Why is quality of service (QoS) needed for voice over Wi-Fi (VoWi-Fi)?
Wireless bandwidth is shared by everyone, and whoever gets on the network first gets the bandwidth. Voice packets need low latency, low jitter and some bandwidth. If I have a bunch of different wireless users without QoS, and someone downloads a large file, that can completely overload the wireless system. If users don't get their bandwidth, they have a poor quality call. How is QoS different on a wireless LAN than it is on a wired LAN?
On the wired side, you can [implement] a virtual LAN, effectively allowing you to get bandwidth as you need it. On a wireless LAN, you have to go back to the old Ethernet hub scenario where bandwidth is shared. Also, interference in the wired LAN is minimal, but on the wireless LAN, it can be a significant problem. Microwave ovens can interfere with 802.11b/g systems. Multiple access points using the same channel can cause interference. This can be a problem for businesses in a high-rise office building if another business on a floor above or below is using a wireless network. What kinds of QoS products are on the market today?
Most vendors have implemented a voice queue, where -- from the access point to the phone -- voice traffic is tagged and sent ahead of data traffic. But if you have six or seven phones on the same access point, they can start stomping on each other's signals. Right now, an access point has control over the airwaves. It sends a beacon frame out to tell the device it can start a transmission. The access point is making sure the channel is empty, which works most of the time. But there is no way for the client device to do the same thing in the other direction. Are certain WLAN architectures, such as a wireless switch or gateway that uses a thin access point, better for voice?
All major vendors have a voice queue, which needs to be turned on. If a business has a heterogeneous environment, then it will have to go to a third party vendor.
It makes little difference if a company has fat or thin access points. With fat access points, a client device sends out a broadcast. If the fat or intelligent access point has too many phones already using it, it will not respond and another access point will have to. In a centralized architecture, the thin access point will send that message back to the central controller, which then decides which access point is best. Are there drawbacks to implementing VoWi-Fi now, before standards are ratified?
All of the access points shipping now will be upgradeable to [an 802.11e subset called Wireless Media Extensions (WME)], or 802.11e. The big thing is the increase in processing power that is required at the access point itself.
In the next couple of months, WME be released by the Wi-Fi Alliance. The Wi-Fi Alliance is ahead of the Institute of Electrical and Electronics Engineers (IEEE) in developing a common voice QoS standard for manufacturers. Should businesses wait for the standard to be ratified?
Most access points available today can be upgraded to WME and 802.11e. Businesses can start deploying them today, as they don't need the voice [features] to function right away. They can add the software upgrades as needed.
The real issue is on the client side. Many of the new client radios on [Wi-Fi] phones have the ability to be upgraded. Businesses should make sure that vendors guarantee to do software upgrades. What about those that have already deployed wireless networks?
The first thing you have to do is call the vendor and see if your access points can be upgraded. Some can and some can't. Businesses need to decide if they need to deploy voice-enabled WLANs throughout the enterprise. Most likely they can start upgrading the access points in critical locations. And for smaller deployments of 100 access points or less, they can be replaced without breaking the bank. Another strategy is to just put in a second network for voice using 802.11a, assuming your existing network is 802.11b/g.