x
Optical/IP

Light Reading Exposes Privates

Once again, Light Reading's editors have reduced all the blood, sweat, tears, and cash that our industry pours into its startups into a nifty little file we affectionately call Light Reading's Top Ten Private Companies. The List's mission, as always, is simple: Survey the telecom private company landscape and try to handicap which companies are headed for liquidity. Light Reading uses an assortment of popular metrics such as funding, revenues, customer connections, and reseller deals to help figure out which companies are most likely to either go public or get bought.

In addition to the usual facts and figures, though, Light Reading also picks companies using our network of well-placed sources, our access to top-notch industry research from Heavy Reading, some subjective assessments -- such as, "Does the CEO's breath smell?" -- and a few cases of contemplative refreshment.

It's that combination of unique industry insight and world-weary gut feeling that has made the Top Ten Private Companies List a hotbed of debate in years past. Occasionally we find unloved gems that the top-tier VC world appears to shun (see WaveSmith Networks), and yet other times we fall under the spell of the Silicon Valley glamour boys who turn out to be duds (see... well, that list is too long to publish here). The current List is a more eclectic mix than usual, with one gigantic Asian equipment conglomerate, a sleepy phone switch vendor, and a big money, big hype, transport company.

The companies on the List this time around are (in alphabetical order): You'll have to actually read the List to find out the order, why we made each pick, and what companies came close. But if the very List itself stirs you, feel free to let us have it on the boards below or in the following poll:

As with everything published on Light Reading, the feedback loop is ongoing. If you have a comment on the companies picked, the ones missed, or the list in general, you may also send a note to [email protected].

— Phil Harvey, News Editor, Light Reading

corvisalum 12/5/2012 | 1:47:00 AM
re: Light Reading Exposes Privates a few points re: scott, doug and stephen's comments -

- regarding the claim that they use optional edfa's to provide customized distances. is it possible they need edfa's because their power levels from the lasers are so low they cannot build a system w/o edfa's? do they offer any system with any reasonable span length where they can eliminate edfa's altogether? if edfa's are not eliminated altogether, this takes one key saving they profess to bring to the table.
- do they offer a full suite of transponders? if not, what "real" services can they transport? if you take away the transponder market, what is their real market size?
- i have talked to many folks who have looked at the indium phosphide technology and cannot envision high yields. the yields will dictate if they can truly generate savings for carriers and profits for themselves. it is always easy to build a few devices, go into trial and lose money on the first few systems. if they are banking on improving yields over time, my advice to carriers: run away from them and continue with current approaches.
- finally, given that most of the national size networks have already been built and are not lacking for capacity, why bother solving problems which, while albeit cumbersome (been there, tried doing that), are not relevant?
- if infinera's plan is to go to the rbocs, my advice: be very careful - you don't have the ops support to even start your discussions with them.

- long-haul is no fun, rboc's less so. tier-2 carriers in europe? what is the total market size again?

not to let you misconstrue this post as totally negative :-), i am quite excited if this concept works out (given the personal pain i have gone through in making long-haul systems practical and workable in the field). the business model is what i would worry about.

ca
douggreen 12/5/2012 | 1:47:06 AM
re: Light Reading Exposes Privates Scott, et all,

One of the things that I find interesting about Infinera is the similarity of their contrarian approach with that of Lightera. (surprise, surprise, since most of them came from Lightera). While everyone else was talking about all optical switches, Lightera was developing an electrical cross connect with STS grooming. I can immaging this analogy being stressed in Infinera's pitches to investors since Lightera was so succesful.

THe reason that they were right in the case of Lightera was that the technology they were replacing in long haul networks was being used primarily to switch at the STS level. So, the equipment fit in with the existing services better. The carriers had much less to change than they would with optical switches. (It didn't fit in so well with the RBOCs, who needed VT and even DS-0 level switching).

In other words, Lighteras contrarian approach was in line with the carriers existing network architecture versus the prevailing marketing hype.

