Enterprise video use is exploding, creating an unending need for bandwidth on the wireless access network. In response, many businesses have invested in fast but expensive 802.11n wireless LANs. So talk of reinvesting in even faster 802.11ac gigabit wireless networks is unnerving, and many wonder if it's really necessary.
The 802.11ac standard, which the IEEE isn't expected to ratify until 2013, will double the throughput of 802.11n. It's also expected to improve performance of video and other media-rich applications because it will include multi-user, multiple input, multiple output technology (MU-MIMO). MU-MIMO enables access points (APs) to send multiple streams of data to multiple clients simultaneously, while current Wi-Fi networks only allow single-client access.
But 802.11ac implementation will require a costly rip-and-replace, and for now, 802.11n can still keep up with even the heaviest of video use. This leaves network managers wondering if they really need to make the investment.
With AC, you can provide up to 160 MHz channels, so you could get up to 600 Mbps per second, per channel, and you can bond eight of those channels. It's extremely high raw throughput.
director of product marketing, Aruba Networks
In this interview, Robert Fenstermacher, director of product marketing at Sunnyvale, Calif.-based Aruba Networks and who serves on the IEEE 802.11ac committee, explains the highlights and drawbacks of 802.11ac gigabit wireless.
What is the difference in the 802.11ac and 802.11n standards? Is there really a need to invest in new Wi-Fi technology at this point?
Robert Fenstermacher: With 802.11ac, you get a lot more overall throughput because you have the ability to combine more spectrum. When we went from 802.11a to 802.11n, we went from 20 MHz channels to 40 MHz channels, so it got you to 150 Mbps per second, per strain. With 11.n, you can bond up to four streams, which gives you 600 Mbps per second, per radio. Most wireless vendors today provide an AP that bonds three streams, which is 450 Mbps per second, per radio. That's a lot of bandwidth.
With AC, you can provide up to 160 MHz channels, so you could get up to 600 Mbps per second, per channel, and you can bond eight of those channels. It's extremely high raw throughput. I think the expectation of the initial silicon and APs will be around a gigabit of throughput. That's more than double the bandwidth and throughput of current 802.11n access points.
But there are other factors to take into consideration. The most obvious one is that this is going to be client dependent, so you would need to have 802.11ac-capable smartphones and tablets and laptops in order to support that higher throughput. It's not like you put in an 802.11ac AP and all of a sudden all your problems are solved. If I have an iPhone 4, it’s going to have the same throughput as it does today, because it has an 802.11n chipset.
I've heard that 802.11ac will have benefits for video optimization that go beyond greater bandwidth.
Fenstermacher: We have clients today that are supporting IP television on demand across large college campuses. Their highest throughput requirement is around 4 Mbps per second, and 802.11n does a very good job. But 802.11ac helps as you have more density of clients. If you have a 1000-seat auditorium where you want to stream video to every student, this will help.
When you have faster 802.11ac clients within a dense environment, they can get on and off the airspace more quickly. It basically allows more available capacity for all of those users.
More on 802.11ac gigabit wireless networks
Understanding 802.11ac gigabit wireless: A primer
Gigabit wireless migration: A planning guide
Cisco 802.11ac gigabit wireless available next year
But the most important component is multi-user MIMO, assuming that it gets put into the standard.
Until now, Wi-Fi has been a shared medium, so only one person is accessing the air at any given moment. With MIMO, you can send the same data stream to an individual client simultaneously, and that helps with reliability. Now, with multi-user MIMO, you can use that technology to send different data streams to multiple clients at once. It's almost like moving toward a switch-like environment over the air that you have never had before with Wi-Fi. This will help with density, too.
Will bring-your-own-device (BYOD) programs drive 802.11ac deployment?
Fenstermacher: As we are seeing far more devices come into the enterprise -- not just corporate-issued, but employee-owned -- you will see density drive 802.11ac deployment. If you look at the types of devices that employees are bringing in on their own, they're optimized for multimedia, and they support much higher resolution video. This will also drive the need for 802.11ac over time.
What other challenges will users face with 802.11ac?
Fenstermacher: Multicast is a challenge for both 802.11n and 802.11ac. With multicast, you send the same video stream simultaneously to multiple devices. This is common for handling live video feeds (for example). That way, if everyone is watching a sporting event, each person receives that same stream. While there are a lot of efficiencies that can be gained by not duplicating video streams, the problem with sending multicast over wireless is that the 802.11 standard dictates that you select the lowest available data rate.
There are some things that vendors have done to improve that problem. Aruba, for example, maintains the multicast stream all the way down to the access point. Then, when we send that video stream from the access point to the client, we convert it to unicast, which gives us all of the benefits of a much higher data rate over the air.
The other challenge is [ensuring] you have enough wired capacity to handle a gigabit wireless network. If you have an enterprise that hasn’t upgraded its wired infrastructure, it will become the bottleneck. A lot of enterprises that built out wireless in the past as a convenience network didn’t plan to handle this kind of capacity.