FortiWLC-200D & Fortinet AP822i WAP's for new wireless network

[lafleur1977]

Hello,

 

I am currently looking at upgrading our network and wireless network. Currently have HP switches, 1GB backbone, 100MB desktop. Connecting to that is a netgear wireless solution with injectors to WAPs or mini poe netgear switches. It was initially okay, reasonable coverage and working, nothing fancy though. Now we have a full blown 1:1 chromebook rollout (1000 devices aswell as the 500 windows PC's and laptops) and its struggling.

 

I have been working with one company who have done alot of good work in the past so where my initial choice. They have proposed what seems like a sensible solution - upgrading current HP switches to 1GB, some pow where needed, a new core 10GB core switch between our core and some of the larger cabinets. In terms of wireless they are recommending Fortinet (Previously Meru) with a FortiWLC-200D & Fortinet AP822i WAP's for new wireless network. I have spoken to several other schools in my area who have Fortinet and have recommended it. I just wondered what other peoples thoughts on Fortinet are currently? Has anyone used Fortinet with Chromebooks?

 

We have done several site surveys and really got into some detail planning so I am confident it will be a good solution and good value.

 

Thanks

[AlanD]

HP switches...they're great...maybe not the "Bentley" solution (=Cisco) but certainly Rolls Royce. Do you not think as a school you need to look at a Ford Mondeo type of solution....I'd be taking a good look at Unify Switches ...possibly netgear both of which would work fine - and infinitely better than what you currently use. Don't expect a solution provider to be rushing to recommend either - because their margins would be really small..and they probably have no experience of using or installing them. Get a test switch (you could sell it on Ebay later if you decided against it...and probably get 80% of the cost back). We use 10GB netgear core switches with 2x 1GB LAG connections to edge switches...with vlans...and it just works....

 

Wireless...well I'd be very wary about Meru. I don't think their single channel architecture gives you the kind of data bandwidth/density you need in a school and if you are going to pay for such a "premier" solution - I'd be taking a really good look at Meraki (the web interface will blow you away) - and/or for that matter Aerohive, Aruba or Ruckus. Personally, I wouldn't choose any of those until you have tested a Unify HD access point (no annual charge...and half the price of premier brands). Is it as good? Probably not....but you'd require a stop watch and some exhaustive testing to spot the difference - only to dsiver there was little in it.

 

Keep in mind that (a) all access points have the same chipsets and 802.11ac2 software at their heart (otherwise they wouldn't be 802.11 compliant. And they all do load balancing, band steering and radius authentication etc etc. (b) All those super fast data rates quoted in the literature are largely ficticious. You are not going to get gigabit rates delivered to your chromebooks and/or replace wired desktop suites with domain wireless laptops that work just as well. Think somewhere between 100Mb/s and 200Mb/s for the whole access point in terms of wire speed. And halve the figure is you have more than 30 clients hanging off a single access point because of the size of the collision domain. © You will need the access points in the centre of the ceiling of most classrooms. Out in the corridor is not going to give you any of the high "ac" data rates - because they only work for 3 to 4 metres in open space at those rates...and going through even thin walls or bodies will drop the signal to noise ratio to make such speeds impossible.

 

..and no; your supplier isn't going to tell you all this bad news...because otherwise they wouldn't be able to sell you their over priced stuff and continue to charge you every year for cloud subscription costs.

 

..and don't get taken in by the promise of MU-MIMO delivering data to four clients at once....it only works "one way" - download and with overheads the best theoretical improvement is x1.8 And ... in reality you get around x 1.2 and then ONLY when all clients are ac capable...with mixed clients its actually slower than non MU-MIMO (...bet they won't tell you that either...)

 

Make absolutely sure - that you test a Unify HD access point. You will be able to afford twice or three times the number..and there are no on going costs. If you decide against it some people sell them on ebay for more than they paid for them.....but given the very large number of schools using Unify very successfully...I'm pretty certain you will keep it/them.