As much as Infinera (and their investors) would like to think that the same contrarian approach will yeild the same results, they will face an uphill battle. The existing networks have been optimized for the last 8 years around the use of optical amplifiers and OADMs. In many cases, their contrarian approach is now going against the carriers thinking. They are now asking for a change versus trying to accomidate the existing strategy.

That doesn't mean they won't succeed. If the savings are really huge, they might. They face more of an uphill battle than Lightera, however, in spite of some of the similarities in philosopy.


Scott Clavenna 12/5/2012 | 1:47:07 AM
re: Light Reading Exposes Privates Stephen,

Good point about eletrical interfaces. I miswrote, meaning 1310 nm optical interfaces. And your point there is right. They will in cases have to deal with SONET.

'tis a tough challenge, indeed.

Scott
OpticOm 12/5/2012 | 1:47:16 AM
re: Light Reading Exposes Privates Well, the most likely scenario, will be that a big carrier will like the system, but will not be particularly eager to buy from a start-up.
The word about the technology and system would then be passed to an incumbent equipment provider, who will buy Infinera right away.
It happened in the pastGǪ.
And about the most likely buyer(s), I think you can figure yourself...
stephenpcooke 12/5/2012 | 1:47:17 AM
re: Light Reading Exposes Privates Scott,

Doug has added some perfectly valid points that I now don't have to. It is obvious that he HAS been there.

I will add my 30 second analysis for posterity anyway...

- Be careful about the terms of your NDA. I don't recall seeing anything about electrical muxing on their website.
- VERY few ADMs have electrical interfaces on their high-speed (or 'line') side. Interfacing to ADMs requires things like SONET/SDH compliance, protection switching, link monitoring, alarms, etc.
- The only real interface available is GigE. Either they are mapping it into POS or muxing to 10GigE optical.
- 10G interfaces will only travel 80-100km on modern-generation fiber on PMD-'quiet' spans (ie: last 3-4 years) without external components. This means substantial thought concerning route selection.
- If you need to send 10 channels of 10G 100km you still need to demux the traffic at the other end (ie: you still need the ADM/router/whatever which you could have bought a native optical interface for in the first place). Where is the business case?
- Optical add/drop can be a pain but, with current 2-stage opamps you can select the wavelength to add/drop in the mid-stage and not have to demux the other n wavelengths just to mux them up again. The point is that you still have a single opamp (EDFA) versus a per wavelength regen.
- You have to look at the numbers as well from a carrier point of view. WhyisWhy made some excellent points in this area but I would add things like OSS integration, training of field personnel, etc.
- this has now become a greenfields installation, therefore there is an enormous amount of competition and is nowhere close to EDFA replacement.

The technology seems to have promise, the system is still questionable in my mind, however the only minds that count are those of the carriers (unless they go the component route).
whyiswhy 12/5/2012 | 1:47:19 AM
re: Light Reading Exposes Privates If you only need to send 10 channels of 10G 100 km, you reduce or eliminate the need to include much in the way of dispersion compensation and the like.

That's the hype. It's true but not the whole truth. Whole truth is more like: Nobody wants to do 10 channels, they want to do 80 channels (broad brush I admit). And 80 channels means there is another layer of AWG (loss) involved, and the distance shrinks to about 80Km at best with NZDSF.

Anything else needs an EDFA (or gag, Raman), and what have you gained cost-wise or simplicity-wise?

This is at its best a Metro-ring technology, not something for LH. I agree with Doug, the architecture for that was decided awhile back: ROADM's with express channels.

I reiterate: Infinera should go to 50 micron MM, 10G Ethernet, and sell to enterprizes who are much less picky about reliability and qualification....they just need cheap.

If they can get a foothold there due to lower price, they can grow into LH from the bottom up. China and India would eat it up, if the price was low enough. It's a classic low ball approach, but not one VCs like one bit, since they have to stick in there with operating losses for many years.

