Wes

Might be a little late but...... Extreme Networks make some very slick products and are some of the most reliable in terms of up-time not only because the hardware is stable but they also have a modular OS, called XOS which can automatically restart processes without needing to reboot the whole OS. Extreme's main market has always been ISPs and top of rack switching in data centers for this very reason (and the fact they have a switch chassis which has up to 196 40Gbps ports!). Recently they have brought out a range of gigabit PoE and Non PoE switches, which are all under 10" deep (so great for wall cabinets) at a very competitive price. They also offer 10Gbps uplinks and stacking at 40Gbps in these switches(I'm referring to the X440 range of switches). Personally we have had a lot of success with these in secondary schools and colleges because the price is very similar to HP, the hardware comes with a lifetime warranty too but the main draw is the smarts you can get out of XOS (absolutely ALL Extreme switches run the same OS, which is nice and easy compared to HPs plethora of different OS versions of Commware and Provision). XOS features Identity Manager so the switch can directly query AD or any LDAP data base and apply a role to the user at the edge port - basically .1x without the need for complex set up or multiple databases. They run EAPS (which is like a super fast spanning tree, which can re-route traffic in a link fails without losing a ping, and is great for eliminating the chance of network loops created by over enthusiastic cleaners 'tidying' cables by plugging the loose end into another wall socket!). The there's Universal Port, so the switch will run a script depending on what's plugged into the port - you can use this to configure VoIP phones, WAPs, IP cameras etc - anything you can talk to via XML, and put them automatically into the right VLAN, QoS etc with zero config. All of this saves a lot of time and hassle from a management perspective. Best of all this is all included in XOS at no extra charge and the Ridgeline Management software is free for up to 10 nodes (a single node can be a stack of up to 8 switches). All of the switches stack too btw.

£20 saved!

Sure there's a place for Unifi, and many schools will get exactly what they need from the system and never have any troubles at all. There's a lot of rhetoric going around - 'it's good 'cos it's cheap - don't get suckered into features you don't need' (I'm not accusing anyone other than Ubiquiti's marketing dept. of this btw). There's no conspiracy between other WiFi vendors to overcharge schools or anyone else for that matter. There was never a cartel meeting where the price of WiFi was set, Unifi isn't threatening the Magic Circle by giving away the trick. It's cheap because it does less (less may well be enough for a lot of cases). The £50 WAPs are not intelligent, they have a single 2.4GHz radio and they don't have a firewall. Ubiquiti have recently released the pro version WAP, which added a 5GHz radio and I believe the price is now £180 per unit, plus you'll need the injector at £20 - so £200 per unit all in. Simply adding a 5GHz radio has increased the price 4 fold, imagine what the price would be if they added a firewall, some band steering and load balancing as well as proper roaming. Suddenly a £300 enterprise grade WAP doesn't look so bad, in fact it looks darn cheap. The enterprise vendors have features which are designed to make use of 5Ghz (where you have 19 non overlapping instead of the 3 offered in 2.4Ghz) by encouraging the clients that can into that band (most clients will just pick up on 2.4GHz and ignore 5, even if both are present). There's no Bluetooth, microwave etc in 5Ghz so there's cleaner air and clients and WAPS spend less time waiting for clean air in order to send data, which improves performance for everyone. You can also bond channels in 5GHz up to 450Mbps data rates. If you've got 450Mbps wireless, I think you have a solution that will last 5 years at least - I don't see a lot of desktop users on wired access at more than 100Mbps, and Gigabit Ethernet has been cheap for years. Unifi only offers 5Ghz at £200 per AP, so at best two thirds of the price of your enterprise solution, not the dramatic figures being marketed. I would say if you want to buy the £50 UAPs and have 100Mbps (it only has a 10/100 Ethernet port so let's not kid ourselves it'll run at 300Mbps like the marketing says) 2.4GHz WiFi, expect to rip it out in 2 years or less, so why not buy UniFi (aside from the hassle of finding somewhere to plug in those tens of PoE injectors and re-cabling again in 2 years time) but the with Pro version WAP the price no longer makes taking the huge hit on features worthwhile, in my opinion.

