|
|||||||||||||||||||||||||||||
|
First Glimpse Of 802.11a It's finally in silicon. The long-awaited 54Mbps technology that's expected to light up the 5GHz radio spectrum is available and being tested by at least one chip supplier and its customers. Learn how it fared in our offices.
Now, the livelyand hitherto theoreticaldebate on the relative capabilities of 802.11a and 802.11b can be waged on the basis of real-world observation. Atheros Communications paid a visit to 802.11 Planet late last week and gave us a demonstration of its single-card, "reference design" 802.11a components, based on early production runs of the Atheros AR5000 chipset, which will be shipping in quantity to OEM customers soon. Atheros has been using the hardware, based on its two-chip solution, in its own internal network for some time. The demo consisted of the famous Madonna BMW video beamed from a notebook server to an 802.11a access point and then to another notebook workstation. We carried the workstation around our offices with some freedom within a range of 75 feet or so with no deterioration in quality until very substantial impediments (heavy concrete walls) interrupted the signal. Following the film credits, product manager James Chen discussed with us the company's early, in-house testing findings, at some length. Today, we'll look at three key issues: 802.11a's effective range, power consumption, and the cost of the hardware. Then we'll look at some additional topics, including a "turbo" mode that allows data links at up to 108 Mbps! Circles within circles According to Chen, typical 802.11b deployments use a cell radius of about 65 feet (based on access points with power in the 15dBm range), which ensures reliable coverage at the maximum data link rate of 11Mbps. Beyond 100 feet or so, an 802.11b link steps down to 5.5 Mpbs, and at roughly 175 feet, to 2 Mbps. With 802.11a, on the other hand, Atheros testing established stepdown points at about 20 feet (from 54Mbps to 48Mbps), 40 feet (to 36 Mbps), 75 feet (to 24 Mbps), 85 feet (to 18 Mbps),135 feet (to 12 Mbps) and 175 feet (to 6 Mbps). The points Chen chose to stress are that at the 65-foot "sweet spot" distance, 802.11a appears to hold about a 3.5x data-rate advantage, and even out at the 225-foot distance, it betters 802.11b by about 3 to 1. (At certain other points in the curve, the advantage is less.) Cost questions Assuming radio activity of anywhere between 10 and 30 percent of operating time, Atheros calculates that the energy cost of sending/receiving a megabyte of data (at maximum data rate) is between four and nine times greater for 802.11b equipment than for 802.11a. Then there's the question of the cost of the hardware itself. Product manager Chen was quick to point out that as a chip supplier, Atheros does not control pricing at the end-user level. (Although the company's decision to use relatively inexpensive CMOS chip-making technology was consciously aimed at minimizing hardware costs.) Nonetheless, Chen said, the companies that will be selling products based on the Atheros chipsetwhich include Intel, Proxim, and TDKhave stated intentions to price their hardware "similarly" to 802.11b equipment. We won't have to wait long to find out what this means. Preliminary findings based on testing of early prototype components using Atheros Communication's AR5000 chipset suggest that the five-fold speed increase comes at a smaller penalty in effective radio coverage than theoretical predictions suggested. While Chen demonstrated the company's "reference design" components—access point and PCMCIA network interface card—in our offices, we discussed the performance data. Chen touched on other important facets of the 802.11a specification—most importantly those relating to the broadcast spectrum allocated to this technology. Here's the crux: Whereas the spectrum allotted to 802.11b in the 2.4GHz range affords only three clear transmission channels, the portion of the 5GHz spectrum set aside for 802.11a "indoor" applications has room for eight clear channels. (In addition there is a further allocation set aside for higher power, "outdoor" applications.) This has implications not only for dense deployments in corporate, campus, and other public venues, but for raw performance as well. Beehive space With 802.11b (Wi-Fi), which, again, is limited to three clear channels, deployments of more than three contiguous cells are likely to be subject to some performance degradation (up to as much as 50 percent) due to "co-channel interference" (CCI) between cells operating on a given channel. This is because with Wi-Fi, there's no way to avoid duplication of channel usage more than one cell diameter away. And the closer together the cells, the more interference. With 802.11a's eight channels, however, it's a relatively simple matter to arrange a cell grid so that access points using the same channel are at least twice as far apart—and the overall density of cells using any given channel is roughly one fourth as great. This should greatly reduce the effect of CCI, if not eliminate it altogether. Other spaces Another application that Atheros has already implemented is a "turbo" networking mode that, theoretically doubles the basic 802.11a data-link speed to a whopping 108 Mbps, or roughly the equivalent of Fast Ethernet. Atheros doesn't claim this speed for turbo mode, rather splitting the difference and specifying 78 Mbps. Any way you slice it, however, this is great throughput for a wireless LAN. End
|
|
|||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||