Posts Tagged ‘MIMO’

Wireless networks that follow you around a room, optimize themselves and even talk to each other out loud

Tuesday, October 8th, 2013

Graduate students at the MIT Computer Science and Artificial Intelligence Laboratory showed off their latest research at the university’s Wireless retreat on Monday, outlining software-defined MIMO, machine-generated TCP optimization, and a localized wireless networking technique that works through sound.

Swarun Kumar’s presentation on OpenRF – a Wi-Fi architecture designed to allow multiple access points to avoid mutual interference and focus signals on active clients – detailed how commodity hardware can be used to take advantage of features otherwise restricted to more specialized devices.

There were several constraints in the 802.11n wireless standard that had to be overcome, Kumar said, including a limitation on the total number of bits per subcarrier signal that could be manipulated, as well as restricting that manipulation to one out of every two such signals.

Simply disabling the Carrier Sense restrictions, however, proved an incomplete solution.

“Access points often send these beacon packets, which are meant for all clients in a network … you cannot null them at any point if you’re a client. Unfortunately, these packets will now collide” in the absence of Carrier Sense, he said.

The solution – which involved two separate transmit queues – enabled OpenRF to automatically apply its optimal settings across multiple access points, distributing the computational workload across the access points, rather than having to rely on a beefy central controller.

Kumar said the system can boost TCP throughput by a factor of 1.6 compared to bare-bones 802.11n.

*

Keith Winstein attacked the problem of TCP throughput slightly differently, however. Using a specialized algorithm called Remy – into which users can simply input network parameters and desired performance standards – he said that networks can essentially determine the best ways to configure themselves on their own.

“So these are the inputs, and the output is a congestion control algorithm,” he said. “Now this is not an easy process – this is replacing a human protocol designer. Now, it costs like $10 to get a new protocol on Amazon EC2.”

Remy works via the heuristic principle of concentrating its efforts on the use cases where a small change in the rules results in a major change in the outcome, allowing it to optimize effectively and to shift gears quickly if network conditions change.

“Computer generated end-to-end algorithms can actually outperform human generated in-network algorithms, and in addition, human generated end-to-end algorithms,” said Winstein.

Even though Remy wasn’t designed or optimized to handle wireless networks, it still handily outperforms human-generated competition, he added.

*

Peter Iannucci is a researcher looking into highly localized ways of providing wireless Internet, which he refers to as room area networks. Having dismissed a number of technologies as insufficient – Bluetooth was too clunky, NFC had limited uptake – he eventually settled on sound.

Iannucci’s acoustic network – which he has dubbed Blurt – uses high-frequency sounds to transmit the ones and zeroes of a network connection. It’s well-suited for a network confined by design to a small space.

“Acoustic networks provide great low-leakage properties, since doors and walls are intentionally sound-absorbent,” he said. “[They] work over moderate distances, using existing devices, and they don’t require any setup for ad hoc communications.”

Iannucci acknowledges that Blurt isn’t without its problems. Given that sound waves move about a million times slower than radio waves, speed is an issue – he said that Blurt can handle about 200 bits per second when using frequencies inaudible to humans, with more speed possible only at the cost of an audible whirring chirp, reminiscent of old telephone modems.

But that’s really not the point – the idea would be more to do things like verify users of a business’ free Wi-Fi are actually sitting in the restaurant, or any other tasks involving heavily location-dependent network services.

Source: networkworld.com

802.11ac ‘gigabit Wi-Fi’ starts to show potential, limits

Monday, October 7th, 2013

Vendor tests and very early 802.11ac customers provide a reality check on “gigabit Wi-Fi” but also confirm much of its promise.

Vendors have been testing their 11ac products for months, yielding data that show how 11ac performs and what variables can affect performance. Some of the tests are under ideal laboratory-style conditions; others involve actual or simulated production networks. Among the results: consistent 400M to 800Mbps throughput for 11ac clients in best-case situations, higher throughput as range increases compared to 11n, more clients serviced by each access point, and a boost in performance for existing 11n clients.

Wireless LAN vendors are stepping up product introductions, and all of them are coming out with products, among them Aerohive, Aruba Networks, Cisco (including its Meraki cloud-based offering), Meru, Motorola Solutions, Ruckus, Ubiquiti, and Xirrus.

The IEEE 802.11ac standard does several things to triple the throughput of 11n. It builds on some of the technologies introduced in 802.11n; makes mandatory some 11n options; offers several ways to dramatically boost Wi-Fi throughput; and works solely in the under-used 5GHz band.

It’s a potent combination. “We are seeing over 800Mbps on the new Apple 11ac-equipped Macbook Air laptops, and 400Mbps on the 11ac phones, such as the new Samsung Galaxy S4, that [currently] make up the bulk of 11ac devices on campus,” says Mike Davis, systems programmer, University of Delaware, Newark, Delaware.