 

(actually I use Aerohive...over 100 APs .... and just use their command line code to deploy....becuase there is no way I would be paying for their subscriptions costs...which are great for a small business - perhaps one with several offices....but don't seem to be the right kind of cost model for a school...but I'm not suggesting this is the solution for everyone)

[AlanD]

Forgot to say....I'd be wanting to check - by looking at some graphs/switch data ...exactly what the problem with the existing system is. No point in investing in 10G core and new wireless - if the real problem is that your external internet bandwidth is limiting...or the server is just providing the data fast enough. It can be surprisingly difficult with random access to get 1Gb/s from a RAID array of non SSD disks....

[speakercon]

We recently had a fortinet wireless system installed (not my choice!) and I wouldn't recommend it to my worst enemy. Constant problems since it was installed and still going on. Devices loose wireless connection for no apparent reason, then wont reconnect, some laptops and devices refused to connect full stop however all were happy to work with the Merika AP I have and all used to work perfectly on our old cisco aironet wireless. Fortinet were no help they said it was our equipment that was outdated and incompatible. Firmware seems to be at the root of the problem in my opinion. I’d put it all in the bin if I could.

[AlanD]

... And keep in mind that all new access points that do ac2 require 802.11at capable poe+ switches. Yes, they are 802.11af "compatible" but are only so by turning off a radio, or multiple streams, and or second ethernet socket. So if you want the benefit of higher speeds and capacity don't bother investing in access points unless you invest in what is driving them. It's also worth mentioning that buying devices with single stream wireless cards tends to make the expensive access points a waste of money too.

[njc235]

... And keep in mind that all new access points that do ac2 require 802.11at capable poe+ switches. Yes, they are 802.11af "compatible" but are only so by turning off a radio, or multiple streams, and or second ethernet socket. So if you want the benefit of higher speeds and capacity don't bother investing in access points unless you invest in what is driving them. It's also worth mentioning that buying devices with single stream wireless cards tends to make the expensive access points a waste of money too.

 

Lots of inaccuracies here. The Fortinet AP832 operates both radios with only 802.11af Poe. 802.11n clients operate faster on the newer wireless because they are afforded more time by the faster clients being served more quickly.

[AlanD]

Lots of inaccuracies here. The Fortinet AP832 operates both radios with only 802.11af Poe. 802.11n clients operate faster on the newer wireless because they are afforded more time by the faster clients being served more quickly.

If we are nit picking about inaccuracies...I don't think I said they "ALL" could only operate one radio - merely that they "ALL" made some compromises and did not fully function with all bells and whistles on 802..11af. And its not that 802.11n clients operate any faster...its just that they get more air time....assuming a mixture of clients and that they are not fighting to get MU-MIMO because mixed clients more than undoes any advantages that might bring. And in a busy environment - its quite possible that that "ac" clients won't get a wider channel - and may not have the signal to noise ratio to benefit from the higher QAM (Unless the student is sitting under the access point) ...so in reality - if not theory - schools may not see much advantage.

 

 

Yes, I am trying to be provocative (and possibly failing) ...the point I am trying to make is that new wireless systems are oversold....and too much is promised. The greatest advantage of ac and ac2 are the wide channels (160MHz) - but - in a dense deployment of a classroom block - with several classrooms using sets of laptops - you can't use wide channels (Yes - its great for a home - or to cover a single office space...or even an assembly hall). In fact - you are even unlikely to deploy using 80MHz channels because there are so few of them - you run into the same kind of problems as you had with 2.4GHz radios....and "noise" from distant radios (and clients) is treated as DFS in use - reducing the number of DFS channels during the day one by one until there are none left. And while some high end tablets support two streams - almost all the "school" laptops - are single stream.

 

Hopefully 802.11ax - with its two way MU-MIMO and other enhancements might provide a better "school "solution....but its always the case that there is something better coming next year. So a new school deployment will have to invest in "ac2". But I'd be very weary about replacing a well deployed "n" wireless and expecting miracles.

