Wireless is already ubiquitous in any modern home or business these days, yet it’s one of the areas that probably most upsets employees, managers and IT staff all alike. There’s an assumption that business WiFi must be easy because anyone can purchase a cheap home wireless router and set it up at home, so how hard can it be to do the same thing for a business?

Actually for a small business where you’re just providing connectivity for one or two people with their work laptops and maybe a mobile phone or two in a small office, it’s probably not too difficult at all. You probably can get by with a home/small business router - in the UK these days the one that comes ‘free’ with your ISP connection will probably be just fine - and if the connection is a bit ropey in the office kitchen when the microwave’s running then who cares?

But what about more complex networks? The increase in complexity between what will ‘make do’ for a small business and what you need for a larger business is much larger than you might think. So I went on the internet, and I found this it didn’t take long to see examples of people who were just thrown in at the deep end. I’m going to discuss a few of these and a recent project I’ve completed for my employer to try and clarify things for people who end up in this situation from now on.

There's still a problem with WiFi projects

I've talked in the past about BYOD projects being fraught with difficulties, and they still are. I previously dismissed the technology behind a wireless implementation as being "the least difficult and least interesting part" and I stand by that. However, this is an area that causes a lot of problems where people are implementing wireless systems without thinking through their requirements properly.

Some time ago, a post on edugeek asked for guidance on implementing a wireless network in a small primary school. This is a fairly common request, as there are many similar posts on Server Fault. In these  cases we see people who understand just enough to know that they need help being left to build networks that vary in complexity (the network the server fault poster is trying to build is much larger than the school one being discussed on Edugeek), but in both cases (and with all respect to both posters, who are trying to make the best of their situations) I don’t think they realised just how far out of their depths they might be at first. I’m an experienced network engineer who has worked on wireless projects in the past and I still didn’t hesitate in going to suppliers who could supply us with specialised consultancy when we upgraded our wireless network this summer.

Before you even get that far, you need to be thinking about what the actual goals of your wireless project are. The solution to a requirement to only allow specific devices in a tightly defined area (inventory picking terminals in a warehouse, for example) will probably be totally different to a requirement to allow all the staff, students and visitors to a University to connect their mobile phones and tablets to the Internet.

With that in mind, the questions you need to be asking should include these ones:

How many connections do you anticipate on this wireless network?

How “dense” will the connections need to be (e.g. 30 devices in one classroom or department office is a very different prospect to 30 devices spread out over a large campus).

What will the connections be used for? (Again, 15 people streaming video in one room on one AP is a different prospect from 15 people using a lightweight, web-based line-of-business across a campus).

Do you need to support “roaming”? In other words, will users be sitting at their desk using a laptop wirelessly and that’s pretty much it, or will they be walking around the building using a tablet, or streaming a VOIP call to a phone?

How will the system be managed? How will it be secured? (It’s amazing how many medium sized organisations install individual, unmanaged APs around their site and give everyone the same password when they walk through the door).

What is the impact of a WiFi outage in an area? Or over the whole campus?

Do you need to support ‘BYOD’ for personal devices? Or do you need to support wireless access for visitors?

If the answer to either of the last two questions above is yes, then how many? What level of support can be provided for connection issues? How will users enrol? What do they do if that process fails? It’s probably better to not offer a facility than offer one that can’t be supported and which people grow to hate because of lack of support.

What changes to the rest of your network infrastructure are needed to support the numbers of mobile connections arrived at above? If you’re ordering 80 Wireless Access Points then where will they be plugged in? Do you need to upgrade your switch infrastructure to support POE?

How do any changes noted in the above point fit into your strategy for maintaining and improving the general network infrastructure? It’s pointless to install a wireless network that can support 1500 devices if the wired LAN it connects to is overloaded already and suffers a meltdown on busy days as it is.

In the end, we moved away from the Aruba Clearpass solution for BYOD we first considered and implemented eduroam. If you’re in education then this is what you need to be doing too. If you’re in education and need a fairly “open” BYOD solution then this is the proverbial “No-brainer”. For business, the holy grail of “just say BYOD and magic will happen somehow” remains elusive but solutions such as Clearpass do exist if you can get away with a locked down environment.

Now all you need to do is control how those personal devices access corporate data without treading on the toes of your users, who probably don’t want their entire phone erased when they leave the company because they once read an email about the corporate BBQ from home… Not to worry, I’m sure you included the cost of something like Microsoft Intune in your budget.