Archive for the ‘Architecture’ Category
This was a question that I asked the audience during my presentation at today’s Information Systems Security Association (ISSA) meeting in Phoenix:
I received a following set of questions via email and thought that rather than only respond to the one individual, but post it on WiFiJedi.com for general consumption.
Here’s what I received:
I have recently done a deployment where the customer is using 802.11n in both 2.4 and 5 GHz range. I have configured a WLan with that is providing coverage for all the radios including a/b/g/n on a 4402 controller in WiSM Module. The clients are also unable to run on all the radios. Everything is working fine, but we are seeing clients that have been connected to either a 2.4ghz radio or 5ghz radio, tend to switch between them when running for a while. I am trying to understand this scenario, since there is no setting on the controller that I can find would allow the preference of the radio. we are currently running the latest 6.0 code on the WLCs. Secondly the data rates are set to 24Mbps or higher for both radios, that means 802.11b is not allowed. The client machine was sitting at one position for 24hrs that means the user was not moving around hence the roaming should not be involved I think. Lastly we did the survey for 802.11a coverage and I don’t think there is any issue with the coverage, since it connects to 802.11a at excellent. The question really is that once the client connects to 802.11a on excellent and is running on that radio for an hour perfectly and the user is not moving at all why is it that it switches itself to 802.11g and then again to 802.11a?? I wanted to get your opinion as if do you know any bugs or vulnerabilities by having both radios enable?
And here is my response:
Thank you for your message. Unfortunately, at this time most/all of the roaming decisions are made by the stations (laptops, etc.) and not the infrastructure (Access Points and WLAN Controllers). The IEEE is devising a standard to change this, but it will take some time.
What type of stations are you using? Are they laptops, or something else such as handheld scanners, Voice-over-WiFi phones, etc? If they are laptops, are they Windows or Macintosh? If you are using Windows based laptops, you may be able to set a preference for 5 GHz within the client driver. For example I have an Intel 4965 AGN adapter — if I right-click on my wireless adapter and select “properties”, then click “configure” I can select the Wireless mode to be 802.11a/n only. I can also set the roaming aggressiveness so that it roams less frequently. Macintosh computers have a natural preference for 2.4 GHz and it is more difficult to encourage them to connect to something at 5 GHz – in a case like that, you may consider adding a 5 GHz SSID on its own VLAN. Again, these roaming decisions are made by the station, so your best bet is to look at the laptop settings to see what you can tweak.
One other potential “gotcha” that came to mind was how you enabled 802.11g only. The data rates have to be carefully managed for full interoperability between the station and the infrastructure. There are two types of data rates – the “basic” rates and the “supported” rates. The basic rates should include all 802.11g rates, even below 24 Mbps (therefore you should double-check that the basic rates include 6,9,12, and 18 Mbps). The requirements for supported rates are less strict.
What makes up a wireless LAN (WLAN)? Is it just the Access Point(s) and any associated WLAN controller(s)? Does it include anything else? Personally, I think of wireless LANs as a system. In my mind, there are three main parts to this system:
- The infrastructure
- The clients
- The environment
This includes not only the access point(s) and the WLAN controller(s), but any of the routing & switching gear used for uplinks. Also included in the infrastructure would be any of the network infrastructure / services leveraged by the wireless LAN (examples include RADIUS, Network Time Protocol, Syslog, etc.)
The IEEE 802.11 specification refers to these as “stations”. Clients/stations could be computers such as laptops or tablet PCs. Other examples of wireless stations include dual-mode phones, handheld scanners, etc. When thinking about the clients, realize this includes not only the hardware, but the software (such as a client driver or management supplicant).
This most commonly refers to the Radio Frequency environment. The RF environment can be evaluated with a spectrum analyzer, which can show you the signal level of your APs, the noise floor, and any interference in the area. Interference can come from other Wi-Fi devices, or non Wi-Fi devices (such as microwave ovens, cordless phones, and Bluetooth). I also think of the Physical environment, which is the type of construction materials used, the amount of vegitation in the area, the terrain/line-of-sight, etc.
I often get called out in troubleshooting situations. In many of these cases, the infrastructure has been evaluated as a potential source for issues (for example, are the authentication/encryption settings correct), but the clients and the environment have been neglected.
In these troubleshooting scenarios, it is imperative to take a thorough look at all three components. Not only should you take a look at the configuration file of the infrastructure, but ensure that wireless clients have the most up-to-date drivers. Look to see that clients power-save and roaming agressiveness settings are appropriate for the application. Double-check to see there is no interference from Wi-Fi or non Wi-Fi sources, etc. In short, evaluate your WLAN as a system.
As you *may* have noticed, I took a hiatus from blogging over the summer. Now, it’s time to get back to work! “But Douglas”, you say, “It’s November. Summer was over long ago.” To that, I will remind you that I live in Phoenix. The high temperature is 94 degrees today. So in that sense, summer isn’t *entirely* over for me, but I’ll start back with blogging anyway…
Yesterday, I wrote a post for Computerworld titled: “It’s Not About ‘Full Bars’, Stupid”. It actually combines two of my passions: NFL football and wireless. The wireless portion discusses wireless’ hidden weakness: network congestion.