-Why

Scott Clavenna 12/5/2012 | 1:47:20 AM
re: Light Reading Exposes Privates Doug,

Good points. I think Infinera has addressed the first one by including optional low-cost optical amps (booster, inline and pre) to give some flexibility in span length, and to provide an option for traversing a hut optically if there is no anticipated add-drop anticipated ever (middle of the desert or the like). They definitely have to prove the low-cost claims, however.

On the second point, my guess is that they will not focus on RBOCs initially (or IXCs like MCI, who believe in the all-photonic future), but look to regional operators, Euro city carriers, and those folks building networks to serve Tier 2/3 metros. This is a potential area of opportunity for Infinera since they can support lots of add-drop on a regional network without much advance network planning.

The question is what's riskier: waiting for RBOCs and fighting their religious wars or going after the remain facilities-based regional CLECs and euro city carriers to build a footprint and prove the economics? Neither choice really makes you jump out of your seat, but the promise of this technology is keeping investors committed and relatively sanguine on the prospects.

Scott
douggreen 12/5/2012 | 1:47:20 AM
re: Light Reading Exposes Privates Scott,

Going after smaller regional carriers sounds good in theory. My experience is that startups are at a disadvantage. Larger companies can put together "package" deals that include installation, training, and other types of equipment. Smaller carriers also often depend on using well known equipment to legitimize themselves to their own customers.

I wish them luck, however. I've been there.
douggreen 12/5/2012 | 1:47:21 AM
re: Light Reading Exposes Privates Scott,

A couple of points.

1. For long haul networks, one would still want to have things like FEC, pre-amplification, etc. Distance between huts is an issue, regardless of whether you put an amplifier there or de-mux and put an OEO into the equation. In fact, the loss on the de-mux (which is not there with an amp) has to be thrown into the equation. This is not an issue on the total route, but on the individual span. Every re-gen site adds a huge cost, so span distance is a big cost factor.

Second, if the sweet spot for these guys is regional networks, then the sweet spot will likely be RBOCS. This is both good and bad. It's good because a true electrical ADM will fit into the Telcordia network management models much easier than a ROADM. It's bad because the RBOCs aren't likely to qualify a new system as quickly and are much more "religious" about eliminating OEOs. Who wants to go back to his boss now and say: " I know I've been spending all my time on ROADMs for the past 2 years and we are just about ready to use them, but I want to hold off another two years to look at this new technology." In the real world, religion and fear are factors.

I am very intrigued with the technolgy and think it has promise. I just hope their investors have a lot of patience given that this company has already been around a while.

We'll have to wait and see.
Scott Clavenna 12/5/2012 | 1:47:22 AM
re: Light Reading Exposes Privates "I hope that you have NDAs with Infinera and that there is something behind this because these statements are pure fluff. Anyone who says things like "The value over existing EDFA-based DWDM systems is that an operator no longer has to worry about wavelength band planning, gain management, optical signal conditioning and the like." either understands very little about 10G transmission, has bought the farm being sold by a smooth-talking marketeer, or the laws of physics have been repealed."

Stephen,

True enough. I do have an NDA with Infinera, so have had a good look at the technology and system in operation.

This is the key thing to get across in this discussion: Infinera is building a long-haul/regional optical transmission system. Therefore, it's role is just like any other DWDM transmission system: it takes in electronic signals from ADM, routers and switches, electronically muxes them up into 10G signals (if necessary), muxes those optical using WDM onto a single fiber, then launches them down the long haul route.

The difference between Infinera and a typical EDFA- or Raman-based DWDM system is that the system isn't optimized around eliminating electronic regen (using advanced FEC, super amps, chromatic and PMD compensation and the like to extend optical tx distances) but instead is optimized around reducing the cost of OEO regen using photonic integrated circuits.

Since it performs OEO regen at every node (80-120 kms apart typically), the challenges in the optical domain are greatly reduced.

If you only need to send 10 channels of 10G 100 km, you reduce or eliminate the need to include much in the way of dispersion compensation and the like. The costs of signal conditioning are reduced or removed for the system, therefore. A limited amount of optical amplification is required in some cases, but very low-cost EDFAs can be used to compensate for loss in the optical muxing stage, rather than expensive amps that must account for accumulated noise, transients, etc. along an entire route.