A long-time Aruba Networks WLAN customer, the university has installed 3,700 of Aruba’s new 11ac access points on campus this summer, in a new engineering building, two new dorms, and some large auditoriums. Currently, there are on average about 80 11ac clients online with a peak of 100, out of some 24,000 Wi-Fi clients on campus.

The 11ac network seems to bear up under load. “In a limited test with an 11ac Macbook Air, I was able to sustain 400Mbps on an 11ac access point that was loaded with over 120 clients at the time,” says Davis. Not all of the clients were “data hungry,” but the results showed “that the new 11ac access points could still supply better-than-11n data rates while servicing more clients than before,” Davis says.

The maximum data rates for 11ac are highly dependent on several variables. One is whether the 11ac radios are using 80 Mhz-wide channels (11n got much of its throughput boost by being able to use 40 MHz channels). Another is whether the radios are able to use the 256 QAM modulation scheme, compared to the 64 QAM for 11n. Both of these depend on how close the 11ac clients are to the access point. Too far, and the radios “step down” to narrower channels and lower modulations.

Another variable is the number of “spatial streams,” a technology introduced with 11n, supported by the client and access point radios. Chart #1, “802.11ac performance based on spatial streams,” shows the download throughput performance.

802.11ac

In perfect conditions, close to the access point, a three-stream 11ac radio can achieve the maximum raw data rate of 1.3Gbps. But no users will actually realize that in terms of useable throughput.

“Typically, if the client is close to the access point, you can expect to lose about 40% of the overall raw bit rate due to protocol overhead – acknowledgements, setup, beaconing and so on,” says Mathew Gast, director of product management, for Aerohive Networks, which just announced its first 11ac products, the AP370 and AP390. Aerohive incorporates controller functions in a distributed access point architecture and provides a cloud-based management interface for IT groups.

“A single [11ac] client that’s very close to the access point in ideal conditions gets very good speed,” says Gast. “But that doesn’t reflect reality: you have electronic ‘noise,’ multiple contending clients, the presence of 11n clients. In some cases, the [11ac] speeds might not be much higher than 11n.”

A third key variable is the number of spatial streams, supported by both access points and clients. Most of the new 11ac access points will support three streams, usually with three transmit and three receive antennas. But clients will vary. At the University of Delaware, the new Macbook Air laptops support two streams; but the new Samsung Galaxy S4 and HTC One phones support one stream, via Broadcom’s BCM4335 11ac chipset.

Tests by Broadcom found that a single 11n data stream over a 40 MHz channel can deliver up to 60Mbps. By comparison, single-stream 11ac in an 80 MHz channels is “starting at well over 250Mbps,” says Chris Brown, director of business development for Broadcom’s wireless connectivity unit. Single-stream 11ac will max out at about 433Mbps.

There are some interesting results from these qualities. One is that the throughput at any given distance from the access point is much better in 11ac compared to 11n. “Even at 60 meters, single-stream 11ac outperforms all but the 2×2 11n at 40 MHz,” Brown says.

Another result is that 11ac access points can service a larger number of clients than 11n access points.

“We have replaced several dozen 11n APs with 11ac in a high-density lecture hall, with great success,” says University of Delaware’s Mike Davis. “While we are still restricting the maximum number of clients that can associate with the new APs, we are seeing them maintain client performance even as the client counts almost double from what the previous generation APs could service.”

Other features of 11ac help to sustain these capacity gains. Transmit beam forming (TBF), which was an optional feature in 11n is mandatory and standardized in 11ac. “TBR lets you ‘concentrate’ the RF signal in a specific direction, for a specific client,” says Mark Jordan, director, technical marketing engineering, Aruba Networks. “TBF changes the phasing slightly to allow the signals to propagate at a higher effective radio power level. The result is a vastly improved throughput-over-distance.”

A second feature is low density parity check (LDPC), which is a technique to improve the sensitivity of the receiving radio, in effect giving it better “hearing.”

The impact in Wi-Fi networks will be significant. Broadcom did extensive testing in a network set up in an office building, using both 11n and 11ac access points and clients. It specifically tested 11ac data rates and throughput with beam forming and low density parity check switched off and on, according to Brown.

Tests showed that 11ac connections with both TBR and LDPC turned on, increasingly and dramatically outperformed 11n – and even 11ac with both features turned off – as the distance between client and access point increased. For example, at one test point, an 11n client achieved 32Mbps. At the same point, the 11ac client with TBR and LDPC turned “off,” achieved about the same. But when both were turned “on,” the 11ac client soared to 102Mbps, more than three times the previous throughput.

Aruba found similar results. Its single-stream Galaxy S4 smartphone reached 238Mbps TCP downstream throughput at 15 feet, 235Mbps at 30 feet, and 193Mbps at 75 feet. At 120 feet, it was still 154Mbps. For the same distances upstream the throughput rates were: 235Mbps, 230M, 168M, and 87M.

