Consider the IEEE 802.3bp
Reduced Twisted Pair Gigabit Ethernet project currently underway. As the
name implies, the effort is defining delivery of Gigabit Ethernet over
less than the standard four pair used today—perhaps one or two pair.
One of the primary justifications for this project is networking
applications in vehicles. It has been forecasted that a nominal 300
million ports in automobiles will be shipped by the end of this decade. I
would say that is fairly high volume, wouldn't you?
And once we
have these networks in these vehicles, it is easy to envision the
various applications that will be developed. Can't you see the
commercials now—the kids in the back of the car screaming, and the
parent turning around to plug in some new toy into the automotive
network. The screaming and tears stop instantly as the kids start
watching streaming video or chatting with their friends. Okay, maybe not
instantly, but you get the idea.
As a parent with two kids who
have the electronic toys of today, the family trips I take now would be
so much easier if there were more ports in the car for them to charge
these toys. So how about Power over Ethernet for my car too?
example illustrates a few lessons that the Ethernet community should
consider. First, there are still applications out there for Ethernet
outside of the enterprise and data center. There are lots of places
where a mature networking technology could make an immediate impact.
benefits of networking are well known, and I don't see any reason to
re-iterate them here. However, the next lesson we can draw from this
example is this: Ethernet has a wealth of technology that can be
leveraged into a multitude of applications. And let me be clear—not
everything Ethernet does has to go 100m over UTP cable and connect with
an RJ45 connector. I would argue that mindset has limited the industry
from exploiting Ethernet's catalog of technologies.
would love to see a small form factor Ethernet connection for the
millions and millions of consumer products now being shipped. It is
time that the Ethernet industry recognize its own potential and start a
true consumer focused Ethernet project.
John D'Ambrosia is chief Ethernet evangelist, CTO office, at Dell.
Not long ago, "everyone" was telling us how the economy had fundamentally changed, only because some day traders were raking in millions for a short period of time. But in that time, these pundits couldn't see that the free-for-all was a mere blip in history, and that the economy would soon stamp out those who provided no added value to it.
Today's blip is handhgeld gadgets, and the pundits seem to think that these handheld toys are all there is and all there ever will be in the future. So if a technology doesn't immediately apply to the toy, that technology is supposedly "in decline."
Ethernet speeds keep going up because the carriers are using Ethernet in place of SONET, for their trunk lines. And the demands for capacity of the trunk lines and in all manner of core ISP networks are still going up. As people migrate their TV watching to the Internet, I can only predict a more demand for bandwidth in core ISP networks. So there's no question in my mind that Ethernet has to continue to grow, and that 802.11 is simply a different discussion, for a different set of applications.
IEEE 802.11 is merely a last-few-meters connection link. It ain't the main event. It is cabled Internet networks that do the heavy lifting.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.