RAM | Reseller Advocate Magazine: Issue 78

If a Little is Good,
N is Better
Does The Latest Wireless Keep Its Promises?
By Chris Angelini

I COM FROM A FAMILY OF PEOPLE WHO like to tinker with cooking recipes. A perfectly good dish was always fair game for improvement, and if experiments turned out to weaken the end result, then the problem was always one of insufficient addition, never subtraction. “It needs more of this or that,” Mom and Dad would always say. (Coincidentally, neither one was particularly thin.) The household mantra was “if a little is good, more is better.” Is there a better summation of the American way?

Certainly, this philosophy holds true with wireless bandwidth. Having to move 10GB of data off of a notebook via 802.11b (11 Mb/s theoretical) is a good occasion to get through 200 pages of a novel or take up a new hobby. Going with 802.11g (54 Mb/s theoretical) is better, but only in some ways. Part of the problem has been that wireless networking vendors state unrealistic speeds based on absolutely optimal conditions free from interference—conditions you’re unlikely to find this side of Earth’s moon. No walls, no obstructions, no radio interference, no magnetic interference—nothing. The 11 Mb/s of 802.11b frequently only yielded 4 to 5 Mb/s under normal indoor use. In turn, standard 802.11g systems hit in the 20 to 25 Mbps range. The rule of thumb has been to take the industry’s theoretical specs, divide by half, and then pull off a few more percent for good measure. If a little pessimism about the wireless bandwidth is good, more is better.

Why the Long Face?
Vendors tried to put lipstick on the pig. For a while, it seemed that every major player had its own “turbo” mode, a scheme whereby two 2.4 GHz wireless channels carried separate data streams, thus approximately doubling the throughput. Real-world measurements often landed in the 50 to 60 Mb/s range. Going back to our 10GB transfer, that’s roughly 10,000MB, or 80,000 megabits (Mb). Eighty thousand divided by even 60 Mb/s is still more than 22 minutes of transfer time. Did I mention that different wireless chipsets from Atheros and Broadcome frequently resulted in incompatible turbo modes when mixing brands?

To make matters worse, in the approximate words of Haley Joel Osmet: “I see dead spots. Lots of them.” I’ve now run Wi-Fi in two houses, never mind in friends’ homes, and it seems absurd that a technology with a rated distance of 300 feet can’t survive the 80 feet and few walls from my office to my spot on the bed. Not my wife’s spot—that works fine. It’s my spot four feet over that gets the terrible reception. If it doesn’t exist already, I’m going to create a maxim here and now that says, “Signal strength at any given physical point is inversely proportional to the importance of that point’s reception performance in everyday life.” You can quote me.

Next up to bat in 2005 were the myriad vendor interpretations of MIMO (multiple-input multiple-output), often wrapped up in the pseudo-standard of “pre-N.” MIMO uses multiple antennas to capitalize on a radio phenomenon known as multipath propagation. As one radio signal moves from point A to point B, it’s likely to encounter objects along the way, which will result in the signal skewing. Different parts of the signal will reach the destination at slightly different times. MIMO turns these skews into new data paths, essentially cramming more data streams into the same transmission provided there are two or, preferably, three antennas at each end. With better signal reception, theoretical speeds of up to 240 Mb/s, and fewer dead spots, MIMO seemed a sure-fire winner. Unfortunately, between incompatible chipsets (how the industry loves a good format war) and the looming specter of real 802.11n on the horizon, first-gen MIMO products proved so popular that you couldn’t sell one on eBay today if you tried.

Skip ahead to the arrival in mid-2006 of Draft-N, the follow-up to pre-N that was supposedly so in tune with IEEE specs that it had to be forward-compatible with whatever minor variance the standards body finally ratified. Well, I personally tested several of these Draft-N products, and they were almost universally terrible. Interoperability was poor, and while Draft-N carried an official throughput rating of 300 Mb/s, my real-world tests showed speeds in the 50 to 70 Mb/s range. (Second-gen BIOS updates pushed this up into 80+ Mb/s territory, but the days of the 50%-minus-a-pinch performance rule were clearly over.) In other words, I was seeing numbers even worse than those I’d measured with 802.11g turbo gear, and reception coverage, while better than traditional Wi-Fi, was far from the building-wide panacea I’d expected. To all you resellers who pitched and deployed these Draft-N products based on vendor promises, you have my every sympathy. It was a disaster.