“We rechecked that several times, to make sure we were doing it right, says Aruba’s Jordan. “We knew we couldn’t get the theoretical maximums. But now, we can support today’s clients with all the data they demand. And we can do it with the certainty of such high rates-at-range that we can come close to guaranteeing a high quality [user] experience.”

There are still other implications with 11ac. Because of the much higher up and down throughput, 11ac mobile devices get on and off the Wi-Fi channel much faster compared to 11n, drawing less power from the battery. The more efficient network use will mean less “energy per bit,” and better battery life.

A related implication is that because this all happens much faster with 11ac, there’s more time for other clients to access the channel. In other words, network capacity increases by up to six times, according to Broadcom’s Brown. “That frees up time for other clients to transmit and receive,” he says.

That improvement can be used to reduce the number of access points covering a given area: in the Broadcom office test area, four Cisco 11n access points provided connectivity. A single 11n access point could replace them, says Brown.

But more likely, IT groups will optimize 11ac networks for capacity, especially as the number of smartphones, tablets, laptops and other gear are outfitted with 11ac radios.

Even 11n clients will see improvement in 11ac networks, as University of Delaware has found.

“The performance of 11n clients on the 11ac APs has probably been the biggest, unexpected benefit,” says Mike Davis. “The 11n clients still make up 80% of the total number of clients and we’ve measured two times the performance of 11n clients on the new 11ac APs over the last generation [11n] APs.”

Wi-Fi uses Ethernet’s carrier sense multiple access with collision detection (CSMA/CD) which essentially checks to see if a channel is being used, and if so, backs off, waits and tries again. “If we’re spending less time on the net, then there’s more airtime available, and so more opportunities for devices to access the media,” says Brown. “More available airtime translates into fewer collisions and backoffs. If an overburdened 11n access point is replaced with an 11ac access point, it will increase the network’s capacity.”

In Aruba’s in-house testing, a Macbook Pro laptop with a three-stream 11n radio was connected to first to the 11n Aruba AP-135, and then to the 11ac AP-225. As shown in Chart #2, “11ac will boost throughput in 11n clients,” the laptop’s performance was vastly better on the 11ac access point, especially as the range increased.

802.11ac

These improvements are part of “wave 1” 11ac. In wave 2, starting perhaps later in 2014, new features will be added to 11ac radios: support four to eight data streams, explicit transmit beam forming, an option for 160 Mhz channels, and “multi-user MIMO,” which lets the access point talk to more than one 11ac client at the same time.

Source:  networkworld.com

Cheat sheet: What you need to know about 802.11ac

Friday, June 21st, 2013

Wi-Fi junkies, people addicted to streaming content, and Ethernet-cable haters are excited. There’s a new Wi-Fi protocol in town, and vendors are starting to push products based on the new standard out the door. It seems like a good time to meet 802.11ac, and see what all the excitement’s about.

What is 802.11ac?

802.11ac is a brand new, soon-to-be-ratified wireless networking standard under the IEEE 802.11 protocol. 802.11ac is the latest in a long line of protocols that started in 1999:

  • 802.11b provides up to 11 Mb/s per radio in the 2.4 GHz spectrum. (1999)
  • 802.11a provides up to 54 Mb/s per radio in the 5 GHz spectrum. (1999)
  • 802.11g provides up to 54 Mb/s per radio in the 2.4 GHz spectrum (2003).
  • 802.11n provides up to 600 Mb/s per radio in the 2.4 GHz and 5.0 GHz spectrum. (2009)
  • 802.11ac provides up to 1000 Mb/s (multi-station) or 500 Mb/s (single-station) in the 5.0 GHz spectrum. (2013?)

802.11ac is a significant jump in technology and data-carrying capabilities. The following slide compares specifications of the 802.11n (current protocol) specifications with the proposed specs for 802.11ac.

(Slide courtesy of Meru Networks)

What is new and improved with 802.11ac?

For those wanting to delve deeper into the inner workings of 802.11ac, this Cisco white paper should satisfy you. For those not so inclined, here’s a short description of each major improvement.

Larger bandwidth channels: Bandwidth channels are part and parcel to spread-spectrum technology. Larger channel sizes are beneficial, because they increase the rate at which data passes between two devices. 802.11n supports 20 MHz and 40 MHz channels. 802.11ac supports 20 MHz channels, 40 MHz channels, 80 MHz channels, and has optional support for 160 MHz channels.

(Slide courtesy of Cisco)

More spatial streams: Spatial streaming is the magic behind MIMO technology, allowing multiple signals to be transmitted simultaneously from one device using different antennas. 802.11n can handle up to four streams where 802.11ac bumps the number up to eight streams.

(Slide courtesy of Aruba)

MU-MIMO: Multi-user MIMO allows a single 802.11ac device to transmit independent data streams to multiple different stations at the same time.

