Optical switches may have the potential to revolutionize telecom networks in the long run, but right now most optical switch manufacturers are in a bit of a pickle, according to the latest report from the Optical Oracle, Light Reading’s subscription research service.
Called “Switch Hitting: OEO vs OOO," the report identifies different types of optical switch and reviews their market prospects. It then analyses the strengths and weaknesses of the major players by examining their financial figures. The major conclusions are these:
Ciena Corp. (Nasdaq: CIEN) is the only company making steady money out of optical switches. Sales of its CoreDirector switch are expected to represent 30 to 35 percent of fourth-quarter revenues, scheduled for announcement on Thursday (Dec 13). That compares with a mere 12 to 15 percent in the previous quarter. However, the report suggests that Ciena’s margins will come under pressure now that competitors are catching up.
Corvis Corp. (Nasdaq: CORV) has “alarming” inventory metrics that are “far worse than any competitor,” according to the report. Inventories as a percentage of cost of goods sold in the most recent quarter reached 1,182 percent, nearly twice Sycamore’s 673 percent and more than 10 times the normal mid-90s range achieved by Lucent and Nortel.
Lucent Technologies Inc. (NYSE: LU) can’t escape its general problems, which shroud its prospects in the optical switch market with uncertainty..
Nortel Networks Corp. (NYSE/Toronto: NT) plans to ship its OPTera Connect HDX optical switch this quarter and says that it’s being “evaluated by customers.” No news yet on who the customers are or whether they like the HDX. All the same, it might give Ciena’s CoreDirector a run for its money. The future of Nortel’s PX photonic switch, derived from its acquisition of Xros, is much less assured.
Sycamore Networks Inc. (Nasdaq: SCMR) is running out of time, according to the report, which pours cold water on recent speculation that Sycamore might make an interesting acquisition. Its large cash balance ($728.5 million in the most recent quarter) may be inviting, but this has to be set against Sycamore being “nowhere near break-even profitability.”
Tellium Inc. (Nasdaq: TELM) is improving revenues, but this gets short shrift in the report, which points to “a myriad of red flags” in the company’s financial figures. This includes “unacceptable” inventory metrics and operating cash flow that was not only negative to the tune of $58.3 million in the first nine months of 2001 but was also worse than the comparable year-ago figures. “A very narrow customer base and a lack of profitability exacerbate the aforementioned concerns,” the report adds.
— Peter Heywood, Founding Editor, Light Reading http://www.lightreading.com
Editor's Note: Light Reading is not affiliated with Oracle Corporation.
re: No Riches From Optical Switches Hello I am new to this industry, but I'm an investor. I have been following a company called. Symbol: LGMR "Light Management Group", www.lgmr.net. The switch looks good, but it appears they are ahead of the industry.
Could you provide some other company names who are building sound based optical switches, or other companies building optical swithces.
Your comments have been impressive. Especially number six: Sixth, a point that Huber does not appear to understand (and neither does opto apparently), although others in Corvis have explained this in more than one presentation, the real reason why optical switching is only beginning to take off but will accelerate, is because the network capacities still can't easily fill wavelengths, so grooming switches are still needed, but they are now just reaching the point where it make sense to provision point to point wavelength services and hence all optical switches.
re: No Riches From Optical Switches Any mention of Brightlink or Calient in this report? Do they break out the different market opportunities between OEO w/grooming (STS-1), OEO non-grooming, and all optical?
Maybe you should just look at any article published here as a trigger to start an interesting discussion, rather than expect the article to consider all aspects of the problem, which is a pretty tall order in this environment anyway. Sometimes I don't even read the article and just go straight to the discussion. It's fun and very informative!!
re: No Riches From Optical Switches There has been more than enough confusion in the switch market and Lightreading is not helping. The distinction should never be OEO versus OOO. The distinction is whether the switching is done in a 1310 or 15xx environment.
