Has wireless come of age? Will 802.11n be the silver bullet?

This post is sparked by an article on CNet which covers a recent Burton Group report comparing Gigabit Ethernet to 802.11n, the latest version of WiFi. So – will it happen?

I’m still not convinced. It’s not because 802.11n isn’t a good technology, it’s because I don’t really like the idea of sharing, at least when I need guarantees as to my level and quality of service (read: bandwidth).

We have a ‘pre-n’ access point and some devices where I work and it’s great, it’s fast and it works – but it’s still not a ‘standard’. If you follow the Wikipedia link above you will find a timeline of the .11n draft which started back in 2004. I still remember discussing it when it was first proposed and it sounded oh so wonderful, theoretical 270 Mbit/s easier and better security… It would make right what was so wrong of the current .11 standards of the day (at that point being mostly .11b with a smattering of .11g gear and the glimmer of .11a on the horizon – which incidentally still hasn’t really taken off). The specification is still in the IEEE process and it’s expected in March 2009 and has been through a couple of drafts now, the 2.0 version of which has seen manufacturers such as Belkin, Linksys (Cisco) and Apple become impatient and release their own ‘pre-n’ gear.

While it’s good that the market is driving urgency into the process, it must also be recognised that the consumer is the potential loser here should radio design and spectrum management change significantly between the premature release of the pre-n and the eventual release spec due in 2009. The manufacturers won’t care, they will just build new kit – and they get to save themselves the anguish of handholding ‘early adopters’ through upgrading the firmware of their pre-n devices to bring the device into line with the released spec.

Now – to the nuts and bolts of my concerns (let’s focus on the technology and ignore the customer). There are five major things wrong with wireless which I’d like someone to reassure me over:

  1. Shared Spectrum (thus bandwidth/throughput)
  2. Ease of Setup for normal people
  3. Security of Networks
  4. Connectivity / Coexistance with ‘legacy’ a/b/g standards
  5. The growing number of ‘connected devices’

So, let’s knock them off one at a time (as best we can anyway). The 2.4GHz and 5GHz spectrum is shared that means that if (which over a 3-7 year period will turn into WHEN) my neighbours also get .11n gear, they too will be wanting to use that spectrum, along with my (and their) legacy .11b/g devices, and DECT phones, and baby monitors etc. etc. Of course, the reach of .11n is just over 5 times that of it’s .11a/b/g predecessors so, while I can see one of my neighbours access points currently, and need to walk outside and to the back hedge to pick up another two, with .11n I will be exposed to a whole heap more access points, all wanting to use the 2.4 & 5GHz spectrum around my home. That then raises the question of how can this cross over be mitigated, how easy can we, as an interested/responsible community, make it for end users to switch on their gear and tune it for their (and only their) environment?

So ease of use is the next point we need to cover. “It’s working I don’t need to do anything more” seems to be the attitude of many users, and I base this on my semi regular drives around the streets in my suburb with a Laptop on the dashboard and NetStumbler running. There are a frightening number of access points out there not only broadcasting their SSID, but broadcasting the default name of DLINK or Linksys. Logic would predict that, if someone has neglected to turn off the SSID, or at least change the default SSID for their access point, then the odds are also good that they have not changed the administrative password for their device, which then exposes them to people not only ripping off their bandwidth, but also accessing their system, opening up holes in their home network and potentially (if they’re “bad people(tm)”) implementing a war pharming attack on their modem.

You may be interested in seeing what size the issue is where you live, there are some good resources here to help you with that, just bear in mind that in many countries, use or even possession of these tools may be considered an offence. It’s not good practise to go changing peoples settings, or to leech their bandwidth, so I’d recommend you educate yourself in the regulations covering your corner of the globe before you set out on your investigation, and if you do find open nodes and have an overwhelming urge to do something about it, how about just dropping a few notes into the letter boxes surrounding the location of the strongest signal advising them that their AP can be seen and accessed from the road.

If it’s easy to setup, this ease also needs to extend not only to the radio strength to avoid crossovers with other nearby networks, it also must account for the security of the home network.

