Silva
Chipmaker Marvell announced last week the introduction of its TopDog 11n-450 802.11n Wi-Fi chip which
boasts speeds of 450Mbps. The vendor will be demonstrating the new chip at its
booth at the Consumer Electronics Show this week.

450Mbps? Yes, we have
reported speeds for 802.11n, based on the draft standard, as being in the
neighborhood of 300Mbps, so you may be wondering how Marvell can accomplish
this accelerated throughput. The answer is 3X3 configuration – instead of the
more standard 2×2 configuration – in which three transmitters and three
receivers on the 90nm chip increase the theoretical throughput by 150Mbps. All,
of course, in controlled testing environments. Overall throughput and realized
gains are not likely to mirror these results in any real-world application.

The design, while increasing the speed at which the chip can pass and receive
data in theory, is unlikely to reflect the linear increase in throughput the
additional transmitter and receiver would seem to imply. Additionally, the true
benefit of the new chip will only be realized in a symmetrical application with
TopDog-enabled hardware at each end of the data interchange. Furthermore, the
additional gear on the silicon will increase the price of production and – at a
minimum – allow Marvell to demand a premium from a chip. The result: a
questionable benefit in the most favorable RF environments, and one that’s
negligible in noisy RF implementations – that is, most real-world
implementations.

So what’s the big deal? This advance will likely be implemented with an associated
price-tag increase in consumer gear first, enterprise gear later, all with a
mere potential for increasing data
throughput. Will 3×3 become a standard chip design? Perhaps, however intelligent and directional antennae – currently  an infrastructure design element common to only one or two infrastructure vendors –  will likely need to precede wide-scale 3×3 chipset integration to realize the performance benefits promised by the new chip design.