Second, add-drop should be greatly simplified, no? Optical add-drop is a pain, by anyone's account. It messes with the wavelength plan, it disrupts the gain profile of a given route, and can be quite disruptive if not planned into a network from day one. Add-drop with Infinera occurs electrically, so nothing is really disrupted; it was going to regen electronically at that node anyway, so why not peel out the signals you need or add more. There is no effect, per se, on the optical layer of the network in that case.

The concerns around Infinera are valid, for certain: the long-haul transport market is soft, to say the least; carriers don't like to buy from startups; photonic integrated circuits are rather risky technologies. But the goal is laudable: build network infrastructure for optical transport that addresses the challenges around a reconfigurable optical layer by addressing the costs of OEO regen rather than the technical constraints of optical engineering.

Scott

whyiswhy 12/5/2012 | 1:47:32 AM
re: Light Reading Exposes Privates "On backwards compatibility, it is more backward compatible than most any DWDM system today, since everything remains at the electronic level, meaning typical SONET and GigE/10GigE interfaces. It actually plugs in quite nicely, as far as I can see. But of course the word isn't in yet. It is mostly promise at this point, but an interesting one..."

And if the carrier elected to rely on dispersion compensation, either in their modulation format, and/or with fixed or tunable DCM's, how does an "Infinera OEO module" work? And what if the point to point distance (and dispersion) between Tx and Rx changed?

I mean, wouldn't the Infinera module take the dispersion (carefully introduced mind you) and throw it away! That's presuming it could even detect a useful (high Q) eye in the first place!

There is not one good PHY-level optical com guy in the company! Perkins is an internet protocols guy! Welch is a laser diode guy!

Scott, it would take a wholesale adoption of the Infinera "OEO node solution" to make it work. I.e. no dispersion compensation, no chirping, etc. In other words, a total greenfield start from scratch "solution"!

My advice to Infinera: Do it in 50 micron multimode, with mixed 1 and 10GE/channel, and push it in metro and/or enterprize and/or SAN. That might have a chance, if it was dirt cheap, and included an electrical "grooming" ADM (OEOEO: three leg) as a variant.

It's the only greenfield out there.

Maybe not Kestrel, how about Centerpoint?

-Why
stephenpcooke 12/5/2012 | 1:47:33 AM
re: Light Reading Exposes Privates Doug,

I am sure that there are good applications for their technology and I have suggested one (SR switch interfaces). I feel that, assuming their technology is good, they may be missing the boat.

I have designed network equipment and troubleshot huge optical networks and can relate to your comment about misguided business cases. One wonders how some of these decisions get made. Sometimes they are the right technology in the wrong market.
douggreen 12/5/2012 | 1:47:33 AM
re: Light Reading Exposes Privates Stehpen,

There is some validity to Infineras claims in the fact that electronic regeneration can eliminate the need for dispersion compensation as well as PMD. You probably wouldn't replace all EDFAs. It would seem to be most useful in places where you need OADMs along the routes.

The issue, as always, is that networks are complex, and when you save money in one area you often generate expense in another. There is no way to tell until you see real routes modeled by a network planner who is given the equipment specs and costs.

I can't count how many business cases I've seen that look great when you look at a single network element, but fall apart when real network modeling takes place. That doesn't mean that Infinera won't come out looking great. It just means that we've not seen any proof yet. I'll be convinced once a carrier network planner gives some hard savings numbers in a real network configuration.





stephenpcooke 12/5/2012 | 1:47:35 AM
re: Light Reading Exposes Privates Hi Scott,

You've lost me so I'll have to go through your post point-by-point:

I'm not sure where your strong opinions on Infinera are coming from. Infinera is a selling a system, not a technology or a device, and the system is based on photonic integrated circuits that allow full electronic regen at every node (or inexpensive EDFA to extend span lengths to a couple hundred kms). That said, this system is not a replacement to EDFAs, but an architectural alternative to DWDM systems that use EDFAs.

