TEF Q&A – Michael Bennett, Lawrence Berkeley National Lab
What are the top three things you’d like to take away from the Technology Exploration Forum (TEF)?
I’m interested in hearing what kinds of problems end user’s are running into and what kind of solutions they are seeking. I’d also like to get a sense of where Ethernet is headed from here, especially in terms of higher speeds – what are they going to be and how soon do we need them? And, I’d like to make connections with people that I might not otherwise meet face-to-face.
What is your role at the TEF?
I’ll be the moderator of the panel on the Role of Power in Networks. One of the things I’m looking forward to is getting the conversation going between the panelists Una Song and Bruce Nordman, who both work in the energy industry, and Bob Felderman from Google, who will share his views about the things he thinks are missing in the industry.
I’m hoping that the panelists and myself will be able to indicate to the people in the room the state of energy efficiency in the industry, discuss where we would like to see it go, and talk about opportunities available to close the gaps or further develop solutions that are in the works.
What topics or issues will be of particular interest to attendees?
We’ll start by painting a broad stroke from general to specific. In general, the panel will provide an update on numbers presented in the past. In specific, Bruce, who is an energy researcher, will discuss energy use by network equipment.
One of the things that needs to be repeated is that solving the energy problem requires a multi-pronged attack. There is no one-size-fits-all solution. The panel will cover the solution space and serve as a needs analysis to identify the things that look like they have been solved or that we are making progress on, and shine a light on other areas where we see there is still work to be done.
When we talked about putting together the panel, Bob Felderman pointed out that there are a lot of devices out there, but there is no software to help manage them. He will be talking to a room full of experts that will be able to discuss and explore this problem
We will have this environment that is rich with really bright people who are looking for really smart ways to solve problems. If a good idea comes up, light bulbs will come on and people will start talking and their ideas will have real potential to take off.
Will people really feel free to share their ideas?
That’s where the Ethernet Alliance and IEEE play an important role. Most people attending will be members of either or both organizations and they will share ideas under a certain set of rules that allow them to discuss almost anything without sharing everything. The TEF is almost like a focus group except we are covering a lot more ground. Nevertheless, we will be focused and organized around Ethernet. But even under the umbrella of Ethernet, there are a lot of things to talk about.
What’s more interesting — the past, the present or the future?
I think people will be most interested in the future. I have my own curiosity about things that I hope to find out about. Energy Star is an energy-efficiency incentive offered by the EPA to help people competing in the market by getting a label. I’d like to know when we are going to see Energy Star labels on networking equipment.
Energy Efficient Ethernet, IEEE 802.3az, was published in October of last year and we are seeing products that are poised to hit the market. So, we want to find out how much of an impact these products are going to have and find out if a device has it, can it get an Energy Star label, and what types of devices are eligible for a label.
What results are you hoping your panel will produce?
I’d like to discuss things that are being done in front of people who can take what we say and further develop worthy ideas. For example, Bruce is going to talk about a relatively new idea that has to do with Power over Ethernet. People have heard hallway conversations about his ideas, but those conversations have never really gelled. The TEF is a great environment to toss out an idea, start talking about it, and discuss if things should be moved forward to initiate the process that moves the idea into the IEEE.
I also hope we can identify gaps in the energy solution set with respect to Ethernet. If any are sufficiently interesting and have a high value proposition, we can gauge whether or not people are interested in working on them. That is what I’m looking for but the other panelists have their own set of goals.
Are companies pursuing energy efficiency simply for competitive advantage or are they just as concerned about the amount of energy being wasted by all the networking equipment being deployed?
It’s a little bit of both. For example, Verizon has a publicly pronounced energy efficiency policy. If you Google the words Verizon and green, you’ll find dozens of announcements on what they are doing to reduce energy usage and its cost.
As much as companies want to save the planet, they also want to reduce their operating costs. When they think about installing networking and computing equipment they realize that for every watt required for operation, another watt is required for cooling. They are looking for innovative ways to reduce their carbon footprints and their costs in order to accomplish their goals.
How much energy are we talking about?
We did forecasts on 802.3az Energy Efficient Ethernet and based on the market coverage we were able to come up with pretty big numbers. When you look at just the device itself you are talking about saving maybe a half-a-watt to one watt per port depending on the speed of the device. But, if you have billions installed and turned on 24×7, then it adds up quickly. A conservative estimate that was made said that about $450 million a year in energy savings could be achieved once the market was completely saturated with 802.3az devices. We won’t know how close the forecast was for years to come.
When will we see 802.3az products being used?
A few network equipment companies offer it and now the Mac Book Pro supports it . Apple is the first major computer vendor to use it, and we expect the others to follow suit. I know that there are many network equipment vendors supporting the feature in their products. Maybe by this time next year there will be enough products in the market that we can do a case study or two.
What about the existing installed equipment?
The industry standard lifecycle is about five years. In addition, there is a lot of wireless equipment coming on line. These days, enterprise network managers are asking themselves whether or not they need to install wired networks. Fortunately the 802.11 suite of wireless products have built-in energy efficiency. That is going to affect us.
Wired networks will be around for a long time, but it will be interesting to see number of copper Ethernet ports shipped to wireless ports, or equipment. A computer plugged into a jack represents a port, but wireless devices can be used to support many devices including cell phones, laptops, printers, etc. If there is a big transition from wired to wireless taking place, perhaps more focus should be placed on what is being done in wireless.
So there are a lot of open questions. We will probably have as many questions as we have things to say on our panel. Having a room full of end users and vendors is a great environment to ask those questions and get feedback.