Cable Tech

Siemens Sees Ethernet Everywhere

Siemens Information and Communications Networks Inc. today cast its vote on the issue of whether Ethernet will end up replacing Asynchronous Transfer Mode (ATM) in broadband access networks everywhere.

Siemens believes it will, and it unveiled its strategy for being a big player in the resulting equipment market at the Broadband World Forum, in Seoul, South Korea, today (see Siemens Presents Ethernet DSL Strategy).

The strategy includes Siemens ICN acquiring a 40 percent stake in Dasan Networks Inc., a Korean Ethernet access equipment vendor, for an undisclosed sum.

Dasan is a big beneficiary of Korea's exploding broadband service market. The company's sales have rocketed from $10 million in 2001 to $58 million in 2003. It now has 220 on staff, and its customers include KT Corp., Hanaro Telecom Inc., and Dacom Corp., together with some Japanese carriers.

Siemens and Dasan will jointly develop products. Dasan will develop high data-rate switching systems (VDSL and ADSL II+), and Siemens will develop aggregation switches. A key part of the package, called the Siemens Surpass hiX 5400, is a network management system.

The rollout of higher bandwidth services, requiring up to 50 Mbit/s, is driving demand for Ethernet rather than ATM access networks, according to Bernd Grossmann, director of business development for Siemens ICN's carrier Ethernet solutions. "When you get to higher bandwidth services, the cost of [ATM] bandwidth is too high," he says. "The same technology is going to be used for business and residential services."

Christian Unterberger, Siemens ICN's president of carrier networks, says most invitations to bid for Asia/Pacific broadband networks now call for IP/Ethernet rather than ATM. This applies to the complete connection from users to broadband remote access servers (B-RASs) -- that is, between the user and the DSLAM and between the DSLAM and the B-RAS.

Unterberger expects the same thing to happen in Europe in 12 to 18 months' time. Others aren't so sure.

"Just because it's happening in the Asia/Pacific doesn't mean it will happen everywhere," says Graham Beniston, an analyst with Heavy Reading and author of its recent report, Next-Generation DSL Equipment: The Path to Profitability. Beniston points out that cities in the Asia/Pacific are characterized by high density housing and short line lengths, which make VDSL deployment attractive. This isn't so much the case in Europe and the U.S.

European and American incumbents also have other issues to consider, notes Beniston. They already have huge installed bases of ATM infrastructure, so shifting to new technology could be expensive and disruptive. In the U.S., RBOCs want to generate new revenues by rolling out advanced services, using ATM to ensure quality. They're not convinced that IP/Ethernet is up to that task, even though Siemens says it's got schemes in hand to address this issue.

Issues such as this are likely to mean that the broadband access market remains "pretty fragmented" for the foreseeable future, according to Peter Linder, technical director of LM Ericsson (Nasdaq: ERICY). Ericsson itself has been a big promoter of Ethernet in broadband access networks. It recently clinched a deal with Copper Mountain Networks Inc. (Nasdaq: CMTN) to address carriers wanting to stick with their ATM infrastructures (see Ericsson Mines Copper Mountain).

— Peter Heywood, Founding Editor, Light Reading

Archives of Related Light Reading Webinars:

Page 1 / 12   >   >>
truelight 12/5/2012 | 1:51:11 AM
re: Siemens Sees Ethernet Everywhere Siemens is a slow moving giant that has finally admitted the move to Ethernet access.
arch_1 12/5/2012 | 1:51:07 AM
re: Siemens Sees Ethernet Everywhere The article says that Ethernoet is OK for Asia but not US and EU. US/EU last-mile providers want to use ATM to provide value-added services.

Wake up, people! The customers don't WANT value-added services from last-mile providers. They want cheap, fast IP, and nothing else. With cheap, fast IP, the customers can get value-added services over IP without reference to the last-mile providers. The providers can continue to ignore this, but if they do they will eventually be bypassed by alternative last-mile technologies, starting with cable.

Good IP connectivity is now a fundamental part of the infrastructure of a technical society. To the extent that good IP connectivity is cheaper in Asia than in the US/EU, Asia will be able to out-compete us in the technical fields. The dinosaurs can perhaps slow the uptake of cheap IP and squeeze out a few more years of monopoly profits, but only at the cost of our technical competiveness.
tmc1 12/5/2012 | 1:51:07 AM
re: Siemens Sees Ethernet Everywhere WOW... they decided that ethernet is going to be everywhere. I would have never thought that! How revolutionary. These guys are so far ahead of their time.
technonerd 12/5/2012 | 1:51:05 AM
re: Siemens Sees Ethernet Everywhere Paradyne and Nortel made excellent broadband access systems ("MVL" and "Etherloop") in the late 1990s but they were rejected by the RBOCs because they were based on Ethernet and not ATM. It didn't seem to matter that the PHY layer was indestructible, i.e., operated over long, dirty copper with no cross-talk.

It always mystified me that the CLECs ignored this stuff, and that the so-called gigabit ethernet carriers never used it for the access side of things. They were focused entirely on fiber access to big buildings, and totally ignored a much bigger copper access market that was easily addressed with these technologies.