As I have said in previous posts, I know only as much about Infinera as was contained in the articles. I was making a suggestion that they focus on selling the technology or devices instead of the system because I couldn't see the use for the system as described.

With an Infinera system, wavelength assignment doesn't matter because wavelength multiplexing is only used for fiber gain between nodes, not wavelength-based networking.

Here's where you lose me completely... "With an Infinera system..." Are you assuming that Infinera equipment and only Infinera equipment will be used in this network that you describe? If so, how does the traffic get into the Infinera system? I even looked at their website and it says butkus about how this system would be used. Here are some basics:

- Integration is a two-edged sword, it is often great for getting more functionality into smaller packages but it generally comes at a cost, yield and flexibility.
- Carriers buy a bunch of multiplexers to get their traffic aggregated up to 10GB/s rates. These multiplexers come from companies like Nortel, Lucent, Cisco, etc.
- These multiplexers have transmit wavelengths that are on each carrier's individual wavelength plan. Most wavelength plans are on the ITU grid but not all are used and many are skipped here and there.
- 10GB/s links are not generally loss-limited (ie: bad things happen to the signal before the power is gone). That is why there are these other components (dispersion compensators, etc.) that attempt to clean up the signal so the signal can be interpreted accurately at greater distances. This is especially true on older fiber plant (ie: most fiber in use today).

10 channels of 10G get wavelength multiplexed together, then further multiplexed with seven other sets of ten onto a single fiber then transported to the next node, where they are all demultiplexed and regenerated electronically.

So, you have 80 10G wavelengths travelling down the fiber and you want to de-multiplex them to electrical (I hope that they at least terminate the SONET Section OverHead so that BIP-8 values can be examined and errors reported somewhere) just to clean up the signal...? Ballpark what is the cost of an 80 wavelength Infinera rack including demultiplexers and power? How does that compare with a gain-flattened, single circuit pack EDFA? Use an EDFA from Nortel or others and you will even get an Optical Service Channel for remote monitoring and control purposes.

Currently there are extremely few, I would guess zero, single fibers around that are carrying 800 GB/s worth of traffic. There are many more that have 80 GB/s but those wavelengths may be distributed throughout the ITU grid. It is quite conceivable that a carrier would have only one wavelength in the range of each Infinera Photonic Integrated Circuit (PIC). This would require a fully populated shelf for 8 wavelengths; doesn't sound too efficient to me.

With this kind of solution, a network operator can build arbitrary, rather than esoterically engineered, optical transport networks. They get the benefit of arbitrary add-drop at any node, and no limitations on network size or scope. The value over existing EDFA-based DWDM systems is that an operator no longer has to worry about wavelength band planning, gain management, optical signal conditioning and the like.

I hope that you have NDAs with Infinera and that there is something behind this because these statements are pure fluff. Anyone who says things like "The value over existing EDFA-based DWDM systems is that an operator no longer has to worry about wavelength band planning, gain management, optical signal conditioning and the like." either understands very little about 10G transmission, has bought the farm being sold by a smooth-talking marketeer, or the laws of physics have been repealed.

This is what is meant by digital optical networking rather than analog optical networking. It is digital in the sense that the analog part of the network (WDM, optical amps, signal conditioning) has been either reduced or eliminated altogether, and true networking/routing functions take place in the digital/electrical domain rather than the analog optical domain.

Scott, now I understand some of your comments but, unfortunately they point to you lack of understanding of data transmission. 1's & 0's are transmitted via analog signals in all cases. We call them digital because we only sample 2 different states. If you look at the optical waveform (also referred to as the optical eye) that travels down the fiber, it is anything but 'digital'. Things like rise time, fall time, extinction ratio, pulse width, etc. all play extremely important parts in the transport of data. If any one of them are neglected the link simply won't work. When you are talking about 10G transmission you have a bit period of 100 picoseconds but a realistic eye opening of about 30ps. This leaves VERY little room for error or tolerance for signal distortion.