(Slide courtesy of Aruba)

Beamforming: Beamforming is now standard. Nanotechnology allows the antennas and controlling circuitry to focus the transmitted RF signal only where it is needed, unlike the omnidirectional antennas people are used to.

(Slide courtesy of Altera.)

What’s to like?

It’s been four years since 802.11n was ratified; best guesses have 802.11ac being ratified by the end of 2013. Anticipated improvements are: better software, better radios, better antenna technology, and better packaging.

The improvement that has everyone charged up is the monstrous increase in data throughput. Theoretically, it puts Wi-Fi on par with gigabit wired connections. Even if it doesn’t, tested throughput is leaps and bounds above what 802.11b could muster back in 1999.

Another improvement that should be of interest is Multi-User MIMO. Before MU-MIMO, 802.11 radios could only talk to one client at a time. With MU-MIMO, two or more conversations can happen concurrently, reducing latency.

Source:  techrepublic.com

LTE Advanced is coming, but smartphone users may not care

Thursday, June 20th, 2013

Faster network speeds are better, right?

LTE Advanced is coming soon to the Samsung Galaxy S4 smartphone, offering double the downlink speeds of LTE (Long Term Evolution).

But U.S. carriers still have to upgrade LTE networks to operate with LTE Advanced, and their plans to do so are vague.

Even if U.S. networks were completely LTE Advanced-ready, some analysts question whether buyers would pay much more to upgrade their smartphones for a model with the LTE Advanced speed advantage. There’s unlikely to be the same scramble for LTE Advanced as there was for LTE-ready smartphones such as the iPhone 5, which provide 10Mbps or more on LTE downlinks on average, boosting previous speeds by three times or more over 3G, analysts said.

One analyst is especially skeptical of LTE Advanced’s value. LTE Advanced in a smartphone or tablet “is not important to the user, especially in the U.S. where carriers have been marketing LTE or 4G for years now,” said Carolina Milanesi, an analyst at Gartner. “The novelty has worn off. To tell customers that LTE will be even faster … is nice, but not life changing.”

Jack Gold, an analyst at J. Gold Associates, said consumers don’t understand what LTE Advanced is. “Will users actually see that much improvement? Will they notice anything all that different in their user experience? For most, probably not,” Gold said Tuesday. “Carriers [and manufacturers] are really trying to find advantage to keep the market excited about their networks. Will users buy into it? Remains to be seen.”

JK Shin, co-CEO of Samsung Electronics, said Monday that the Galaxy S4 with LTE Advanced capabilities will go on sale in South Korea in June and with wireless carriers in other countries later on. He said that a three-minute download of a movie using existing LTE technology would take just over a minute to download over LTE Advanced.

Of the four largest U.S. wireless carriers polled on Monday, T-Mobile USA said it was planning on LTE Advanced network support, although it has not announced a schedule.

Verizon said it plans three network improvements that will support LTE Advanced, some of which will be ready and “invisible” to customers later this year.

Sprint said it has already deployed some elements of LTE Advanced, but didn’t elaborate on a schedule or other details. The carrier said LTE Advanced will give customers greater speed, capacity and improvements in video quality, and will help lower Sprint’s costs to keep unlimited data plans.

AT&T is also expected to move to LTE Advanced, but didn’t respond immediately to questions about timing.

The upgrade cost to deliver LTE Advanced is expected to be minor compared to the many billions of dollars it cost to upgrade networks to LTE with new antennas and switches.

Analysts said that while Samsung will introduce the GS4 in South Korea on LTE Advanced networks, the value of LTE Advanced could be far less exciting in the U.S. Smartphones are getting wide acceptance in the U.S., where a majority of Americans own the devices, according to a Pew Research Center survey. Smartphone makers and carriers face the challenge of marketing new features, such as faster network speeds, to lure buyers into trading in their old devices for new ones.

“We are getting to the point where selling innovation is hard,” Milanesi said. “That innovation today is about user experience, convenience and incremental benefits — not transformational ones.”

Such incremental benefits are hard to sell in a store because the features take longer for salespeople to demonstrate, she added. The sales rep might be showing the improvements on “a phone that otherwise will look like all the rest — or even worse — like the previous generation,” Milanesi said.

Gartner and other analyst firms noticed a new trend that started in the fourth quarter of 2012: U.S. smartphone owners were keeping their devices longer, holding them beyond a two-year contract rather than upgrading before the end of the two-year period to get access to newer hardware, software or network speeds.

“Some users might hang onto their smartphones and get a tablet instead,” Milanesi added.

As a result of both lower perceived innovation in new smartphones and the hype to buy tablets, smartphone lifetimes are lengthening, Milanesi said. “That means for a market like the U.S. where we have a replacement market, [sales] growth will decrease,” she said.

At Verizon, LTE Advanced is viewed as an improvement that will be invisible to customers, Verizon spokesman Tom Pica said. Verizon already has rolled out LTE to more than 400 cities, and LTE Advanced will mean customers “continue to find the consistent reliability and high speeds they have come to expect” from Verizon, he said.

