First, let me apologize for not being here for the past six months, but, man, things have been busy!
Ethernet is exploding all over the place. Often, I get accused of only paying attention to things going faster. Okay, i will admit that there is some truth to that, but I think many of us forget that speed is a relative thing.
For example, introducing Gigabit Ethernet into a new application may not sound exciting, but when you consider that Ethernet is currently run at 100 Mb/s in networks in automobiles, the 10x pickup in speed sounds more exciting. And it doesn’t hurt that by the end of the decade this may represent 300 million ports of Ethernet. Click here for more details.
Or consider that the IEEE will hold a call-for-interest at its July 2012 Plenary to discuss the next generation BASE-T project. Click here for more information.
So right now standards are focused on making all sorts of electrical interfaces run faster and exploring optical solutions that will use less power, cost less, and enable higher density. Hmmm, maybe we just need to turn the speed dial up on the process? Tempting thought, but it really takes time to solve these problems and for the industry to reach consensus.
So, John, you ask, why all this talk about speed? Well, frankly, it has been on my mind all year, as the IEEE 802.3 Ethernet Bandwidth Assessment Ad Hoc (BWA) has been working on its assessment, based on the data collected last year. Final approvals are expected to come at the IEEE 802 July 2012 Plenary, as well as an overview tutorial on the ad hoc's findings. Click here for more information on the IEEE 802.3 BWA Ad Hoc.
I can't spill any of the beans yet, but what I can say is that the data tsunami that is pushing 40GbE and 100GbE into networks throughout the Ethernet Ecosystem is starting to build pressure and highlight other bottlenecks in the ecosystem. Currently, I am on my way home from the WDM & Next Generation Optical Networking Conference that was held in Monte Carlo (you have to go there once!). The debate between 400G and Terabit Ethernet is already raging.
I am pretty sure the second half of this year will be busier than the first half. Please say a little prayer for me!
John D'Ambrosia is chief Ethernet evangelist, CTO office, at Dell.
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.