I don’t want to put any “spolier” information here, so you’ll just have to head over to Computerworld Blogs and check it out for yourself.
This is a guest post that originally appeared on CWNP.com – the post generated a LOT of comments, so I thought I would pull it out of “the valut” and re-publish here at WiFiJedi.com
I recently ran a poll on my blog (WiFiJedi.com) about frequency band utilization for 802.11n deployments. Here are the results:
In what frequency band do you plan to deploy 802.11n?
- 2.4 GHz = 17%
- 5 GHz = 12%
- Both (2.4 GHz and 5 GHz) = 18%
- Not sure – why does it matter? = 8%
- No plans to deploy 802.11n = 45%
Personally, I found the results surprising for two reasons. First, I was surprised by the large number of respondents who said they have no plans to deploy 802.11n. I wonder what factors are keeping them from deploying 802.11n? Price? Security? Reliability? Scalability?
Second, I was surprised by the low number of respondents who chose a pure 5 GHz 802.11n deployment. I believe there are numerous advantages to deploying WLANs in the 5 GHz band, especially when it comes to 802.11n.
Consider the following:
Number of Available Channels
There are only three non-overlapping channels in the 2.4 GHz frequency band. Channels 1, 6, and 11. There are 23 non-overlapping channels between the 5GHz lower, middle, and upper bands.
Because there are many more non-overlapping channels in the 5 GHz range, it can deliver greater total capacity. 802.11g networks offer 54 Mbps of capacity on each of the three non-overlapping channels in the 2.4 GHz spectrum. This equates to a total capacity of 162 Mbps. 802.11a has the same speed, 54 Mbps, but offers a total of 1.24 Gbps of capacity across its 23 non-overlapping channels. This holds true for 802.11n networks as well. With speeds of 150 Mbps per channel, there are 450 Mbps of 802.11n capacity with 2.4 GHz use and 3.45 Gbps of capacity with 5 GHz use.
The 2.4 GHz frequency band is crowded with interfering devices. Other Wi-Fi access points, microwave ovens, cordless phones, Bluetooth devices, baby monitors, etc. all make for a noisy environment. This degrades the overall signal-to-noise ratio (SNR). The 5 GHz band is considerably cleaner in most areas – see for yourself with a spectrum analyzer!
With nearly 8 times the number of channels to chose from in the 5 GHz band, planning is far simpler than in the 2.4 GHz band. I realize that most enterprise grade solutions have some sort of auto-channel or automatic radio management feature to assist with this, but co-channel interference remains a concern, especially in tight spaces or high-density environments. The additional choices in 5 GHz minimizes the possibility that two APs will utilize the same channel in the same areas.
Sure, some organizations need to support legacy devices in the 2.4 GHz spectrum. However, I don’t think it makes sense to deploy 802.11n networks in this band. One of the main technical improvements available in 802.11n is channel bonding. There is only space for one bonded channel in the 2.4 GHz band which, if utilized, would increase the probability of co-channel interference and make channel planning even more complex.
Last week, at the Gartner Mobile & Wireless Summit in Chicago, Michael King and Timothy Zimmerman gave a presentation on Next-Gen WLANs. In their presentation, they said that 802.11n networks are faster, cheaper, more secure, more reliable, and better managed than the wired infrastructure deployed in most enterprises today. Additionally, they said enterprises should align networking investments to an all-wireless office. I agree with them. But to maximize your chance of success, do it in 5 GHz.
They are so sleek. Their simple, single radio design with an omni-directional antenna seems so practical. Tunneling all the data from each and every AP back to a centralized WLAN controller seems so… secure.
Not! April Fool’s! 🙂
I couldn’t let the day slip away without some foolishness, which I decided to publish on my blog.
Here’s the truth about thin APs:
- Single or dual radio designs severely limit the capacity of your WLAN. In a multi-radio array, each radio operates on its own channel, which offers its own capacity. Remember, wireless is a half-duplex, shared medium.
- Omni-directional antennas limit the coverage of a particular AP. The FCC limits output power. It is more effective to use directional antennas that take the same amount of output power and focus it in a given direction, thereby increasing the range. An array of radios, each tied to its own directional antennas can be arranged to provide 360 degree coverage. This is exactly how modern cell phone towers operate.
- Centralized architectures create a potential bottleneck in the WLAN, especially with the advent of 802.11n and its increased capacity. Distributed architectures allow for more efficient process of the data. Management traffic is still split out and handled centrally. The WLAN industry is replicating the same lessons learned from the wired network world. At first, all the intelligence in a wired network was at the core router. Then, wired networking matured and the intelligence was pushed out to edge switches for better performance.
At any rate, Happy April Fool’s Day!