Later in 2013, Verizon will deploy small cells and AWS (Advanced Wireless Services) spectrum as part of LTE Advanced capabilities. A third step, involving advanced MIMO (multiple input, multiple output) antennas for devices and cell sites, is in the plans, but no schedule has been announced.

AWS uses two spectrum bands in the 1700MHz and 2100MHz channels to increase network capacity for heavy data users but not necessarily speed. Verizon already sells seven devices that support AWS, including the GS4, the Nokia Lumia 928 and the BlackBerry Q10. Small cells can increase network capacity and network reach.

Source:  computerworld.com

Corning taps into optical fiber for better indoor wireless

Monday, May 20th, 2013

Bringing wireless indoors, which was once just a matter of antennas carrying a few cellular bands so people could get phone calls, has grown far more complex and demanding in the age of Wi-Fi, multiple radio bands and more powerful antennas.

DAS (distributed antenna systems) using coaxial cable have been the main solution to the problem, but they now face some limitations. To address them, Corning will introduce a DAS at this week’s CTIA Wireless trade show in Las Vegas that uses fiber instead of coax all the way from the remote cell antennas to the base station in the heart of a building.

Cable-based DAS hasn’t kept up with the new world, according to the optical networking vendor. Though Corning is associated more often with clear glass than with thin air, it entered the indoor wireless business in 2011 by buying DAS maker MobileAccess. That’s because Corning thinks optical fiber is the key to bringing more mobile capacity and coverage inside.

The system, called Corning Optical Network Evolution (ONE) Wireless Platform, can take the place of a DAS based fully or partly on coaxial cable, according to Bill Cune, vice president of strategy for Corning MobileAccess. Corning ONE will let mobile carriers, enterprises or building owners set up a neutral-host DAS for multiple carriers using many different frequencies.

Though small cells are starting to take its place in some buildings, DAS still has advantages over the newer technology, according to analyst Peter Jarich of Current Analysis. It can be easier to upgrade because only the antennas are distributed, so more of the changes can be carried out on centralized gear. Also, small cells are typically deployed by one mobile operator, and serving customers of other carriers has to be done through roaming agreements, he said.

However, some DAS products based on coaxial cable are limited in how they can handle high frequencies and MIMO (multiple-in, multiple-out) antennas, Jarich said. Some vendors are already promoting fiber for greater flexibility and capacity, he said.

Going all fiber — up to the wireless edge, at least — will make it easier and cheaper for indoor network operators to roll out systems that can deliver all the performance users have come to expect from wireless networks, according to Corning. That includes more easily adding coverage for more carriers, as well as feeding power and data to powerful Wi-Fi systems that can supplement cellular data service, the company says.

Wireless signals don’t travel the same way inside buildings as they do outdoors, so one antenna can’t always cover the interior, regardless of whether it’s mounted in the building or on a nearby tower. A DAS consists of many antennas spaced throughout a structure, all linked to a base station in a central location. Most types of DAS use coaxial cable to carry radio signals in from the distributed antennas.

However, those copper cables get more “lossy” as the frequencies they have to carry get higher, meaning they lose a lot of their signal on the way to the base station, Corning’s Cune said. That has left coax behind as new frequencies are adopted, he said. For example, coax isn’t good at carrying the 5GHz band, which is crucial in newer Wi-Fi equipment, Cune said.

MIMO, a technology that uses multiple antennas in one unit to carry separate “streams” over the same frequency, is another big limitation of DAS, according to Corning. MIMO antennas for better performance can be found in newer Wi-Fi gear based on IEEE 802.11n and 802.11ac, as well as in LTE. A coax-based DAS with MIMO antennas needs to have a separate half-inch-wide cable for every stream, which is a major cabling challenge, Cune said.

Corning ONE links each antenna to the base station over optical fiber, converting the radio signals to optical wavelengths until they reach the base station. Fiber has more capacity than coax, can handle higher frequencies, and requires just one cable from a MIMO antenna, Cune said. Because of fiber’s high capacity, it’s relatively easy to bring other mobile operators onto the DAS.

The system is based on optical fiber, but it can be extended over standard Ethernet wiring to provide backhaul for Wi-Fi access points. Each Corning ONE remote antenna unit that’s deployed around a building will have two Ethernet ports to hook up nearby Wi-Fi access points, which can use the fiber infrastructure for data transport to wired LAN equipment, Cune said.

Corning ONE is in beta testing at one enterprise and will have limited availability beginning in late June, after which orders can be placed, Cune said. It is expected to be generally available two to three months later. The company expects its main customers to be mobile operators, though most of those operators will arrange multi-carrier services, he said. Enterprises and large building owners increasingly will step in to buy and deploy the DAS, Cune said.