But now there is "Ethernet in the First Mile." This is another knee-slapper, by the way, as if calling "the last mile" by a new name, "the first mile," will make it new. Who knows, maybe it will. I want to see how Morgan Stanley, et. al., tart this one up when the first "ethernet in the first mile" concept wants to go public.

Yes, ethernet access makes all kinds of sense. It always did. For the voice side of it, you just run PCM tunnels through the lower frequencies.
SuperChargeAccess 12/5/2012 | 1:51:04 AM
re: Siemens Sees Ethernet Everywhere I could really use some help understanding the reasoning for using ethernet in the first or second mile(or last and second to the last). From what I know and from what I've read, the deciding factor is cost. And the largest cost factor is the aggregation/switch. This switch must carry voice and various types of data and it most likely will be located in a cabinet in the field so it must be temperature hardened. Also since it must carry voice, it must be redundant, highly reliable, withstand lightning, and utilize some level of QOS. So this is not an off-the-shelf LAN quality switch.

My question is then, if you create a ethernet switch that incorporates the robustness and the QOS that is needed, why should that ethernet switch be any cheaper than an ATM switch that has the same capabilities?

ATM has come from a background of high levels of robustness and powerful QOS capabilities. It was designed from the ground up to carry voice and data on the same pipe. Ethernet has come from a low cost, low performance background which has given it the perception of always being low cost. There have been layer after layer of changes, added protocols, etc trying to make it robust enough for access.

I maintain that when ATM and ethernet switches are designed for use at the same point in the network with identical specifications that the difference in cost will be small and may even go in favor of ATM due to it's small, fixed packet size.

Whether you use VoIP or not does not affect this decision since both ATM and ethernet carry IP and the QOS requirements will be the same.

Since the internet backbone is ATM and the first and second mile is usually ATM, why the big push to squeeze in ethernet?
truelight 12/5/2012 | 1:51:02 AM
re: Siemens Sees Ethernet Everywhere
The communications infrastructue system is changing and the economics for success are based on cost and ubiquity. When you switch wireless carriers do you really care what technology the use - nope. Just the right service and the price you are willing to pay.

ATM as a technology and innovation is done - finished. ATM has served its purpose that's why many of the 'technical' leaders are now on Ethernet issues.

Fact not fiction
truelight 12/5/2012 | 1:50:59 AM
re: Siemens Sees Ethernet Everywhere Siemens is like that - slow to change, slow to acknowledge their own shortcomings, slow to adopt new technology and market trends. They are just really really .......boring slow, snails and stubborn Germans, let me tell you they are stubborn, I mean so stubbborn they refuse to acknowledge the WWII crimes etc, now that is stubborn or they will deny it's their fault, something about the culture and inbreeding.

The interesting news is when the shift focus on a new area they are big enough to make an influence.
arch_1 12/5/2012 | 1:50:58 AM
re: Siemens Sees Ethernet Everywhere SuperChargeAccess makes three points:
1)ATM and ethernet technology should have equivalent cost at the initial aggregation point.
2)ATM is intrinsically QoS and voice capable.
3)ATM is already used in internet backbone.

1) Whether or not the two technologies "should be" cost-equivalent in this role is irrelevant. They are not price-equivalent, and the reason is not technical, it's cultural. ATM comes from the Telco "big-brother" mentality. The last-mile providers want to use it to "add value," while the customers just want fast cheap IP. There are many examples of this. A cheap ethernet firewall/router costs $50.00 or less. The cheapest router with a T1 interface costs $500.00 or more. Why? Please do not tell me that a T1 framer "should be" more expensive than a MAC that runs nearly 60 times faster.

2) ATM is intrinsically QoS and voice capable.
Who cares? Fast cheap Ethernet is also QoS and voice capable, and I'm not locked into the last-mile provider's services.

3)ATM is used in the internet backbone.
This is completely irrelevant. The IP datagrams sent on the first mile are extracted from the ATM cells and forwarded into the internet as IP. Certain internet routers are interconnected via ATM because this happens to be the cheapest installed legacy mechanism for certain routes. Other routes have Frame relay, and still others have SONET. The routers don't care: they just encapsulate the IP in whatever is available. Actually, modern routers do "care." It's cheaper to use less encapsulation. If you were building a new network from scratch, it would be IP only, with all ATM switches and FR switches replaced by routers.
fiber_r_us 12/5/2012 | 1:50:57 AM
re: Siemens Sees Ethernet Everywhere Few of the Tier 1 Internet providers use ATM to interconnect Tier 1 cites today... It is all POS using unprotected optical wavelengths. This is true for:

1) AT&T
2) Sprint
3) Level 3
4) WilTel
5) Qwest
6) WCOM (aka MCI/UUnet, though these guys still have a lot of legacy ATM in the access)
7) and a bunch of smaller player

The RBOCs (which are not Tier 1 as they generally don't cover enough of the US) still use ATM for DSL aggregation, but use POS for longer inter-city routes.
Peter Heywood 12/5/2012 | 1:50:54 AM
re: Siemens Sees Ethernet Everywhere My impression is that, with incumbent carriers, the issue of whether to move away from ATM is unlikely to come down to a comparison of technologies.

It's more likely to come down to whether it's worth having to replace all your OSS systems, retrain all your maintenance and operations staff, redo all the other issues like storing spares etc.

Page 1 / 12   >   >>
Sign In