Here is a short question & answer tutorial on optical transmission:

Q: Why are modulators used in the optical domain (external modulation) instead of simply turning the lasers on and off directly (direct modulation)?
A: Turning lasers on and off at high data rates (eg: 10G) causes additional pulse distortion due to a phenomenon called local heating. This problem is very pattern sensitive (ie: gets better or worse depending on the number of consecutive 1's or 0's transmitted). With bit periods of 100ps and non-optimal fiber plant there is simply no room for additional pulse shrinkage.

Q: Are all of the modulators on the Infinera PIC the same?
A: They had better not be. Electro-optical modulators are based on a process of constructive or destructive interference in the optical domain. The incoming optical signal is split and fed through two different paths. One path can have its effective optical path length changed by the application of an electrical voltage. The path length is changed in one arm by EXACTLY 1/2 wavelength for destructive interference (ie: optical '0'). With no voltage applied the optical paths are identical, therefore constructive interference at the output (ie: optical '1').

Q: If the modulators were all the same in every detail what would suffer?
A: If the optical path length were not changed by EXACTLY 1/2 wavelength then the interference would not be complete and there would still be light transmitted in the case of an optical '0'. This means that the extinction ratio would suffer. As there are 10 different wavelengths in an Infinera PIC each modulator has to be mated with its own specific wavelength.

I will re-iterate, I know virtually nothing about Infinera or its technology.
dbostan 12/5/2012 | 1:47:37 AM
re: Light Reading Exposes Privates I think the technology is ground breaking. They will go a long way, if they can deliver, i.e., it is an operations matter now.
Scott Clavenna 12/5/2012 | 1:47:37 AM
re: Light Reading Exposes Privates We'll see about the kestrel comparison.

On backwards compatibility, it is more backward compatible than most any DWDM system today, since everything remains at the electronic level, meaning typical SONET and GigE/10GigE interfaces. It actually plugs in quite nicely, as far as I can see. But of course the word isn't in yet. It is mostly promise at this point, but an interesting one...

Scott
whyiswhy 12/5/2012 | 1:47:40 AM
re: Light Reading Exposes Privates ...or a dream, Scott?

Ever notice companies call their products "solutions" when they are too expensive or complicated to call them "products"?

Pretending service providers are going to re-learn and re-engineer their systems to put this stuff in is really smoking it! We all know how much greenfield is going in...zero. There is still a ton of dark or under-lit fiber out there.

If it is not backward compatible with existing systems, i.e. not an EDFA replacement, and it would appear it is only marginally so, it will be ten years before it sells in volume....if ever. By then, there will be something even better.

Optical FDM is a good idea too, from a pure communication theory POV. It even made good sense for PON. But it was and is not backwards compatible with TDM DWDM, at least not easily. The only systems out there are hertiage HFC, and they are going the way of the dinosaur.

Infinera is Kestrel spelled sideways.

The only difference here is KPCB has so much money, they can flex their keiretsu muscle to make it look like this stuff is hotter than spit!

Global Crossing here we come!

JMHO

-Why
Scott Clavenna 12/5/2012 | 1:47:41 AM
re: Light Reading Exposes Privates Stephen,

I'm not sure where your strong opinions on Infinera are coming from. Infinera is a selling a system, not a technology or a device, and the system is based on photonic integrated circuits that allow full electronic regen at every node (or inexpensive EDFA to extend span lengths to a couple hundred kms). That said, this system is not a replacement to EDFAs, but an architectural alternative to DWDM systems that use EDFAs.

With an Infinera system, wavelength assignment doesn't matter because wavelength multiplexing is only used for fiber gain between nodes, not wavelength-based networking. 10 channels of 10G get wavelength multiplexed together, then further multiplexed with seven other sets of ten onto a single fiber then transported to the next node, where they are all demultiplexed and regenerated electronically.

With this kind of solution, a network operator can build arbitrary, rather than esoterically engineered, optical transport networks. They get the benefit of arbitrary add-drop at any node, and no limitations on network size or scope. The value over existing EDFA-based DWDM systems is that an operator no longer has to worry about wavelength band planning, gain management, optical signal conditioning and the like.