Source:  networkworld.com

Super Bowl plans to handle 30,000 Wi-Fi users at once—and sniff out “rogue devices”

Saturday, February 2nd, 2013

When 73,208 fans file into the New Orleans Superdome for the Super Bowl on Sunday, they’ll have to follow the usual rules: no booze, no weapons, no fireworks, and no food (though food and beer can be purchased inside the stadium at exorbitant rates).

They’ll also be prevented from bringing in any wireless equipment that might interfere with the proper workings of the Superdome Wi-Fi network. Lots of time and money went into giving ticket holders a wireless connection that rivals the one in their living rooms, and the NFL doesn’t want anyone messing it up.

“The NFL has a very robust frequency coordination solution in place,” Dave Stewart, director of IT and production for Superdome management firm SMG, told me in a phone interview. “Every device that enters the building has to go through a frequency scan and be authorized to enter. At the perimeter the devices are identified and tagged. If they present a potential for interference, they are remediated at that moment. Either the channel is changed or it is denied access. It’s all stopped at the perimeter for this event.”

In Stewart’s words, the goal is to prevent any “rogue access points or rogue equipment from attempting to operate in the same frequency” as the stadium Wi-Fi network (“rogue” as in “not under the control of the system administrators”).

It’s hard to imagine fans, press, or stadium staff deliberately trying to sabotage Super Bowl Wi-Fi, but some may do so unintentionally. Interference can be produced by “everything from someone operating a network wireless camera to someone operating pyrotechnics equipment that utilizes wireless service to trigger their equipment,” Stewart said. “Imagine if you were to bring in a wireless camera and that wireless camera is tuned to the 2.4GHz frequency range [also used by Wi-Fi] and is continually broadcasting a signal. Anything that’s going to operate in the same frequency range has the potential to cause interference. Some of those interfering devices are minimal, but others are impactful.”

The biggest concern, he said, comes from “non-Wi-Fi-compliant continuous broadcast devices such as wireless cameras.”

The best defense against such “rogue” wireless networking is to prevent the wrong devices from coming into the stadium, but you can’t stop everything. “You can’t stop a laptop from coming in. Working press needs to use that,” he said. Yet laptops can be problematic if their owners try to create their own private Wi-Fi networks. “Anyone who enters the facility with a laptop has the ability to become a rogue by going to ad hoc [wireless networking] mode,” Stewart said.

That’s why wireless security doesn’t stop when the game starts; the Superdome will use spectral analysis equipment to detect interference. “We’re always monitoring the network. So we have a plan in place if there is an interfering signal to identify that and remediate that problem,” Stewart said.

So if you’re broadcasting a rogue wireless signal, well, you might just get a tap on the shoulder from a Superdome employee. Isn’t it more fun just to watch the game, anyway?

Is the TV timeout over yet?

Well, maybe not. The NFL manages to spread 60 minutes of clock time across three hours in a typical game. What with time running off the clock between plays and the typical play lasting about four seconds, an average game ends up with only 11 minutes of action. And given how long Super Bowl halftime shows last, the game might not be over till Monday.

So fans have plenty of time to check their e-mail, upload pictures to Facebook, or get instant replays and game-related information on their mobile devices. While the NFL’s strict control over wireless equipment might sound draconian, it’s in service of the greater good: Wi-Fi for everyone who wants it.

The Superdome already had one of the most robust cellular networks among football stadiums, because 18 months ago AT&T built a carrier-neutral distributed antenna system on site to boost mobile signals. But that wasn’t enough. Cellular providers want Wi-Fi in places like the Superdome because it offloads traffic from the cellular network, and fans like it because they’re less likely to drop their connections or wait for videos to buffer.

The Superdome (or the Mercedes-Benz Superdome, to get the branding right) previously had Wi-Fi—but only for staff, press, and systems like ticketing. “That was not getting us where the NFL wanted us to be relative to the Super Bowl needs,” said Superdome General Manager Alan Freeman. The new, Super Bowl-scale Wi-Fi network was just put in this season, with trial runs in a couple of late-season Saints games and in the Sugar Bowl. The Super Bowl will be the first time the network is publicly advertised as available to all fans, so the load will be greater. No password will be required to get on the Wi-Fi network.

More than 700 wireless access points will distribute signals inside the Superdome. Another 250 access points will provide Wi-Fi outside the stadium, including in parking lots and in Champions Square. (Another 300 access points are in the adjacent New Orleans Arena, which hosts the city’s pro basketball team.)

During the Super Bowl, the network will be able to handle up to 30,000 simultaneous connections, which should be enough. At last year’s Super Bowl in Indianapolis, Wi-Fi from 604 access points supported 8,260 simultaneous connections at its peak, while 12,946 attendees were on the Wi-Fi at some point during the game. 225GB of data was downloaded and 145GB uploaded, with peaks of 75Mbps down and 42Mbps up. (We’re told the cell network for all carriers at last year’s Super Bowl handled another 560GB of data total, including downloads and uploads.) Usage is expected to be higher this year, but it’s impossible to predict exactly how much it will increase.