Companies such as Ciena (CoreDirector), Corvis (Baylight/OCS), Nortel (HDX), Sycamore (SN-16000) and Tellium (Aurora) are building switches that switch 1310 since their equipment are sandwiched in between DWDM transponders, connecting optical point-to-point links to each other.
In this environment, the switch fabric is electronics (aka OEO) and the long term survivors are those with STS-1 grooming capabilities (sorry, Tellium) since they must be backward compatible with voice circuit. So in a sense, these are not G«£opticalG«• switches but more accurately, G«£optical-layerG«• switches replacing SONET add-drop multiplexers and digital cross connects.
Optical(-layer) switch is the only bright spot in an otherwise dismay industry (CoreDirector growing from 10% to 30% of CienaG«÷s quarterly revenue is a sign of tremendous success, not failure). But in any case all of these manufacturers have also announced product roadmap that would include an all-optical photonic switch fabric (aka OOO) in a hybrid (Ciena' terminology) or multiband (Corvis' terminology) architecture to manage express traffic (typically 70% to 80% of total traffic).
However, since these OOO switches are also behind transponders, they also operate in an OEO environment. However, in this environment, the value proposition of the photonic switch is not the elimination of optical-electrical conversion but that they can switch wavelength much more cost-effectively.
On the other hand, all-optical photonic switches are also being deployed in a 15xx environment (i.e., no DWDM transponders). These can be found in the long haul applications (Calient, Innovance, Lucent/Lambda-Router, Nortel/Xros/PX, Photonex, Solinet, etc.) or in the metro (Opthos, Movaz, Newtork Photonics). Obviously this group is mainly startups and they tend to be the pioneers, making inroads in a market that does not yet exist. So only time would tell if there is a future.
So in a sense, although most of us have not seen the Optical Oracle report, one can already see that it is misguided. If their targets are Ciena, Corvis, Nortel and Sycamore, they couldnG«÷t possibly draw any conclusion on the future of OEO versus OOO since these are OEO manufactures and they have all committed to deploying OOO as an add-on feature.
In fact, the more successful they are in deploying OEO switches, the more likely they will deploy OOO switch to distinguish each other and to gain competitive advantages.
To them, this discussion is never about OEO versus OOO. OEO is about grooming STS-1 voice circuits onto wavelength and OOO is about switching wavelengths. They do not compete, they complement.
It never ceases to amaze me the lack of understanding Lightreading has in the market that they are supposed to serve. I just hope that someday Lightreading can be a little less sensational and a bit more investigative. But then I guess I would probably have more luck with the National Enquiry.
re: No Riches From Optical Switches Since Lightreading is not doing their job and mine is not very interesting at the moment, I guess I can continue.
The point about all-optical networking is not about eliminating OEO conversions. We need OEO conversion in order to extract revenues (we as consumers do not pay for photons, we pay for electrons).
In fact, the point about all-optical networking is about eliminating G«£non-revenue generatingG«• OEO conversions such as regeneration, switching and conversion. The more G«£non-revenue generatingG«• OEO conversions that we can eliminate from the optical network, the cheaper it is to operate the network.
Companies that have extended the reach of optical network without requiring 2R or 3R regeneration by using technologies such as Raman amplification are solving the first problem.
OOO switches (for example, 2D and 3D-MEMS, liquid crystal, bubble switch, etc.) are solving the second problem. But deployment of all-optical switches will be very limited if the problem of wavelength blocking (i.e., lack of wavelength conversion) is not also solved simultaneously.
On the other hand, today OEO optical switches (i.e., CoreDirector and the like) are being deployed very successfully not because they are designed to change the network into an all-optical network but to actually put G«£networkingG«• in the optical network.
Up to now, the optical network industry is a really a point-to-point transport industry. If you start with a fiber, the two ends of the fiber are DWDM terminals (made by Nortel, Ciena, etc.). Connecting to the DWDM terminals are SONET add-drop multiplexers (ADM, made by Fujitsu, Alcalel, Nortel, etc.), which are then connected to Digital Cross Connects (DCS, made by Tellabs).