 

Too often access points are deployed too sparsely in corridors, so that none of those high speed (QAM) connections at 5GHz make it through the walls - often with clients getting little better than 54mb/s. Replacing these with "ac" access points - isn't going to make much (yes - it will make some) difference - and certainly not unless there are so few you can use wider channels.

 

I strongly advise schools to get some test kit installed. Don't just test with a handful of clients in one room with an access point - try two rooms fulls of client laptops with the access point in a third room (unless you are going for a 1:1 classroom deployment of APs). Compare with what you have now AND compare with several over suppliers. Yes this will take time. Don't buy kit from supplier "X" because you have always done that....or that posh school down the road has those access points. And don't imagine that by paying more you are getting something better. You MAY have to pay more to get something better....but paying more does not guarantee that. And you will only know if you have actually done the leg work and tested it.

[mjk]

I run Fortinet Wifi successfully with 1500 Chromebooks. We have AP832's an MC4200 and (mostly HP A5120's).

Don't pay too much attention to those who say it won't work - I can guarantee you they've just not configured it correctly or don't have a decent install company who can help support them.

 

You should plan to have 2Gb/s to each AP as each AP can provide nearer 1200Mb/s. 10GB/s backbone is sensible if you have more than a few AP's connected to a switch.

Your leased line (and backup) capacity should be reviewed too.

[AlanD]

I'd be interested to know whether you use their "Single Channel Architecture" or just use them as "normal" access points. I bet you don't get anything like 1200Mb/s in real world (wired) data speeds....partly because I bet your clients don't support 160MHz channels...and while dual stream might a double UDP transmit rate - you rarely get more than 20% more in actual data dates - assuming clients have dual stream wireless cards.

[mjk]

I'd be interested to know whether you use their "Single Channel Architecture" or just use them as "normal" access points. I bet you don't get anything like 1200Mb/s in real world (wired) data speeds....partly because I bet your clients don't support 160MHz channels...and while dual stream might a double UDP transmit rate - you rarely get more than 20% more in actual data dates - assuming clients have dual stream wireless cards.

 

Yes we do use the single channel architecture. The majority of our clients only support 2x2 mimo 802.11ac so an individual client won't get near, I've only really seen 3x3 on newer macs that the kids bring in and none of our chromebooks support it yet, but I', sure they will in future.

[AlanD]

There are a number of "meru" related threads...glad to hear its working OK for you...because that would not always seem to the case.

 

http://www.edugeek.net/forums/wireless-networks/179826-802-11ac-dfs-channels-radar.html

 

While I understand how MERU manages the "single channel" solution - and can understand that for various reasons it was probably a good solution for longer range 2.4GHz and a single trolley of laptops moved between a classrooms in a block...I'm not convinced that it is the way to go when you have realtivley short range 5GHz and much higher density of devices in every classroom at the same time. I have head of a number of schools who have abandoned the single channel bit...and if you are not going to use that feature - you may as well take a good look at a selection of different wireless solutions...because they are different to use - and considerable different in price when you look at Unify ....with little difference in throughput.

[lafleur1977]

Hi all,

 

Thanks for your replies. They have given me a few things to consider. With the solution above, I have checked the model of WAP and looks like my provider had made a mistake on their quote and intended to supply the AP832 model. However, when I asked them they have said a new universal AP is now available, will do the same price and is better. Not sure which model yet but the specs are here: https://www.fortinet.com/content/dam/fortinet/assets/data-sheets/FortiAP_U_Series.pdf

 

I have a meeting with them tomorrow to discuss.Any thoughts on the universal option?

 

Thanks

[njc235]

Personally, I'd stick with the 832 for now as it's a proven model. Yes the UAPs are more current but early adoption sometimes has its costs.

[AlanD]

...and I'd still be taking a VERY good look at what else is out there...because you may be shocked by how good they are (Meraki) or how much you could save (or indeed give a much better dense coverage) by looking at Unify. It would be negligent of you not to consider a range of options.