Sailboat you were one of many that doubted INFN's commercial viability so your thoughts, comments and opinions are not creditable. So I would stay far away from any INFN related topic. You don't know what you are talking about. That's obvious and proven. My last comment on this topic. 

Ninja, it is and always has been that you are an Infi pumper.  At least you are consistent.  few years ago you were claiming you did not work there in your posts.  Then later you claimed you had no "insider" knowledge.  Then later you finally admitted you worked there when it was obvious from the start.  Clearly a pumper.

And your assertions that others do not know how to duplicate or improve on Infi's PIC approach is complete BS.  

There are only 2 issues preventing someone else from taking that approach.  1) investment.  No one is going to invest over $350M in a new systems start up today.  2) the approach ties you to only one solution architecture and prevents adopting the latest technology from other suppliers.  Many have analyzed this and decided they do not want those constraints.

But I guarantee there are at least 3-4 groups around this industry who could do what Infinera has with the investment levels Infi received.

Infi was not even the first to do a PIC.  Not even close.  

While I respect and appreciate the work Infi does, and think they are a good bunch of folks over there, it never the less cheapens the perception of them in the market when you post things that are so obviously not true.

You are clearly an Infi pumper. Always have been.

sailboat

Your comment regards power efficiency is not accurate.  You assert that PIC's are more efficient.  While I  agree with you that PIC's pack a lot of stuff into a single optical package, I  disagree with you regards your power assertions.  

Both systems require drivers, electronics, lasers, modulators, detectors.  While I have seen nice forward progress in PIC functionality, I also am quite knowledgeable about this space, and it is absolutely not true that other approaches consume more power.

Yes, many large modules do consume more power, but those are older module designs.  Modules with best of breed discrete components are significantly more efficient.

You can argue PICs save manufacturing costs for the OEM as many functions are inside one co-packaged module.  

But a significant savings in power?  No.

sailboat

I am not a pumper of INFN. I just happen to have spent the last 15 years in the Optical industry and understand optical system technology and cost. What I mean by re-piping is "large deployments". Even though ALU and CIEN have 100G product you will agree that the amount of 100G being installed today is fractional at best. The essence of this article is that the 100G product is not economically feasible at this point in any large deployment. So as you correctly point out ALU and CIEN are the only suppliers shipping 100G product. The price points they offer (due to the discrete components that are at the mercy of using) obviously is not even close to what carriers would require to consider the switch at this time. As far as problems popping up. Sure they will and have. It may be a 100G variation of the PIC but more problems have been resolved as a result of there prior experience developing and solving 10G hurdles then any forward looking unknowns at this point in the development cycle. Based on what I have observed, INFN is an extremely conservative company. So the fact they have announced a launch date and a product GA date tells me they are pretty sure of what hurdles are ahead of them and what's needs to be done. I am sure ALU and CIEN are not looking forward to the INFN announcement. Having a product and it being cost prohibitive is not a good place to be. Driving significant costs down happen within the initial design phase of any product. Once the product is shipping the only things that will bring cost down will be volume, a redesign or material cost reductions. Unless of course they simply want to loose money when they ship product or be satisfied with low margins that will be obvious to all once quarterly earning are released. Something both ALU and CIEN have been historically good at doing for the past 10 years based on historical income statements. INFN current loses are a result of high R&D expenses that are non-recurring at todays levels in 6 months. So again we shall see.

I am not saying there is anything wrong with PIC technology. It obviously has its merits just like the technology Cien or Alcalu use. Duplication is not as easy as one may lead to believe. Going from 10G to 40G or 100G is not a walk in the park. New problems inevitably pop up.  There are many ways to reduce costs, so saying Alcalu and Ciena won't be adept at doing so, is premature and speculative compared to what you are claiming the cost and performance of PIC at 100g offers. It is obvious you are a pumper of INFN and that is fine since as you note there have been many detractors also posting. I understand that.  What do you mean re-pipe their backbone. Are not the solutions alcalu and ciena offer void of repiping You just drop the stuff in with existing fiber that may be old and flakey and there for years, and voila you just increased your throughput. Not sure what you mean by re-piping?  You seem to be suggesting INFN will be getting the bulk of 100g contracts and alcalu/ciena will be left in the dusk. That my friend seems like wishful thinking, especially when only two companies are shipping this and signing contacts today. In that high technology context, 2012 is a century away.  But you are right we will see?

So everything you said is not anything new. The problem is the bizarre statement you claim in my comment is no different then the numerous comments made on this very site by many predicting failure for INFN. The PIC was doomed to never be a commercially viable product and would only work in a very controlled lab environment. Not!

So after 6 plus years of shipping 10's of millions of nodes without a single field failure with the 10G systems I would not question or doubt INFN future success with 100G regardless of how many stages you claim need to be accomplished. Those stages have been accomplished before so duplicating it is simple compared to the initial accomplished of bringing out the 10G PIC itself. You see when no one has ever accomplished a technology doubt can creep into the minds and confidence of the most ambitious and talented company. However, that never happened at INFN. However to date, no one has even attempted to duplicate it or have the technical knowledge to do so now. ALU and CIEN have always and will continue to be at the mercy of the highest level of technology available from numerous discrete OEM component suppliers. They will continue to utilize the same commercially available parts everyone uses. As a result, they will not be able to control costs and bring attractive economic solutions customers need in order to jump to 100G platforms. You are already hearing that from customers now. That's because ALU and CIEN is all they can see today. No one is going balls out to re-pipe their backbone on those solutions. The beauty of the PIC integration is what differentiates INFN from everyone else in the optical industry and what will bring CAPX to the economic cross points required to convince customers to switch to 100G. 2012 is just around the corner and the 100G market has barely been sniffed out. So in time we will all see. 

The definition of CAPX is the process of making an investment in something at a particular time that has potential for growth down the road. More or less.  Even though the Infinera solution may be a competent solution it has not announced or released.  AlcaLu and Ciena are the only two companies actually shipping 100g product that is being used in live networks.  It is far too early to say infinera has the only 100G CAPX solution available when it has not been announced or released. Proof is in the pudding. Infinera have to announce, release and then see how it holds up in customer environments over time. That's three stages.   They are not even at Stage one yet, while the other two, are at Stage three.  How Can infinera have new and existing customers if they are only at lab, field demo and customer trial stages?  Seems kind of a bizarre statement?

Infinera's 100G PIC system to be announced launched in Q411' with production volumes starting 1H12' seems to be the only 100G CAPX solution available. The chip integration alone will eliminate the need for over 200 passive and active optical discrete components. New and existing customers seem to agree based on the field and lab trials that are expanding rapidly including some very large NA Tier 1 players.

Nah, the Huawei thing is just happenstance.

Here's a behind-the-scenes peek: We editors let the video producer pick the camera angle, and we just stand where we're told.  I rarely check what's behind me, and I bet it's the same for my colleagues.

Angles are chosen mainly for variety (and for lighting and sound, of course).  Huawei did pay to have a table at our event (as did MRV) but they just happened to be in the right place, in this case.

...did Huawei have to pay for that ad space? 

The cost of 100G equipment is one issue, scary indeed.  However, scarier than that is the power consumption of a fully loaded 8T 100G system based on discrete components.  80x100G Transponders.  The OPEX required to the HVAC alone to cool it will be enormous, nevermind the actual consumption of the discreet components.