Superdome management thinks it’s ready. In testing, “We’re constantly seeing 20Mbps up and 20 megs down in all areas of the building,” Stewart said. “That, of course, will change depending on the load, but the system is backed up by multiple redundant links to the Internet.”

Verizon Wireless built the Wi-Fi network, and all equipment used came from Cisco. The back-end of the network is handled by two Cisco Nexus 7000 Series Switches (with another two in the adjacent arena), Cisco 5500 Series Wireless Controllers, and Cisco’s 5540 Adaptive Security Appliances. Access points are Cisco Aironet 3500 devices.

This isn’t consumer gear (a business Cisco is getting out of); these are high-density access points, designed for stadiums, with directional antennas that send the signals to just the right places. The Superdome has a ceiling, of course, but it’s far above fans’ heads. If antennas weren’t positioned correctly, signals could be wasted in all that empty air. Using directional antennas lets the Superdome “control the signal and have a seamless handoff from one section to another when a fan roams, or when someone comes online,” Stewart noted.

We need more channels!

The network supports 802.11n and previous Wi-Fi protocols 802.11a, b, and g, using both the 2.4GHz and 5GHz bands. Unfortunately, many fans’ devices are capable only of getting on 2.4GHz.

The 2.4GHz band has 11 channels that can be used for Wi-Fi in North America, but because the channels overlap, the Superdome uses just channels 1, 6, and 11. In the 5GHz band, with its 23 non-overlapping channels, the Superdome network can use just about every available channel (while making sure not to interfere with radar).

High-end mobile devices like the iPad, iPhone 5, and Samsung Galaxy S III support 5GHz. Newer versions of the Kindle Fire support 5GHz, too; Google’s Nexus 7 Android tablet does not. Many phones are still stuck on 2.4GHz as well.

“We’re very anxious and can’t wait for everyone to get on 5GHz,” Stewart said. Even better will be when the world moves on to 802.11ac networks and devices, because the next-generation protocol uses multi-user MIMO (multiple-input/multiple-output) to transmit signals more efficiently. But that’s not happening soon. While some home routers support 802.11ac, they’re not NFL caliber.

“There’s no commercially available high-density 802.11ac equipment that I know of,” Stewart said.

In a worst-case scenario, high numbers of fans streaming video could cause congestion and slow down fans’ connections. “This is not unlimited. There’s no such thing,” said Kelley Carr, co-founder of Cellular Specialities, a consultant who helped oversee the design and implementation of the network for the big game.

What they have for the Super Bowl is probably good enough for this year, though. “We’re all confident it will work, just based on our experience in the past,” said Carr. “As long as 100 percent of the people in there don’t take out their cellular device and switch it to the Wi-Fi network, it should be fine.”

With an average signal strength rating of -60dB, fans in their seats should get a signal comparable to what they would have at home if they were sitting about 20 feet from their wireless router, Stewart said.

This will be the seventh Super Bowl to be hosted at the Superdome since 1978—and such access would have been unthinkable in any of the previous games. Freeman noted that “from a technological perspective, these mega-events keep getting more complex, exponentially in some cases, every year.”

Source:  arstechnica.com

Aruba brings Wi-Fi to wall plates

Thursday, January 31st, 2013

The typical Wi-Fi deployment today involves access points deployed in hallways or rooms as standalone boxes. As the move towards pervasive wireless access grows, so too have the demands on wireless infrastructure. That’s where Aruba Networks (NASDAQ:ARUN) is aiming to fill a gap with a new wall mountable access point.

The AP-93H is a 2×2 MIMO 802.11n access point that can be installed on a standard wall mount for wired Ethernet access. The AP-93H has a gigabit uplink port for high-speed connectivity to the wired network for access. The access point is a dual band radio operating in either the 2.4 Ghz or the 5 Ghz ranges. On the software side the device includes the Linux-powered Aruba OS.

Among the target markets for the AP-93H are hotel and dorm room type deployments.

“Over the past few years, the number of mobile devices have really exploded,” Manish Rai, head of Industry Solutions for Aruba, told InternetNews.com. “I think we have reached a tipping point where it makes sense to increase the capacity and move to an in-room deployment for better coverage.”

Source:  wi-fiplanet.com

802.11ac boosts buzz more than bandwidth

Friday, January 27th, 2012

A great innovation, 802.11ac does more for 802.11n than it does to change the face of wireless networking

The buzz about 802.11ac is in full swing, but don’t believe everything you read.

The newest of Wi-Fi innovations, 802.11ac (still in draft form) looks like it will start making it into enterprise Wi-Fi products as early as 2013 and home products even earlier. It’s already being flaunted as Gigabit Wi-Fi. And for the largest Wi-Fi market, the home, it will be. But will it deliver gigabit speeds for the enterprise? Not a chance.