The ADMG«÷s and the DCSG«÷s are legacy (ADMG«÷s were deployed when the optical world was TDM and DCSG«÷s were deployed even before fibers). If you ever visit a CO, you will be amazed by how much legacy equipment are necessary to connect just two DWDM fibers (it is as if you have two bullet trains interconnected by hundreds of rickshaws).
The new class of optical switches (i.e., CoreDirector which was invented by Lightera, the company that was acquired by Ciena) replaces both the ADMG«÷s and the DCSG«÷s. It maintains the switching granularity of the DCS which is STS-1 (the synchronous transport equivalent of DS3) while at the same time allows a very broadband cross connect up to OC-48.
The end result is phenomenal. Each CoreDirector replaces anywhere from 15 to 25 racks of ADMG«÷s and DCSG«÷s. It does exactly the same function as the legacy equipment (saving both CapEx and OpEx) but at same time can speed up service provisioning from months to minutes (new revenue opportunities). They also allow the carriers to eventually migrate from ring architecture to mesh (better use of network capacities).
CoreDirector (and products like it from other vendors) will change the industry forever and will separate the guerillas from the dinosaurs.
On the other hand, OEO switch that does not have STS-1 grooming capabilities (i.e., Tellium) are not backward compatible and therefore not long term viable (note the shut down of Monterey Network by Cisco and the dropping of the Titan 6700 by Tellabs). Their only potential customers are those with no voice traffic (i.e., CLECG«÷s) which are few and getting fewer.
Interestingly, in the past when traffic is dominated by voice, most are local. Now with traffic dominated by data, most are long distance. Therefore, at these so-called aggregation nodes where the traffic is being handed off from the metro core into the long haul network (and vice versa) which is where the CoreDirector is being deployed, most of the traffic that comes into the node in fact wants to go to the next node (and not into the metro rings).
In other words, on average 70% to 80% of the traffic that go into a CoreDirector are in fact really pass-through traffic (or so-called express or transit traffic) which is why Ciena (and Corvis, Nortel, Sycamore, etc.) have committed to a hybrid architecture where the express traffic can be switched by a OOO switch and the local traffic is groomed by a more expensive OEO switch. This hybrid or multi-band architecture is not driven by the component suppliers but by the carriers (which have less money to spend these days and a stronger desire to be profitable).
But notice that in this case, since the OEO switch sits between two transponders, all photonic circuits are already of the same wavelength (i.e., 1310), so that wavelength blocking is not an issue and deployment of OOO switch is a no-brainer.
On the other hand, in other applications (both in the long haul express layer and in cross connecting metro rings) where you want to eliminate transponders and where the wavelengths are at the native 15xx, then even if we have all-optical OOO switches, we still have problems with wavelength blocking.
So to the extend that OOO switches can be deployed in these architectures, it will be limited. But in another year when we will start to see low-cost tunable lasers coming from a variety of quality suppliers, the all-optical wavelength conversion problem will be solved and we will start to see meaningful deployment of all-optical networking using OOO switches.
Every revolution takes time. This is not about find a villain (and forcing OEO and OOO switch into a death match) but about having the desire to understand the industry that we are supposed to serve.
re: No Riches From Optical Switches You guys are missing the point of the article. It's not a discussion of OEO versus OOO, which is better, or even if they belong in the same market segment.
This is an article on whether or not anybody is making money selling optical switches. Remember, the reason that businesses exist? Sorry, but the multi-billion dollar giant sandbox for engineers to play in no longer exists.
I agree completely with your statements. And in this article, Lightreading missed some points. But anyway, the message that i got from that article is that All Optical Networking is still in its infancy, and will take some time to be implemented.
Another message that is important, is that grooming is still very important on the transport network, even though we now have a lot of pass-through traffic going on.
I don-¶t know how are the COs in USA, but here in Brazil i work on the biggest long distance carrier and STILL, the bandwidth requirements do not make viable a photonic switch. Even if we didn-¶t have the wavelegnth blocking problem.
I guess the Brazilian market i a good representation of the CALA market.