RealTime IT News

Bob Metcalfe, Chairman, Interim CEO, Ember

Bob MetcalfeLove e-mail? Thank Bob Metcalfe.

While he was a graduate student at Harvard, he began inventing hardware and software for Arpanet, the precursor to the Internet. He landed at Xerox PARC in 1973 as "the networking guy" at an interesting time.

Xerox's research center was one of the few places in the world that had a computer on every desk, and Metcalfe got the job of hooking them all up to a laser printer, another new invention.

The PARC crew also wanted to hook up to this Arpanet thing. The solution he and colleague David Boggs came up with evolved into Ethernet, the protocol for connecting multiple computers in what are now known as LANs.

In 1979, he left Xerox PARC to found 3Com, the granddaddy of computer networking.

At first, Ethernet had a tough time competing with IBM's Token Ring, a proprietary networking system that used telephone wires instead of Ethernet's coaxial cable.

But Metcalfe said that fierce competition among startups scrambling for a scrap of the tiny market led to a key innovation: Ethernet via twisted pair cable. And the rest is history.

Metcalfe's career has taken a few twists itself. He spent several years as a journalist, publisher and producer of the Agenda conferences, authoring the book "Internet Collapses and Other InfoWorld Punditry." He joined Polaris Venture Partners in 2001.

As a general partner at Polaris Ventures, Metcalfe was instrumental in funding Ember, a company providing hardware, software and services for embedded wireless networking based on the emerging ZigBee specification. ZigBee is a radio technology for remote monitoring, control and sensor network applications. He stepped in as interim CEO of Ember in July 2005.

In March 2005, Metcalfe received the National Medal of Technology. But he seems prouder of receiving the Medal of Honor of the Institute of Electrical and Electronics Engineers.

Metcalfe spoke with after Ember announced a non-exclusive agreement with STMicroelectronics to jointly develop packages of software, hardware and tools for the ZigBee wireless networking market.

Q: Do Ember and ST plan to offer a complete ZigBee radio, microcontroller and networking software on a single chip?

Ember offers semiconductors and ZigBee stacks and development tools, including the ZigBee system on a chip, the EM 250; also, what I consider a more important product, the ZigBee coprocessor chip.

You can hook it up to the application microprocessor, instead of trying to cram ZigBee into your microprocessor. Then, we provide tools to put ZigBee wireless mesh networking into your application. ST, in some combination, will be selling them or referring customers to us.

Q: What industry problem or need will this agreement address?

We are reciprocal second sources, a term of art in the procurement business. Someone wishing to buy our chips or software can get it from us or ST. We provide technology to people who build it into products, like our home thermostat customer, Golden Power.

So, Golden Power wants to make sure that, if we go out of business, they have someone else to get it from. [Besides that], our engineers will team up with ST's engineers on a couple of products to come out in the next year.

Q: Why not just merge?

We talked about a merger with ST and decided against it. Our high ambition is to serve the entire microcontroller market. About 10 billion microprocessor controllers are shipped every year, provided by dozens of companies.

Being a very fragmented market, if we served only customers of ST, that would be 3 percent. Our goal is to help all microcontroller vendors to become ZigBee-enabled.

What are some of the more intriguing applications that ZigBee networking might enable?

At CES, Eaton announced its Home Heartbeat product, which you'll be able to buy at the do-it-yourself store. Home Heartbeat is a set of sensors you might put on windows or to keep an eye on water flow or temperature. If you're away from home, you get notification by e-mail, pager or IM that something is wrong.

What about Cisco's recent investment in Zensys, maker of the Z-Wave proprietary RF technology? Could this lead to fragmentation of the market -- or even to Z-Wave becoming the industry standard?

Although Z-Wave works perfectly, it's a very low-function, single-channel FM radio. Cheaper, less functional networks that are proprietary have cropped up. The Ethernet analogy would predict that they will prosper but be overtaken by the ZigBee Alliance.

How will Metcalfe's Law, that the value of a network equals approximately the square of the number of users of the system, apply to the 8 billion embedded devices that are only 2 percent networked?

I never pretended Metcalfe's Law is true. It's kind of a vision thing. But it does apply. Take 10 billion and square it. It's a really big number.

We see the ZigBee market initially as people who already have microprocessors wired together, and they're interested in replacing them with something wireless. Phase two begins when people notice products from different vendors that might be interconnected using the standard. When it really kicks in is when you get multi-vendor compatibility.

Q: What made Ethernet win out over competing protocols? Are there lessons for ZigBee in the evolution of Ethernet?

Lots. For one, the notion that a multi-vendor standard is a good idea and eventually triumphs over proprietary alternatives, even if they have a higher installed base and are cheaper, was proven by Ethernet.

How close is ZigBee to becoming the standard for sensor networks?

The battle continues. I've been in it since 1989 at General Electric Co. There are a couple of kinds of networking: consumer electronics, where you carrying video around, for example; and control networking, where ZigBee plays. There are many contending standards and applications. ZigBee is not a foregone conclusion, although it's a leading contender to be the standard in the control space.

But we have Bluetooth and Wi-Fi above us, and they could come down into our space. There are sensor nets and RFID networks, and they could come up into the control space. There are still interesting events ahead.

Could Ember be wiped out if that happened?

The $10 billion per year microcontroller market is huge. There are lots of nooks and crannies, without pretending ZigBee is best for all applications.

But in situations where you have battery-operated devices -- and the batteries have to last for five to 10 years -- where you want low power, and where you don't have 32-bit or 64-bit Pentiums but instead microcontrollers, we think ZigBee is best.

There are several bands of ZigBee, and we've focused on the 2.4 unlicensed international band. That's also where Wi-Fi is.

Q: It's nice that you work with unlicensed spectrum.

Right. And a cool feature of our products is that they coexist with Wi-Fi; they're not jammed by Wi-Fi.

Q: What's your vision of the networked world in the next 10 years? Will everything be networked and able to talk to me and give me information, or is that in the distant future?

Networking isn't done. It just keeps progressing and proliferating. The invasion of video onto the Internet is yet another profound transformation of the Internet, and a lot of things will have to change to do it well ... but here it comes. It's happening now, and it will happen for the next 10 years.

Thanks to that, we'll be consuming much more information and will be able to sub communication for transportation, reducing unnecessary travel.

Q: Do you see any danger to privacy in the ultra-networked world?

There are dangers. I don't worry about them. My job is to connect things, and let other people worry about that. I like what [Sun Microsystems CEO] Scott McNealy says. You don't have any privacy. Get used to it.