Defined for the capacity-rich 5GHz spectrum (495MHz) only, 802.11ac introduces a number of new techniques like advanced modulation and encoding, multi-user MIMO and channel bonding, that theoretically, if you’re talking to a vendor anyway, has the potential to dramatically increase Wi-Fi capacity. The question is, REALLY?

Make no mistake, 802.11ac is a great innovation. But like any great innovation, the devil is often in the details. So here are some details that should help demystify 802.11ac. Here are the key differences to understand:

* Eight spatial streams. One of the biggest Wi-Fi innovations came with 802.11n in the form of spatial multiplexing using a technique called MIMO (multiple input, multiple output). MIMO is the use of multiple antennas at both the transmitter and receiver to increase data throughput without additional bandwidth or increased transmit power. Basically it spreads the same total transmit power over multiple antennas to achieve more bits per second per hertz of bandwidth with the added benefit of greater reliability due to more antenna diversity.

In essence, MIMO lets an access point send multiple spatial streams to one client at a time to increase capacity. 802.11n specified up to four spatial streams.

Now in glorious one-upmanship, 802.11ac will support up to eight spatial streams. Historically it has taken chip manufacturers about two years to add an additional spatial stream (802.11n is only at three right now). While that will surely improve with 802.11ac, don’t look for it to ever get to eight. However it would be a funny sight to see. Just picture an AP with 12 omni-directional antennas (eight for 11ac in 5GHz and four for 11n in 2.4GHz) sticking out of it. Not a pretty picture.

* Multi-user MIMO. Another difference with 11ac is support for “multi-user MIMO” (MU-MIMO). With 802.11n, MIMO could only be used for a single client at any given time, while 802.11ac tries to improve on this by supporting multiple clients.

This allows an 802.11ac AP to transmit two (or more, depending on the number of radio chains) spatial streams to two or more client devices. This has the potential to be a good improvement but is optional. And it’s expected that the first 802.11ac chips out the door won’t support this. What’s more, there’s is a good chance that MU-MIMO won’t ever be supported due to the radio and MAC complexity required.

* 256 Quadrature Amplitude Modulation (QAM). QAM is a way to modulate radio waves to transmit data. 802.11n maxed out at 64QAM so the advent of 256QAM should deliver big improvements in maximum throughput. However, the more complex a modulation scheme, the more difficult it is to achieve. In realistic situations it is highly unlikely that any percentage of client devices would consistently achieve 256QAM. Ouch.

* 5GHz only. Due to how 11ac really achieves all this speed (channel bonding) it doesn’t make sense for 11ac to support 2.4GHz which only has three (of 11) non-overlapping channels. What this means is that devices that want to have 11ac (they all will) will be 5GHz capable. Right now it is a very low percentage that are capable of 5GHz and that is a real shame. Now they’ll be required to do it.

* Channel bonding. An easy and effective method to increase the speed of any radio link is to give it more frequency, or bandwidth. To get more bandwidth, 802.11n introduced us to channel bonding: the ability to take two 20MHz channels and make them work as one — basically a bigger Wi-Fi pipe. This effectively doubled throughput that could be achieved.

Now 802.11ac has mandated the support of 80MHz channels with options to bond eight channels for a total channel of 160MHz.

Even with 802.11n, channel bonding is a double-edged sword. In North America, the 2.4GHz band has 83.5MHz (three non-overlapping channels) of total bandwidth, while the 5GHz bands have a total of 495MHz. That means that 5GHz can carry almost six times the traffic of 2.4GHz plus the added benefit that (for now) the 5GHz band is a much cleaner spectrum.

But don’t count your bits quite yet. What most people don’t realize is that by enabling channel bonding you are actually reducing your overall capacity (see chart).

When designing and deploying a Wi-Fi network for high density, more channels are preferred to fewer, larger channels. Increasing the number of devices occupying one channel in a given area makes reduces the efficiency of Wi-Fi.

This is why people like wires. Because each device effectively has its own channel and there are no other devices occupying that channel (the copper or fiber). So we see staggering amounts of throughput.

If Wi-Fi could have hundreds of channels, and each client would get their own, this would be wireless nirvana. But, as you can see from that chart, we don’t have that many channels, and we sure don’t want to exacerbate the problem by bonding them together if it reduces the overall efficiency of the wireless LAN.

Using 802.11ac in the home is a different story. Bonding channels all the way to 160MHz is preferred given there are few devices trying to access a single AP.

The enterprise is just the opposite. Here, numerous APs are required to support hundreds or thousands of users. And, as much as is possible, those APs should be on different channels.

Ultimately, 802.11ac offers improvements for the Wi-Fi industry primarily because it forces clients to add support for the capacity-rich 5GHz spectrum. Current enterprise APs already support both bands.

Ironically, 802.11ac will prolong the viability of current 802.11n networks. As more and more clients become 5GHz capable, capacity and performance will increase without touching the infrastructure. This is the best news of all.

Source:  networkworld.com