So, security then – an unsecured AP allows a malicious (*#&(* to change your DNS settings so, the next time you head off to do your internet banking you are in fact directed by their rogue DNS server to a site they have setup to look like your favourite bank and, a few seconds later they have your credentials, they display a “failed attempt” message, and then direct you to the real site where you log in again (being careful with your typing) and successfully complete your transaction. The fake site in the meantime has sent your details to them (or someone in a country with more… flexible internet crime laws) and your account is cleaned out. Okay, that’s bad – you lose some money, per haps a lot, perhaps none because you bank has some two factor authentication running and you are alerted to the fraud. But consider this… with access to your local network, the malicious (*#&(* scans the address space, finds the machines which are switched on, including the one where all your digital photos and videos are shared and… deletes them. All your digital memories gone in the click of a mouse – you’d better have a good backup regime, or a very comfy couch and understanding spouse because there’s otherwise no way of getting those images back of your wedding / kids birthdays / special anniversary where you got drunk and… well… the images have gone, or have been compromised and posted somewhere where you may not want them to be.

Connectivity and co-existence… Now (uncaffinated) I’m an easy going kind of guy, but if someone is affecting my ability to stream HD content to the average 2.3 televisions in my home, then I’m going to get grumpy. If my neighbour starts a big P2P download in the middle of an All Blacks test, pumps up his wireless signal to boost his throughput, and kills my bandwidth in the process, then I’m going to get grumpy. If my legacy 802.11b web cams are reserving the 2.4GHz channels for themselves and not playing nice with my .11n gear, then I’m going to get grumpy – and let’s not even discuss the lowest common denominator approach to wireless security, I don’t care if my Webcams can only handle 64bit WEP, I want the rest of my devices, regardless of their a/b/g or n to be using the strongest possible encryption available to them.
Let’s also look at the throughput question. Wireless technologies tend to have nice theoretical maximum rates, and this is what the crayon eaters plaster all over the box when you go to buy your kit. But the reality is that the typical throughput of wireless is around half that of the theoretical maximum – except for .11n, which is around a third of the posted top rate. Now don’t get me wrong – 74 Mbits/s is better than what we’ve had to work with previously, but as I stated in one of the previous paragraphs, we need to remember that this is shared spectrum and, as the spectrum becomes more and more saturated by wireless enabled devices from laptops to set-top boxes, IP security cameras to video and VoIP phones, we are going to get a lot of traffic all wanting to use this shared highway, and for those of you based in Auckland, you’ll know all too well what happens when lots of traffic wants to get from A to B (or A to D, or wherever) all at the same time. This could be a “Very Bad Thing(tm)” for the user experience.

So how many things are going to expect an IP connection? Well, Yamaha, Pioneer and Denon are just three of many examples of manufactures of home audio equipment who are putting IP connectivity into their Amps for streaming media and in home distribution. As ‘no new wires’ connectivity increases, so too will the desire to run multiple zones of media, Korsakov in the kitchen, “Rage against the Machine” in the rumpus room and Lucy Lawless in all her flat screen, DVD box set glory in the lounge. Our appliances will also have data they need to share with the house, air conditioning units need to respond to remote sensors reporting a change, the fridge may be reporting that it no longer detects any bottles of milk inside itself. That’s not forgetting that we still will have data driven devices for VoIP calling, email and playing games all of which will have impacts and differing quality of service requirements on the home network. Then question very quickly becomes not if 74 Mbits/s can deliver a good experience, it becomes – will 74 Mbit/s be enough for all of the devices in your home to co-exist and transmit what they need to, without you noticing.

To summarise what has become a very long post – I’m looking forward to 802.11n. It’s been far too long coming and as an industry we need to shorten the cycle time for standardisation, 5 years is way too long to wait for a new generation of wireless.
I don’t think that 802.11n will solve all of our issues – and believe that there is still the need for a structured cable / wireless connectivity hybrid in the home. Am I wrong? Tell me why!



, , , ,




One response to “Has wireless come of age? Will 802.11n be the silver bullet?”

  1. […] caused quite a stir. It’s been mentioned in the press (ZDNet, PCWorld, Network Computing), challenged in blogs, and defended by its author. Will 802.11n deliver the deathblow for ethernet? You be the judge. […]