Coverage is the easy part - having the capacity to connect large numbers of users with a decent throughput is another challenge. Coverage wise, all APs are limited by law on their power output, physics being the other limiting factor for coverage. I would suggest this many APs have been reccomended as the school has a capacity requirement rather than purely looking for coverage? If this is the case a lot of times you will need an AP per classroom to connect 30-35 concurrent users with a decent throughput for each. The best WLANs will have automated load balancing and band steering as well as some sort of airtime fairness technique in use (such as Aerohive's Dynamic Airtime Scheduling) to ensure that high capacity WLANs are optimized towards actual user experience.

This can't be said of Aerohive as they do not have a controller, they use a shared set protocols to distribute data and control planes across the APs at the egde of the network. The access points each have a processor which is more than adequate to deal with anything the AP chooses to demand of it. Essentially with any shared control plane (i.e. a controller, cloud or otherwise) you're diluting a pre-determined amount of processing power which is dependent on processing data sent to it in this case via an internet connection rather than having a dedicated processor on each AP, dealing with data processing and forwarding at the edge of the network. HiveManager is available as either a VM or cloud service and plays no part in the data processing. It's a pure management tool and in my opinion offers many more useful tools than the Meraki's, fairly limited, features. The most useful being the Client Health Monitor (instant look up of all clients WiFi, network and application health), Teacherview (built in classroom control software) and the stateful firewall built into every access point which can be leveraged in to allow or deny access to private or public addresses as required for each user group without or tied into wired network VLANs.

This outlines exactly the benefit to be gained from a fully distributed, controllerless WiFi architecture. Now I know Cisco make a controller to handle more than 200 APs, so the reason you have 3 is because the WAPs are dispersed accross the six sites right. The WAPs could all be handled by one big controller otherwise. So these 3 separate controllers have no failover controller, so that's your 3 obvious points of failure right there, all things being equal (which they aren't of course) you'll lose 66 WAPs if the controller device fails. For a resilient set up you'd need 6 controllers (feel free to set me right on your network deisgn, I'm making some assumptions on the information given, not huge ones though). Over and above this you need some more software to organise and control the controllers! Controller based WiFi network = WCS Software x 1, Controllers x 3 (ideally 6?), WAPs x 200. Distributed WiFi Network = WAPS x 200 Personally I think the latter is slicker.

Agree. For me, you can keep your single channel/channel blanket/beamforming controller based technologies. Controllers = expense, bottleneck, single point of failure. Distributed forwarding and control are the future, pioneered and perfected by Aerohive being copied with varying degrees of success by Cisco (HREAP), Motorola (wing5), Aruba (Aruba Instant) - all the big WiFi vendors. No one is copying single channel and there's a reason - it's 5Ghz performance is not as good as cell based architectures. So what if the client makes the roaming decision - that's what 802.11r is for! Beamforming - not for me either. I don't have any beamforming chipset in my laptop or iPhone - so what's the use in having one in the AP? Clients are half the solution and the data must flow both ways. I want to get the user device connected fast, give it enough throughput and have some-way of providing an SLA for my trusted users, guests and BYOD devices. I want to be abel to scale the network coverage and capacity without thinking about controller licensing and throughput.

Distributed intelligence and processing in WLANs is the future. No wireless vendor is putting more intelligence into their controllers, the trend is towards the edge the AP with a central management interface and distributed control accross intelligent APs. Look at any of the major vendors and you will see they are doing less and less in the controller but yet you still have to have one, it's still a major expense when considering wireless deployments and it's still a single point of failure. The war is over, distributed WLANs won. In the future, you'll all have one! Beamforming is great in open spaces but it's debatable whether it's of any benefit in a school. It's a one way, transmit only deal at the moment. When you think you have 30 or more clients on any one AP, the clients are where the majority of the airtime traffic is coming from. An IEEE standardised version of beamforming is coming, but it'll be a both way deal with new chipsets in both the AP and the client. Single channel plays on the myth that multi channel WLANs don't work. They do, ask anyone with one. Most managed multi channel systems will allow for densely packed APs, automatic RF and power settings mitigate against channel conflict. All vendors have at least one, slick function that makes them different. I truly beleive the only one that really matters right now and can bring a real benefit to all deployments are the distributed control players.