This is what is meant by digital optical networking rather than analog optical networking. It is digital in the sense that the analog part of the network (WDM, optical amps, signal conditioning) has been either reduced or eliminated altogether, and true networking/routing functions take place in the digital/electrical domain rather than the analog optical domain.

Scott
coreghost 12/5/2012 | 1:47:47 AM
re: Light Reading Exposes Privates Your right about SAIC. Its one of a number of
companies thats an extension of the US government.
It was even started by people who came out of
the national labs. Telcordia came from Bellcore
and Bellcore was part of ATT when it was a giant
monster attached to the government.
stephenpcooke 12/5/2012 | 1:47:48 AM
re: Light Reading Exposes Privates The Infinera device is basically an EDFA replacement.

I don't think that their device can replace ANY EDFAs. On new routes it may be used instead of EDFAs but this is also questionable. The architecture is yield-averse and the wavelength selection is fixed, not promising EDFA replacement characteristics. As I have said before its most likely path of success is short reach switch interfaces, not regens.
pearlh 12/5/2012 | 1:47:50 AM
re: Light Reading Exposes Privates The military world is closer than you think.
This link will take you to a job offer of
TelecordiaGÇÖs (maker of OSMINE) parent
company SAIC.
http://jobs.saic.com/ajobbext3...
PO 12/5/2012 | 1:47:52 AM
re: Light Reading Exposes Privates The average MacDonalds franchise costs $45K, and grosses $1.3M per year.

This $45K is just the franchise fee. Then you've gotta outfit the place ($500K - 1M ?), and you're paying 16.5 percent of gross revenues in various fees (according to the cited article). Even assuming "free" finanacing, cheap rent, and labor costs under control, the ROI isn't quite as sweet as the first blush might suggest.

Don't believe everything you read - especially from CNN. There are better resources for evaluating franchisors.
whyiswhy 12/5/2012 | 1:47:52 AM
re: Light Reading Exposes Privates The Infinera device is basically an EDFA replacement. Let's do the math on quarterly reports and some ROM assumptions.

JDSU did $160M in their latest quarterly report, and they own about 60% of the optical components market. Of that, let's say 30% was in EDFAs. High I admit, but let's go with it. That makes the total EDFA (or its replacement) market about $360M this year.

Presuming they solve the cost/yield and reliability problems, so they can make a gross profit, and they can price to get a fraction of the market.

Tidbit: The average MacDonalds franchise costs $45K, and grosses $1.3M per year. If the $280M was invested in a group of MickyDee franchises, they could expect to gross $197M per year.

http://money.cnn.com/2004/04/2...

Pretty sure no one in optics can do that, not even Cisco (though they will come close), let alone start-ups.

-Why
Curious George 12/5/2012 | 1:47:55 AM
re: Light Reading Exposes Privates Let's see, a capitalization of $280M to date against a total '04 (RHK estimate):
- DWDM TxRx module market of $420M in '04
- Sonet/SDH TxRx module market of $708M
-combined Ethernet/Fiber channel market of $780M

How does this make sense much less make the list of top private companies? Maybe because only a private company could push a business case like this through their board and still keep their jobs...

Seriously, what was the selection criteria that placed them on the list?



PO 12/5/2012 | 1:47:56 AM
re: Light Reading Exposes Privates Nortel's new CEO, William Owens, was second in command of the US military at one stage and folk think that's a good thing.

Huawei's founder, Ren Zhengfei, was in the Chinese military at one stage, and folk think that's a bad thing.

Double standards?


Hmm. LR editors don't seem to differentiate between the US military and the Chinese military as training grounds for business, and they're hypersensitive about disclosure of their own relationships with companies on the list.

Is it any wonder that nobody seems to take the list seriously?
whyiswhy 12/5/2012 | 1:48:00 AM
re: Light Reading Exposes Privates Peter:

Who really thinks Owens will clean up the mess at NT? He is an insider, and an obvious target of shareholder lawsuits. He needs a defense strategy, period. Just asking/speculating, I have no factual basis.