Buyers got hosed. Everyone was angry. The blame game led to the network hardware vendors, who in turn got to blame the chipset makers and all that pesky customer demand. Oh, and the industry consortium behind 802.11n, which seemed to be mired in a sludge of sodden red tape. At long last, a lot of consumers and businesses threw up their hands and decided that yesteryear’s 802.11 a/b/g was good enough, and where that ran out, a few more CAT5e drops could fill in. Draft-N had turned into Draft-Not.

11n Inside
Intel’s Next-Gen Wireless-N (4965AGN) notebook adapter is a PCIe Mini Card adapter with three antenna ports for both the 2.4 GHz and 5.0 GHz bands. This enables not only Draft-N support but also backward compatibility with 802.11 a, b, and g.

I Want My WMV
The Linksys Media Center Extender with DVD (DMA2200) is a rare solution that blends Draft-N speed with Windows Media Center functionality and DVD upscaling to 720p, 1080i, and 1080p. This is what bandwidth was made for!

The Cool CAT
Conventional CAT5 (10/100) will soon prove to be a bottleneck for 802.11n access points capable of better than 100 Mb/s speeds. Make sure your customers have CAT5e or CAT6 cabling in place to handle the heavier load.

Getting Over It
Then something unexpected and often unnoticed happened last year. The spec for Draft 2.0 of 802.11n emerged, vendors updated their products to the new release, and the Wi-Fi Alliance finally swung into action. The Wi-Fi Alliance is an industry body that doesn’t create standards but rather validates that products properly adhere to them. You can find the list of products that have passed Wi-Fi Alliance Draft 2.0 testing at XX.

And what do you know? Products with Wi-Fi Alliance certification are actually interoperable. I tried it out myself. Linksys sent me one of its WPC4400N PC Cards to slap into my antique FIC Centrino whitebook. Not only did the card communicate flawlessly with my D-Link DIR-655 router, but it performed like a champ. I set up a 2GB transfer test between my 802.11b-based whitebook at the kitchen table and the Gigabit-enabled main system in my home office, located one floor up and all the way across the house. Historically, this has been a poor Wi-Fi reception spot. My wireless “11 Mb/s” transfer to the notebook took a leisurely XX; transfer back to the main system took XX. Then I plugged the WPC4400N, restricted its drivers to only perform 802.11g transfers on a single channel in order to mimic a conventional 11g adapter, and obtained transfer times of XX and XX, respectively.

Then I cranked the drivers to full 11n bore and pulled down times of XX and XX—a XX Mb/s average. Also note that the whitebook’s native adapter showed a reception strength of XX out of XX bars in Windows at this spot. Linksys’s driver reported XX bars out of XX when using MIMO-equipped 11n. Now, I also played Devil’s advocate and had Linksys send a basic 5-port Gigabit Ethernet switch (SD2008), figuring this is what many small businesses would use instead of wireless in order to cover a workgroup from a single cable drop. I performed the same test from the kitchen using my wife’s ASUS Gigabit-enabled W1N notebook connected to this switch and back into the home’s CAT5e structured wiring. Transfer times were XX and XX, or a XX Mb/s average.

I’m impressed. I repeated this test at two other well-charted problem points in the house and saw similar results. Draft 2.0 802.11n works. It’s stable, compatible, fast, and able to leap tall buildings. In fact, I walked my whitebook down the block, five houses away, and still had a solid connection (XX out of XX bars) while Web surfing from the sidewalk. This technology is ready for prime time. Apparently, I’m not the first to notice.