Thanks all, invitations to this event are no longer being taken as we have reached capacity.

I think you're getting smoke and mirrors confused with differences of opinion. Your phrase (to me at least) implies some sort of trickery or untruthfulness. Clearly there is a case to be made for each vendors solution, otherwise they just wouldn't be in business too long. Each solution has it's merits and the intention of a case study is to highlight those merits. Think of it more like an advert, which is exactly what it is. However in this case it's backed up by a real life customer, which is more than you get on TV!

You hit the nail on the head. You should demo as many as possible. They will all do a job, some better than others. Which one fits comes down to many factors including cost, deployment scale, your requirement, your RF environment etc. But you can't deny, not having a controller to limit scalbility, pay for, support and license is a massive bonus. They will all demo for free.

Contact Mark Powley. He's the Aerohive channel manager in the UK. [email protected] 07931 371565

Yes, the 4210G was out before the merger was completed (just) and is very much a 3com product. It runs Commware V3 which is the same as your existing 4200Gs. You'll just buy it under a HP part number.

HP bought 3com for it's huge RnD and manufacturing base in China, which was previously the H3C brand 3com were running as a joint venture company with Chinese networking giant Huwawei and aquired a large slice of the Chinese networking market into the bargain. They also have aquired 3com's data centre switching portfolio which allows them to compete better with Cisco in the data centre space. The RnD will continue to pump out from China. As for now, the HP and 3com products have been aligned together into 3 categories V class (Value) E Class (essential) and and A (Advanced) and all have HP style part codes but other thanthat ary still very much two different products.

The Netgear solution won't cope with 30 netbooks imho even the 802.11n system (we have some of those APs here and they're each about the size of a laptop also!). Load balancing doesn't seem like the ideal solution, you need to find out how the Netgear system handles this, you may end up with a bunch of real slow connections and a few good ones. If you want to keep costs down and buy into an enterprise class system which you can use for years to come you could look at one of the controllerless vendors like Aerohive. This keeps the costs of deployment low as there's no wireless controller to buy, the APs as far as I'm aware are the same price as an Aruba, Meru or Ruckus AP. Aerohive is centrally configured and monitored via software network management which is provided on the AP and is not integral to the working of the WLAN. Aerohive also have a neat applet called Teacherview which allows teachers to see what the students are doing on each connected client, point their browsers at stuff etc. There are no feature licenses to buy, everything comes as standard inc. QoS, stateful firewall, guest management, VoWLAN capability, Airtime Scheduling etc etc. Netgear will be cheaper than most but it's providing less. The question is, it enough for your needs. I would question whether any 802.11a/b/g new deployment is a good investment. I would at least go for their .11n kit.

Channel and power optimisation requires constant updates from the cloud controller though right? So no internet, no auto RF/auto power?

Hi all, I would like to extend an invite to a free 'Future of Wireless' event or to give it it's correct title "802.11 Alphabet Soup: The Current Recipe for Your Next Wi-Fi Network" at the new £20m conference centre at Lady Margaret Hall, Oxford University on 22nd October. The speaker is Matthew Gast, author of '802.11 Wireless Networks: The Definitive Guide' (O'Reilly). He is a member of the IEEE 802.11 working group, and serves as chair of 802.11 Task Group M. As chair of the Wi-Fi Alliance's Wireless Network Management marketing task group. Matthew will be discussing the current standardization efforts and what this means for Wi-Fi networks over the next five years. THIS IS NOT A SALES PITCH, it's a technology overview. The event is FREE to all Education IT Professionals, breakfast is served at 9:30am and the event kicks off at 10am. Please PM me to reserve your place, including your full email address,name and orginisation so that I may send you your formal invite. There are 140 spaces, 50 of which remain as things stand.

