Swedes Cut Metro WDM Costs

A new player in the market for metro transport equipment has emerged in the form of Transmode Systems AB.

The Swedish startup only got started in April 2000, but it’s already nailed its first U.S. customer, Stealth Communications Inc., a service provider offering high-bandwidth connections to big corporations in and around Manhattan.

Stealth picked Transmode to provide the optical backbone that underlies its router network after turning down a couple of other conventional metro DWDM solutions. It got close to buying gear from Sorrento Networks Corp. (Nasdaq: FIBR). But Sorrento couldn’t handle the low-cost 850 nanometer VCSELs (vertical cavity surface emitting lasers) that Stealth wanted to use in its router interfaces, according to Shrihari Pandit, Stealth’s president and CEO, who adds that another proposal, from ONI Systems Inc. (Nasdaq: ONIS), was prohibitively expensive.

Transmode got the order, worth $1.5 million in the first year, because of its innovative approach to metro transport networks -- an approach that slashes costs by a factor of three, according to Dirk Lutz, one of Transmode’s founders.

In order to understand Transmode’s technology, it’s best to start by reviewing the current way that most metro DWDM systems work. The more sophisticated ones are ring-based at the physical layer. In other words, the actual fiber links together a number of nodes in a ring. However, the wavelengths running over the fiber follow a logical star topology. There’s a hub, and each node is linked to it using a different wavelength of light.

This has a couple of consequences. First, any communication between nodes on the same ring requires two wavelengths -- one from the source node to the hub and another from the hub to the destination node. Second, when wavelengths whistle past an intermediate node on their way around a fiber ring, they lose quite a bit of power. As a result, the number of intermediate nodes they can whistle past is limited -- and that limits the reach of most metro DWDM rings.

Transmode reckons it's cracked both of these problems by coming up with a way of creating logical wavelength rings over physical fiber rings. In other words, traffic going from one node to another on the same ring does just that, using a single wavelength rather than two, because it doesn’t have to pass via a central hub. As 80 percent of traffic stays within the same metro rings, according to Lutz, this is a big deal.

Likewise, broadcasting only needs a single wavelength rather than multiple ones, and something like 80 percent of metro network traffic is expected to be broadcast, according to Lutz. In other words, another big deal.

Using a lot fewer wavelengths to achieve the same goals has enabled Transmode to use much lower-cost coarse wavelength-division multiplexing (CWDM) in its equipment. This spaces out wavelengths so that much less expensive lasers can be used -- one of the issues that helped it win the Stealth contract. Right now, Transmode’s equipment only supports four wavelengths per fiber, but it’s aiming to offer 16 wavelengths per fiber next year. Each wavelength carries 2.5 Gbit/s of data.

Transmode’s logical rings also help it offer longer reach networks, up to 150 kilometers, because it doesn’t face the problems of light power being eaten away by filters in intermediate nodes. In Transmode networks, traffic is taken off the ring and retransmitted by electrical equipment -- like Stealth’s routers -- at every node.

"We have a patented ability to detect line rates, coding, and types of signals associated with different protocols," says Lutz. "We provide unified physical layer management. We take away the manual process of assigning wavelengths. We offer a way to [easily] serve a lot of access points on a metro ring."

”As far as we can tell, we’re the first ones to deliver a logical ring,” says Terje Hallan, Transmode’s CEO.

Transmode’s transport system is intrinsically more efficient when carrying ring-based protocols such as resilient packet ring (RPR), the spacial reuse protocol (SRP) promoted by Cisco Systems Inc. (Nasdaq: CSCO), and the dynamic asynchronous transfer mode (DTM) technologies promoted by Scandinavian companies such as Dynarc and Net Insight AB (Stockholm: NETI-B), according to Lutz (see Cisco Clarifies Ring Strategy and Fortunes Swing for Swedish Startups ).

Transmode’s technology can also be used to support logical star topology networks, making it suitable for protocols such as gigabit Ethernet, Lutz adds.

Transmode was born out of an earlier company, also called Transmode, that was founded by a bunch of engineers from Ericsson AB (Nasdaq: ERICY) in 1992/93. It's received US$10 million in funding to date, $2 million at its founding from European Equity Partners (EEP) and $8 million in April 2001 from EEP and Amadeus Capital Partners.

Stefan Lindeberg a former European marketing executive with Cisco Systems Inc. (Nasdaq: CSCO), serves on the board of Transmode. Interestingly, he’s also on the board of Lumentis AB, another Swedish startup making metro DWDM platforms (see Lumentis Faces Metro Challenge).

— Peter Heywood, Founding Editor, and Mary Jander, Senior Editor, Light Reading
gea 12/4/2012 | 7:27:30 PM
re: Swedes Cut Metro WDM Costs This article would make it seem as if Transmode was the first to offer wavelength reuse within a metro ring. If that's the case then there are some serious problems here! There are plenty of metro ring products that allow for wavelength reuse. Indeed, its possible that ALL of the metro ring products allow for wavelength reuse.