“With 11n, there’s now some degree of enterprise adoption,” says Ivor Diedricks senior product manager of the Connected Office Business unit at Linksys, “and enterprises generally only start adopting a next-generation technology when it’s mature or approaching maturity. It’s now being embedded into notebooks and end-points. It’s closing in on becoming a standard and should be ratified by the end of this year.”

From an application perspective, the wider pipeline of 11n should interest customers because it enables a broader range of applications. To offer a consumer example, I sometimes like to watch back episodes of “Chuck” at NBC.com (www.nbc.com/Chuck). Viewed over an 11b connection, the experience is passable but prone to hiccups and buffering, which drives me absolutely mental. With 11n, the experience is like watching TV, albeit a little one, so if you believe in the growth of PC media center functionality, believe in 11n. Draft 2.0 also provides better “predictability” in applications such as interactive voice and videoconferencing. Additional technology has been built into the protocol to provide more consistent coverage and throughput. The net result is that if you couldn’t get customers to bite on wireless because application performance was too haphazard, 11n should provide the reassurance you need to start closing sales.

I’ve always liked Linksys products, and I’m going to pump the company here and now for three reasons. First, let’s be honest, OK? They rushed me product to test with, and D-Link has stopped returning my calls. (Was it something I said? O, D-Link, where art thou?) Second, you can’t argue with the twin credibility barrels of Linksys and parent company Cisco. People know and respect these words above any other brands in the business. Third, the Business Products unit is a side of Linksys that most resellers have yet to explore.

“We see that the small business market has been largely underserved,” says Diedricks. “Obviously, lots of people are focusing on the enterprise, lots focus on the consumer, and many competitors simply take the consumer product and sell it into the small business space. Our charter is to create custom products for this marketplace. As we go forward, you’ll see things like supporting multiple SSIDs and VLANs. You see features where, when you have a multi-AP environment, there will be more seamless management and roaming between APs. All of those things are differentiators and, we believe, essential for the small business marketplace.”

You can already see some examples of how Linksys has found that middle ground between consumer and enterprise. The SLM2005 is a 5-port Gigabit switch with smart management capabilities and the ability to receive power either from AC or Power-over-Ethernet (PoE), which is cool when trying to cut down on workgroup clutter. One of these ports might tie into Linksys’s WAP4400N PoE-based 11n access point. But don’t forget to look beyond wireless as you enhance a company’s network. Another of these switch ports might tie to an NSS6000, a stackable, four-drive NAS appliance that supports spanning and AES disk encryption but still steers clear of conventional (and expensive) rackable storage solutions. I like that Linksys is now into corporate-class surveillance and IP telephony gear. That ties in well with the QoS features built into products like the WRVS4400N 11n router, which also supports onboard, hardware-accelerated VPN processing. This, in turn, gives you a reason to go back and push for notebook upgrades with the WPC4400N.

With Linksys business products, you get an end-to-end line of networking products tailored to small businesses. Also keep in mind that Linksys and Cisco offer a “path forward” program, where, as a small business grows and needs to move into more enterprise-oriented hardware from Cisco, it can get trade-up dollars back for purchased Linksys business products. This is on top of all the other education, training, marketing help, and incentive programs you’ll see from Linksys’s channel folks on these business SKUs.

Early on, I pointed out that the “if a little is good, more is better” approach had some less-than-attractive side effects on my folks’ waistlines. When it comes to wireless networking, resellers should also keep an eye out for dangerous bloating. What will users do with all of that extra throughput capability, and what effect will that have further up the building backbone? One stream of “Chuck” is fine, but what if you have 50 users engaged in videoconferencing, telephony, surveillance streaming, tutorial streaming, and so on? You’d better make sure that business has a LAN backbone ready for that sort of throughput. You may need to plant more access points, more local switches, and perhaps even implement 10 Gigabit pipes near the network’s core. All of this adds up to a terrific selling opportunity, but know what you’re doing and make sure you’ve got the right partners at your back.