Excuse me if this is the wrong forum, but I thought your knowledge of mobile devices might be of assistance to our firm.
We are currently undergoing a large scale construction project and redoing all the networking as well. I have been working tirelessly with the builders to future-proof our building for new technologies to come! I was hoping that someone would be able to provide information on how much bandwidth is normally needed for one AP for around 200 mobile clients. We need to figure out how many Cat6 cables to run to each site, so I was thinking 3 inch conduits would suffice, but the construction workers have been complaining that this may be overkill, not to mention a bit too wide to handle in the dry wall. They want to use 1.5 inch but I don't think this is adequate for the number of cables that will need to fit in there to handle our high bandwidth. (fiber backhaul)
Wish us good luck on this deployment!
Comment your opinion below..... any help is greatly appreciated.
Marco & Associates LLC
We use 1-inch conduit at all our network drops, and ran four CAT-6 cables. We have ran six in some conduits, but I consider that tight(as I like to keep room for future-or if a cable needs to be replaced).
The highest spec AP's at the moment use either MGig (which is a single copper cable capable of running up to 10GBaseT) or dual Gigabit cables. Note that you can already get AP's that can generate 2.5Gb/s of load - so the dual Gigabit circuits is not enough.
If you want to use MGig and up to a 5Gb/s connection:
If you want to support MGig you should also consider the MS350-24X, which has 8 x MGig ports.
The MR53 is an example of a Meraki access point you can buy today that has an MGig port. In its case, it runs up to 2.5Gb/s.
ps. What are you using for your fibre? OM4? OS2?
" Note that you can already get AP's that can generate 2.5Gb/s of load - so the dual Gigabit circuits is not enough."
This is a pretty misleading statement. Under theoretical and or perfect environments you could possibly exceed a gigabit per second, but you would be putting your APs at least at 80mhz wide channels. This is not sustainable from a channel planning perspective today however and thus in high density and even medium density offices you will more than likely be shooting for lower channel widths such as 20 or 40mhz. I do not know of a single AP on the market that is actually capable putting close to 2.5gbps on the wire. This is due to the amount of overhead associated with wireless transmissions and the fact that there still remain little to no 160mhz wide capable devices that are also leveraging the 3+ spatial streams.
@WirelesslyWired I guess I don't really agree with you, but that is ok, different people can have different opinions.
I just checked a bunch of our clients, and the Dashboard reported the vast majority of those AP's to be currently operating with 80Mhz wide channels. I really don't see it being that long (maybe 4 years) before operating 160Mhz wide channels will become more common place.
And I can tell you from the throughputs we measure at clients they are definitely getting 80Mhz of channel bandwidth. Maybe your clients tend to have different profiles to my clients. I can only comment on what I deploy and see.
Consider the Cisco Enterprise 3800 series access points with their dual personality radios - which are especially designed for dense environments. Consider the case when both radios operate on 5Ghz and 80Mhz channels and chew 160Mhz of bandwidth between them. Now consider the case where both radios operate on 5Ghz and 160Mhz channels and chew 320Mhz between them. The 2.5Gb/s limitation is reached, quickly.
The person was asking about future proofing. I'll stick with my original comment - I would be aiming to deploy a solution that supports MGig to the access points, and that I don't think dual Gigabit uplinks is going to cut it.
Physics doesn't have an opinion.
I wasn't calling you out on your recommendation to plan for mgig. It will be great for 802.11ax.
I am referencing you saying that today we have an AP that will put 2.5gbps on the wire. It is simply not possible today. Wireless throughput and wired throughput are vastly different as wireless throughput suffers from management frame overhead that can at times hit anywhere from 30% overhead to 60% overhead in contentious environments. The wider the channel the higher the overhead as well. I am not agreeing to disagree as I am stating a fact in how wireless medium throughput operates. Unfortunately 160mhz isn't doable today and is why we have it disabled in dashboard.
Once again I don't disagree with your recommendation to plan for mgig.
By way of introducing an element of future proofing, I run Cat 6A cable ( 10GBASE-T up to 100 metres). It is noticeably stiffer than Cat 6 and requires a larger turning radius, which took some unexpected implementation adjustments, which may render the contractor's objections to 3-inch conduits nugatory.
Virtually all WAP vendors advertise fictional speed/capacity specifications.
They have a system for calculating it.
They calculate the maximum simplex speed for each radio under ideal and theoretical laboratory conditions (WiFi is Simplex) and then add the values together. However, the link between the AP and the switch is Ethernet full duplex. Then when you make adjustments for real world conditions you tend to find that a WAP rated at a 2.53 gbps radio rate is not limited by a 1 Gbps link in the real world.
Having had to audit the December 2016 ITU conference report, I would be overly concerned about 5 GHz use by WiFi; access to that part of the spectrum is sought by a lot of industry sectors and special interest groups, I look forward to obtaining uncontested access to the 60GHz portion of the spectrum. The Koreans and the Japanese are making huge advances in overcming distance limitations. We are probably just as much overdue for a switch from copper to fibre as we are changing from IPv4 to v6.
I use three inch for the main feed to the pole. I then use 1.5 inch orange flex for indoor runs that are through floors and ceilings.