with graphic designs over a Wi-Fi connections those guys can be beasts in pulling data.
Actually, in those scenario's you would be least likely to need more than 1G. More clients will reduce TPUT overall dramatically. Stadium deployments 1 client ends up with 200 to 500Kbps each when pushed to the limit.
If you designed correctly you should not drop so low. I know contention would increase and leave less usable airtime but with the RU in WiFi6 clients I'm pretty sure you could get way more.
Btw have you ever tried limiting your WiFi to less than 1 Mbps? Simple webpages are even slow to load and you would have poor performance. So I do believe you are overreacting a bit.
Also watch out for AP's with 2 5GHz radios effectively doubling the amount throughput by splitting people into two overlapping cells with 100MHz separation between the two. I would have to do the math on it 🙂
Joe, let me preface this by saying its nice to have someone else to talk to about this stuff. Although it seems like we might be going off on a tangent from the original topic, my response was still oriented around the uplink port being 1G or higher as a requirement.
What I am getting that currently its only in lab scenario's, or 80MHz environments (rare) where we can push over 1G, however with more clients on the AP (such as very high density VHD) your overall total AP throughput will decrease due to contention.
Also, it has been many years since I did anything close to a VHD installation, so my 200 - 500Kbps was from memory. So you forced me to dig up some old documents I referenced at the time.
I also have limited experience with Wi-Fi 6 (as I assume most do still), so I can reference this link here which covers it some:
To clarify, I'm talking VHD with tens of thousands of clients. The RF is the limiting factor at such scale, pretty much any AP vendor has the hardware to handle excessive client counts these days, but you won't be able to squeeze much when you have 100 or more clients on a radio, due to the RF itself I mean, since its a finite resource etc.
I can't imagine any scenario in a VHD setup were you would be running anything other than 20MHz, so that alone (just looking at MCS rates) will be the determining factor of total throughput per AP. Even a 3SS client (laptop only in this example, since handhelds don't go above 2x2 that I'm aware of) at MCS11 is 366Mbps, then cut that in half because wifi 😃
20MHz AP with 100 1SS clients
72Mbps * 50% / 100 = 360Kbps
Realistically it would likely be even lower since the 50% is more like 40% due to contention and fall off. So more like 250Kbps each.
To be fair there are a million scenario's that I'm not accounting for, so yes good TPUT can fluctuate all over the place. However when dealing with numbers at such scale, it's been my experience that math has to be performed to get a rough idea of what we are looking at to set expectation with client etc.
Some links from Aruba (which I consider the holy grail for VHD design) from people who have a thousand times more experience than I do with VHD deployments.