Huawei does what they have to to survive, and then some. How much do you think they charged the NSA for the back door keys to those "bad" systems? Just asking/speculating, I have no factual basis.

-Why
coreghost 12/5/2012 | 1:48:01 AM
re: Light Reading Exposes Privates
It might be a double-standard but for the fact
that China is no sort of democracy,
Huawei has a long history of doing business
with the PLA and doing the PLA's dirtywork
in dual use sales overseas. Maybe some believe
that Huawei's interest in "growth markets" like
Iraq and Afghanistan was just part of their
great business strategy. But anyone with any
sense knows those deals were done for reasons
that have nothing to do with business.

I'm not aware of nortel being founded by army
money and I'm not aware of nortel working with
the army to sell equipment into countries under
international sanctions.

Nortel is a bad example. There are american
examples in terms of the large number of companies
that do outsourcing for the US military, NSA,
CIA and national labs with a revolving door
between them and the pentagon. But the american
examples don't really change what Huawei is.
captain kennedy 12/5/2012 | 1:48:02 AM
re: Light Reading Exposes Privates Peter,

You have got to be kidding!
Peter Heywood 12/5/2012 | 1:48:04 AM
re: Light Reading Exposes Privates Nortel's new CEO, William Owens, was second in command of the US military at one stage and folk think that's a good thing.

Huawei's founder, Ren Zhengfei, was in the Chinese military at one stage, and folk think that's a bad thing.

Double standards?
allidia 12/5/2012 | 1:48:06 AM
re: Light Reading Exposes Privates Mozart Edge Networks in MA.? Any news?
truelight 12/5/2012 | 1:48:10 AM
re: Light Reading Exposes Privates Interesting post Coreghost,

If true warning to all carriers - Let it be public and your career will be over in this country.

coreghost 12/5/2012 | 1:48:13 AM
re: Light Reading Exposes Privates
I would not call Huawei a private company. I
would (at best) call it a company where the
actual ownership is not clear. The ties to
the PLA are deep and real.

Those ties have over the years been very
apparent in Huawei's sales to Saddam Hussein's
Iraq, the taliban in afghanistan and in
other places where sales were rather obviously
tied to the political objectives of the PLA
outside China. And the rather odd tendancy
of the Chinese government to publically
defend Huawei when its international operations
get it into trouble.

Huawei is an extension of the PLA. And nobody
with any sense could believe otherwise while
Ren Zhengfei remains in charge.


truelight 12/5/2012 | 1:48:14 AM
re: Light Reading Exposes Privates okay - you got me. What do you think of the top ten?
Mezo 12/5/2012 | 1:48:15 AM
re: Light Reading Exposes Privates truelight,

You're finally exposed as a member of Light Reading's staff.

C'mon, your talking to yourself...four times.

Signed,

Article Talk Readers
truelight 12/5/2012 | 1:48:17 AM
re: Light Reading Exposes Privates All in all this is one of the best top ten lists - by LR I have read in a long time.

All are generating revenue (unsure about Infinera) and have major customer traction.

Well done

truelight 12/5/2012 | 1:48:18 AM
re: Light Reading Exposes Privates Is this really a private company in the true sense or is it a communist military state initiated company?
truelight 12/5/2012 | 1:48:18 AM
re: Light Reading Exposes Privates Great product and team the next extreme and foundry. Watch out for riverstone influences.
truelight 12/5/2012 | 1:48:19 AM
re: Light Reading Exposes Privates With all the washout rounds at Calix its in the hands of a few individuals that now control the company. As a result employees are de-motivated, playing golf and short working weeks are the norm.

The product has started to come out of its teething phase but they still face tremendous pressure on the financial side since the product is very expense compared with Alcatel and AFC. If they go public it will be interesting to compare gross margins. Giving away the product and copying other companies ideas appears to be current trend to win the business.

Very ATM centric - old world DLC.

HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE