Ethernet – the Secret of Its Longevity

Brad Booth
Brad Booth

Mr. Booth, incredibly Ethernet has been around 30 years. What is the secret of its longevity?

Booth: I think there are two aspects of it. One is that it’s fairly inexpensive to use because the focus of many vendors is to drive down the costs, the per port costs. The other aspect is that it’s pretty much synonymous with TCP/IP. And as IP traffic has increased with things like Internet Telephony and Internet Protocol Television, the network protocol that has been carrying IP-based traffic the longest is Ethernet.

Founding members of the Ethernet Alliance include AMCC, Intel, Samsung, and Sun Microsystems, among others. Why was its formation necessary as there have been other similar groups over the years?

Booth: The other groups usually were formed to promote one specific technology that had been developed by the IEEE 802.3 working group. One of the things we discovered was that most of those alliances would disappear once they felt their task was done. But the biggest problem that we were seeing was that the industry did not think that the task was complete. So for example, even though Gigabit Ethernet has been around since 1998, the industry still has many questions about it. What we try to do is keep information on the Ethernet specifications that exist, so that a typical layperson can pick up an Ethernet Alliance whitepaper and the Ethernet spec and might be able to get information that they need.

What do you envision are the upcoming challenges for the Alliance from a technical point of view?

Booth: One of the things we want to do is try and help the member companies figure out what the industry is going to need for some of the next generation of IEEE standards. That’s where it gets challenging because everybody has slightly different views on that. For example, some of our members are in support of 100 Gigabit Ethernet being developed but we also have a few members that would like to see 40 Gigabit Ethernet be developed. Any IEEE standards effort requires some form of consensus building, and the Ethernet Alliance provides a means for our members to have these critical discussions. In some cases, like 40 Gigabit and 100 Gigabit Ethernet, the discussion is about technical feasibility as well as economic feasibility and market potential. Depending on the member’s target market, the crystal ball for predicting what is important to the bottom line will have a different form from one member to the next. The challenge is to get to agreement. The advantage for the market is that once the agreement is made and the project initiated inside the IEEE, the Ethernet Alliance members and other companies have the drive and focus to complete those efforts.

What upcoming technologies are the alliance members currently working on?

Booth: The upcoming ones are what we call higher speed Ethernet which may be both a 100Gig and a 40Gig effort. For a typical data center, these efforts may not have an impact for a number of years. But there are many Internet exchange carriers, the Time Warners and the Cox Cable for instance, that have been using Ethernet in their video data centers or their Internet aggregation points. These companies are finding that 10Gig is not quite enough, so a lot of them are looking at more bandwidth, which is 100Gig. The other one that we’re working on is a concept called Energy Efficient Ethernet. We’re not proposing grandiose modifications to Ethernet to make it more energy efficient. But if you’ve got a link running at 10 Gbps and there’s no data flowing on that link, the question comes up: Can we scale back to the speed of that link and how quickly can we bring it back up depending on the data demands. And finally, there’s 10Gig EPON, which is the next speed enhancement for 1Gig EPON. In various Asian markets, 1Gig EPON is being used for some of the passive optical networks. Companies have also been looking at it in North America for the access market. While 1Gig is taking off, they’re finding that they’d like to see a higher bandwidth capability at the aggregation points, hence the reason to work on 10Gig EPON now.

Would you talk about some of the standards the Ethernet Alliance is promoting, such as CAT 6a cabling, which can support 10 Gbps applications up to the maximum distance of 100 meters?

Booth: We try and promote anything that comes out of IEEE 802.1 & 802.3. So for example, there’s enhanced Power over Ethernet development going on and a number of other projects. One of the things we’re trying to do is also show how the technologies can be used and get our members to develop technical white papers, generally at a high level to explain to the market why the standards committee is developing something like 100Gig or 40Gig Ethernet. And if you’re considering putting in a 10GBASE-T link, which provides 10 Gbps connections over conventional unshielded or shielded twisted pair cables, why you should look at using new cabling, whether it be a Cat 6 screened cable or shielded cable that’s capable of carrying 10Gig or whether you use a Cat 6a cable or Cat 7 cable.

What’s the status of Power over Ethernet (PoE)?

Booth: There’s an existing specification which transmits 13 watts on the cable. Many people find it very interesting because it can be used for VoIP phones: you can drive an Ethernet cable into that phone and you can also power that phone via that Ethernet cable so you don’t have to run any power cords to it. Thirteen watts is enough power for most typical phones but when you look at things like security cameras and some of the newer access points, you realize power requirements have gone up. To be able to power those devices, we need to figure out how we can add more power to the cable. The other interesting aspect is the ability to actually send power over that Ethernet cable that you could feed into the laptop. For example, if you’re in a foreign country and you don’t have the correct power connector but you have access to a PoE port, you can plug in your PoE connector into your laptop and the next day you have a fully charged battery.

Do you think Ethernet technologies have a role to play in the emergence of Web 2.0?

Booth: I’d expect that we’ll see Ethernet showing up a lot more in what we consider the next generation of the Internet. The advantage to that – if we really are moving towards a very strong IP-based type Internet and applications, things like IPTV and video on demand – is the ability to maintain one common infrastructure. This means definitely big cost savings for many customers. Having to change protocols and having to change ways to transmit the data going from circuit-type switching to packet-type switching definitely impacts your ability to carry various traffic.

How important do you think Ethernet was in the development of the Internet?

Booth: Ethernet was fortunate enough to be one of the early protocols used in the universities and corporate environments. That’s why I think Ethernet, even though it didn’t have a direct hand in the Internet, helped people get used to communicating via emails. For example, I think one of the things that was really important in the development of the Internet was that people had already developed the ability to communicate inside their own university or corporate environment. Once those networks were linked together, the ability to communicate over large distances almost instantaneously spawned a strong desire to also have that ability in our personal lives. Ethernet was one of the core technologies in use as the Internet developed and grew into what it is today.

Is there a technology coming down the pipe to replace Ethernet?

Booth: I wouldn’t be doing the Ethernet Alliance if I thought there was something to replace it. There are many other technologies that have come around, but there’s none that has really done what Ethernet has accomplished. As long as Ethernet maintains its focus of providing a reliable, high-speed link at a reasonable cost, and providing what the end customers want, I believe it will be there for a very long time. Primarily, I think the reason is because Ethernet and TCP/IP have become so synonymous that it is very hard to step in and bring in a new technology that would make someone want to go out and replace over 600 million Ethernet ports in the world.