The only thing I can think of is that this product actually muxes and demuxes all wavelengths at every node on the ring...they didn't even bother trying to do ring OADMs. But if that's the case, this inflexibility would have to be justified by some significantly reduced costs, and that should be described more clearly in the article.

You Lightreading folks want to clarify this?
poster 12/4/2012 | 7:27:28 PM
re: Swedes Cut Metro WDM Costs gotta say once again I agree with gea! how about some analysis here! logical rings? what is different about this? what's the difference between this and SONET, besides SONET being a widely accepted standard deployed throughout the world? what's the difference between this and the 15454 with a few transponders? nothing that I can see, except the 454 supports more channels! they have to be doing some minor grooming and muxing at each node. big deal! bottom line: it's a proprietary muxing scheme that allows them to build closed rings - and saves costs becuase they reuse lambdas and pack them more efficiently, supposedly. so what? there's tradeoffs for everything. let's here em LR!

founded in '92 with $10M? and they have a working product? Stealth Communications? come on guys, you are supposed to filter out the bullshit for us!

optical_maverick 12/4/2012 | 7:27:26 PM
re: Swedes Cut Metro WDM Costs article is a little squirrly and the technology at least from the authors viewpoint is somewhat vague.........most importantly is the fact that traffic will have to pass through the hub in order for protection......without protection no major players will consider this technology.

generalcharlesdegaulle 12/4/2012 | 7:27:26 PM
re: Swedes Cut Metro WDM Costs Basically optical rings w. each link running
point-to-point CWDM...nothing more...
Vesting 12/4/2012 | 7:27:22 PM
re: Swedes Cut Metro WDM Costs I am wondering if someone from Light Reading actually wrote this article. The assumption that all traffic in a WDM based OADM has to be logically hubbed is one of the most inaccurate statements I have ever read in this forum. Just about every vendor out there can support distributed UPSR traffic patterns. What have you guys been smoking? Did someone at Transmode write this for you and you are passing on their stupid assumptions? You should be ashamed of yourselfs for not taking the time to do a little research as to their claims.
lightmaster 12/4/2012 | 7:27:15 PM
re: Swedes Cut Metro WDM Costs The stuff about all DWDM rings being hubbed is BS. Anybody with an OADM based system can do point to point. Many are optimized around a hubbed structure for two reasons: 1.For large carriers in the US, most metro traffic IS hubbed because the rings are small. 2. In order to use amplifiers, you have to terminate the ring somewhere. So, most system support low wavelength drop systems and a high wavelength box at a hub. That doesn't mean they are not capable of point to point in a non-amplified system.

80% of traffic is on the same ring, and 80% is broadcast? Maybe in Sweeden.

So, they don't have their facts straight, so what is different about them? They say they are using CWDM, and they may be the first to build a ring out of CWDM with VCSELs on their transponders. This would significantly lower the cost, but also significantly limit their potential market. CWDM supports much fewer wavelengths and can't use amplifiers. It's also very easy to do with off-the shelf components.

If this is the case, the market for the product will be limited to smaller lecs - like Stealth. Plus, it won't be hard to duplicate - assuming the market is large enough for anyone to bother.

P.S. I love Stealths web site. An Internet company really should invest in a real web site.
erbiumfiber 12/4/2012 | 7:27:14 PM
re: Swedes Cut Metro WDM Costs This appears to be their "patented ability to detect line rates, coding, and types of signals associated with different protocols" which seems to be their claim to fame according to the article. If you search "Transmode" as assignee you come up with zero US patents and about 8 published applications. Here's the link to one of the English language documents:

lightmaster 12/4/2012 | 7:27:12 PM
re: Swedes Cut Metro WDM Costs Looks like something I designed in my freshman intro to EE class.
exphoton 12/4/2012 | 7:26:34 PM
re: Swedes Cut Metro WDM Costs LR:

It is quite a shame when you cannot respond to some obvious queries as to whats so great about Transmode. At least withdraw some of the claims - by posting such reports and not taking the time to respond, you a)lose credibility about accurate reporting, b) appear to be shilling for companies who write the articles for you, c)demonstrate that your reporters do not understand the first thing about metro WDM (among other topics), and d) simply perpetrate the notion that this is a junk website with no focus on quality.

bougnoul 12/4/2012 | 7:25:49 PM
re: Swedes Cut Metro WDM Costs Dear Indignant readers:
This painfully reminds us of the 802.3 1 G-bit std. meetings in Vancouver...when similar EU claims were presented that essentially proverd through a lot of verbiosity that "they" have gotten 1kM links going without any such problems we were trying to design! It was some "surface modes" on copper wires...instead of the regular Maxwell-theory dependent current/voltage through the wires at thse frequencies!
We did feel totally dumb.
Is this something of that variety? Or are we still dumb?
Sign In