From Meraki's website: So authenticated users are okay it would seem but any new users will not be able to authenticate. So it would be problematic in education with classes changing every hour or so. I know a certain LEA which have chosen Aerohive's wireless solution for all their BSF schools (these ones are built already!) because every student will use a laptop, day in day out. So you could have 900 students all wishing to authenticate at 9am. Aerohive can handle this as authentication is done at the AP, not via wireless controllers. Meraki would want to push all this traffic out through the internet and therefore through the core, to their servers. Every solution has it's place, I believe Meraki has some merits for sure.

Meraki, certainly. There case studies tend to be very focused on guest wireless hotspots. Aerohive is up and running in the UK, the last year has really taken off for them. They have good case studies on their webiste.

Aerohive will be no more expensive than Ruckus - £425 (in my experience) per AP and no controller to boot. You don't have to have it as SaaS, they do VMware and hardware versions of the HiveManager NMS. Support is the only ongoing cost as it is with any vendor. Moreover there's no licensing required for firewall, WIPS, Voice etc as there are with many controller vendors. They also have a neat feature called Teacherview which allows the teacher to control the clients in their class, turn wireless on and off, check what they're browsing and redirect them to web pages etc. Again, it's included in the price. Ruckus beamforming is good, but is it really a necessity in schools where you're generally in the same room or next door to the AP. Currently this proprietory beamforming, soon standardised beamforming will become available which uses an entirely different chipset and antennas curently in development. I understand you can get demo kit from Aerohive, free of charge. Most vendors will do this, you need to try a few and see what fits best.

I think we've all learned something today

yes, maybe you can, but you shouldn't - it's naughty. Anyway, now we're just playing semantics.

I'm sorry, but I believe you were misinformed or have misunderstood. Aruba certainly has WIPS built into the AP, as do many other vendors but you cannot 'bombard' other transmitters in the unlicensed, public 2.4 and 5Ghz spectrum, which was my point. Anyway, not the main point of the thread..... don't want to argue.

2.4 and 5GHz are unlicensed spectrums (public), you can't just go knocking out other people's wireless. There is no way Aruba does this, it would probably be illegal. I think you need to also look at Aerohive's product. Like Meraki it offers Software as a Service, cloud based management. However unlike Meraki, this is simply a management console and not a 'cloud controller'. APs don't rely on any controllers, they are intelligent and use a set of protocols to create a distributed control plane and local forwarding. All APs come fully featured, out of the box including stateful firewall, RADIUS and WIDS. There's no need to set up VLANs for guest access. Guest traffic can be managed by the APs in the firewall. Also, you won't have the hassel of asking the LA to open up firewall ports for outgoing traffic as the APs use HTTP over port 80 to speak with the cloud based network management. Moreover Aerohive offer hardware appliance and a VMware version of the management platform if you prefer. The APs don't care if they see the management platform or not, it's simply there as a NMS. You could take it away and the WLAN will not be affected. APs in my experience cost no more than any other 802.11n and you don't have the controller outlay, licensing costs or single points of failure associated with controllers.

Meru offers a single channel architecture, which is positioned as eliminating the apparent co-channel interference issues that other multi-channel wireless systems are dogged with (the other 97% of all non-meru wireless systems apparently don't work). Some other vendors (Bzzzzz) offers auto channel selection which eliminates any problems with co-channel interference. Also Meru will need to be 100% certain that no neighboruing wireless is using the same channel, otherwise the whole deplyment will need to be re-configured to another channel. In my opinions Meru does a bad job with high density cleints as it means you need to open up more channels to support more devices. In the 5Ghz spectrum there are 19 channels - why not use them? It's like opening up just one lane on the M25. Scaling with Meru is difficult, each AP having just two radios means any more than 2 channels requires another set of APs. Roaming is achieved with the virtual port functionality within the Air Traffic Control software where a virtual AP follows a client accorss physical APs. There is however 802.311r which defined fast roaming and is incorporated in the Voice Enterprise certification announced by the WiFi alliance which will leave Meru's approach overshadowed. I'm certain Meru works though, I have seen it. But I would question that it is the most forward looking solution out there and in my experience is also quite expensive, certainly at the controller end. Some vendors have eliminated the controller which makes for a much simpler, more scalable and lower cost deployment whilst offering the same or better functions and features. Bee sure you're making the right choice. Bzzzzzz.