Wireless interference can take many forms. From the somewhat obvious cordless and conference phones to fluorescent...
lighting, motion detectors and wireless cameras, a host of devices out there can get in the way of Wi-Fi connectivity and sap the spectrum.
At last month's Burton Group Catalyst Conference in San Francisco, Neil Diener, CTO and co-founder of Cognio, maker of wireless spectrum analysis tools, presented "Understanding the Impact of RF Interference on 802.11 Wireless LAN Performance," to help enterprises understand and avoid interference that can limit connectivity in a WLAN environment.
Since Wi-Fi operates in unlicensed bands, the enterprise wireless spectrum can be shared by other, seemingly innocuous devices like Bluetooth and wireless game controllers. But interference from such devices, or from having too many devices on one wireless network, can have a strong impact.
Diener said there are two common types of Wi-Fi interference: co-channel and adjacent channel. Co-channel interference comes from another access point on the same channel that could be next door or sharing the same cell. Adjacent channel interference can come from an access point on a different channel -- for example, when one access point is running on channel 1 while another is on channel 2. Diener said adjacent channel interference can often be worse than co-channel.
Regardless of what type of interference it is, it all has the same results. It affects protocols by creating collisions, retransmissions and rate back-offs. It hurts the network by creating reduced throughput and reduced capacity while introducing latency and jitter. And it creates a user impact -- users call IT wondering why they can't connect, why their connection is slow, or why voice and video are garbled. All of those factors, in turn, affect IT, which gets socked with a huge number of trouble calls and the resulting high support costs.
But Diener said there are ways to avoid interference -- or at least proactively prepare for it.
He said enterprises should perform an initial full RF survey. This should be done once before deployment or as part of a post-deployment coverage sweep. The goal is to discover any pre-existing problems, enable planning around known sources of interference, and establish an RF baseline. Companies can also perform periodic update RF surveys (once a year or as budget allows) to discover any new devices added into the spectrum and compare any deviations from the RF baseline.
Lastly, Diener said, enterprises should establish, publish and enforce a spectrum policy that defines which devices are allowed on the premises and which are not. This gives employees information on the devices that are allowed, and this, coupled with a firm policy and policy enforcement, can keep the RF spectrum clean, he said.
Chip Greene, senior network specialist overseeing both the wireless and wired networks at the University of Richmond, said the campus has more than 600 access points deployed and is in the midst of upgrading to lightweight access points. The deployment covers every point on campus and serves the students, faculty, staff and approved guests.
"Trying to control the airwaves here is a little difficult," he said.
For Greene, the major interference issues arise with 2.4 GHz cordless phones and wireless conference phones. In one instance, an audio-visual communication device that controls projectors was also guilty of some nasty interference.
In one case, a wireless conference phone in a room below a faculty member's office was tweaking the signal-to-noise ratio, creating intermittent interference and connectivity problems every time the phone was in use, despite the fact that the office was smack between two access points.
"Troubleshooting that was quite fun, because it was so intermittent," Greene said.
Greene uses spectrum analysers to look at the RF spectrum and laptop analysers to track down the interference culprits. With the conference phones, which cost the telecommunications department a good amount of money, Greene was forced to add more access points to the network to compensate for the noise, he said, despite being reluctant to do so.
Overall, however, interference runs the fine line between fun and frustrating, Greene said. It's fun to go out and troubleshoot the interference and investigate its cause, he said, but it can be frustrating when connectivity is compromised and the spectrum suffers.
Greene said he advises companies dealing with interference to "be patient, because RF is a very dynamic environment, it's very fluid. It takes a lot of looking around and asking questions."
Reducing interference could be as simple as removing an offending microwave oven or disallowing the use of Bluetooth within enterprise walls, Diener said. Or companies could replace cordless phones and the like with similar devices that don't interfere. Other options include changing access point channels, increasing the transmission power of access points or controlling transmission data rates.
More extreme options include shielding interference with Faraday cages, grounded shields, RF absorbing paints or window shielding -- but those options can get expensive.
"Companies have to be aware that [interference is] a significant problem," Diener said. "You can't pre-plan for interference, but you can do a sweep to make sure the reality meets what you simulated."
One company Diener worked with, a hospital, had a patient system that used barcode scanners for medications. Those scanners used Bluetooth and were interfering with the Wi-Fi signal. The company was able to work with the scanner manufacturer to have the Bluetooth frequency scaled back, eliminating the interfering noise on the WLAN.
"Whether or not they know it, they have interference," Diener said. "It may be anywhere and it may not be 1,000 devices, but it's there. It's a latent problem. It's like heart disease. You have it whether you know it or not, but you want to do something about it to avoid a heart attack."