* Could Marvell’s faster chipset raise the 802.11n performance bar?
Last week, another possible reason for postponing enterprise deployments of 802.11n WLANs reared its head when chipmaker Marvell announced a 450Mbps chipset that Cisco will likely use in its enterprise access points.
Chipmaker Marvell said its TopDog-450 802.11n chipset, to ship in volume next quarter, will deliver a 1.5-fold speed improvement and 1.6-fold range improvement over the pre-standard 802.11n devices currently on the market. And it looks like Cisco will be using the components, potentially raising the performance stakes for competing business-class WLAN system makers.
“Cisco has been our partner since Day One. It will go with 802.11n across the board to get the whole industry moved from 802.11g and 802.11a,” asserted Sameer Bidichandani, senior director of technology strategy at Marvell. For its part, Cisco said only that its policy is “not to comment on component vendors (in particular, those for future products).”
To achieve 450Mbps maximum theoretical throughput, the Marvell chipset supports three spatial streams using 3×3 multiple-input multiple-output (MIMO) technology, which leverages three transmitters and three receivers. Most products on the market today support 2×2 or 2×3 technology and yield 270Mbps to 300Mbps theoretical maximum speeds. Competitor Atheros has a 3×3 chipset that supports two spatial streams and 300Mbps. It takes the two best streams of the three to build in extra reliability and redundancy. The company has not publicly announced work on a 450Mbps chipset.
Engineering arguments persist as to how many antennas and spatial streams add up to the ultimate performance – further evidence that 802.11n devices are fledgling and in transition. 802.11n has yet to be ratified as a formal IEEE standard. In other words, today’s pre-standard products are young and will likely be soon obsoleted by faster products and those that conform to the final 802.11n standard next year.
These unsettled matters are reasons to hold off on large-scale 11n commitments for now. Spend a good chunk of this year testing 802.11n products and architectures, writing 802.11n RFPs and making spot deployments, perhaps in new all-wireless sites or in network segments where bandwidth is already strained.
