re: GPON Gets a 10G LookI'm not really in this particular branch of the industry, but I'm curious. How are they doing 10G PON?
I like Salira's Multi Wavelenght PON (MWPON) approach mentioned in an earlier article. For example, you could take a 2.5 Gig PON system split to 32 customers and use four different wavelength pairs. Then all you have to do is replace the optics at the customer site and now you have 8 customers sharing 2.5Gig bandwidth on a wavelength pair instead of 32 customers--a total of 10 Gig. Or you could use 8 wavelength pairs and only 4 customers share 2.5G, and you have 20G total. This could scale up to an individual wavelength pair per customer and 32*2.5G=80Gig total. It is an easily scalable system.
And the customer equipment doesn't change--just the optics. And the CO box just muxes several 2.5G PONs onto one fiber.
Again, I'm not an expert here, and am not sure if it is as cheap or easy as this sounds. That's why I'm curious how the vendors are doing 10Gig.
re: GPON Gets a 10G Look"The article gives the link to the 10GEPON draft."
It is password protected. Though notes on the site say that the PHY is running at 10G. Wouldn't that require all new electroncs at the customer site as well as really expensive optics?
In other words, isn't this much more expensive than the MWPON idea?
re: GPON Gets a 10G LookI was impressed by the headline, then read the details. Multiple wavelengths? That's not 10G PON, that's DWDM PON (at least a flavor of it). Shame on LightReading and many other media outlets for letting Ericsson get away with that misdirection.
The IEEE and FSAN/ITU standards drafts both specify that 10G PON will be a single wavelength up and a single wavelength down. This "4x2.5G" hybrid version of WDM PON is nothing new and has been promoted by (among others) Alcalu and Fujitsu for at least a year ... if not longer.
re: GPON Gets a 10G Lookyou are bringing about a good point...
Folks find it hard to talk about 10G PON without confusing or mixing it technically with WDM PON
I make the same error too becuase 1.) device technologies aren't clear enough to those outside the standards bodies and 2.) converastions of one topic always link a borrow or a steal form the other among our rank and file slaves to Physics and the even more mysterious communications market dynamics.
Economically, 10G PON will be awaiting the point in time when an externally modulated source is hitting price points that the PON market can swallow (at least at the OLT (pricey CO, then later on the cheap MSO/SP). In the mean-time, it must be slipped into networks with the added backbone transport type functionality associated w/10G TDM functions. Note these are mainly DWDM networks.
Upstream PON client optical tech focus is about squeezing blood from a low cost crystal to avoid ex. mods.(e.g. getting the most reach and highest rate out of directly modulated light sources at 1310nm with a lot of burden shifting to the electronics).
It's hard to talk about 10G without throwing in the WDM because device technologies being discussed are either 1.) implementing broadband light sources using quazi-traditional DWDM type optical MUX/DMUX components, or 2.) the straight up ROADM/SDH components with the same components and OA's, with maybe a different twist being considered (i.e. photonic integarated cicuit's).
Frankly, this stuff bores the hell out of the many iii-v oe/eo and straight-optics guys. R&D folks would love the funding to come back into the US budgets again. They hope something new can pop up and disrupt a seamingly plodding march. We feel the optical domain for signaling needs to be better tapped real soon.
So let's acknowledge that from the lowly optical layer stand point, the 10G and WDM can bleed together to us physics slaves who tend to have both problems in the same floaty RFP full of questionable promise.
If your focus is on electronic (CMOS) IP, systems level, and all the firmaware/software logical tweaking that lives at and above the PHY layers, then I think maybe your demarcation between WDM and 10G can seam much clearer.
re: GPON Gets a 10G LookIsn't XAUI 4x3.125GHz considered 10Gig? You can argue about terminology, but I think it is much more interesting and useful to talk about technology, market, and costs.
Verizon is busy wiring up the US with BPON and GPON. It looks like they are about to wire up NYC. rjmcmahon is fond of telling us how many years (decades?) it is going to take Verizon to recoup these costs. So just when do you think Verizon is going to rip out all these customer boxes and CO boxes and replace them with much more expensive 'single-wavelength 10G' PON systems?
...especially when there are much less expensive alternatives to get extra bandwidth to the customers that don't require replacing any boxes at all.
I'm suggesting that the IEEE and FSAN commitees are wasting their time with 'single-wavelength 10G' standards work. But like I keep saying. I'm not an expert in these technologies. I'd be interested in hearing alternative viewpoints.
re: GPON Gets a 10G LookSlick press release and interview... but what aren't they saying? It doesn't take much to get from a working 2.4/1.2G OLT and ONT to 10G/1.2G, and not much more than that to get to 10G/2.4G. Siemens (now N-S) announced they had one of those working in the lab last May, done under an EU project.
Now, if that were 10G upstream, it *would* be an accomplishment.
re: GPON Gets a 10G LookNothing is 'easy'. But no fundamental breakthroughs are needed in OLT receiver design to make 2.4G work with 80 ns guard+preamble+delimiter. 10G is a completely different beast.
re: GPON Gets a 10G LookOf course 8 customers per 2.5G (WDM PON) will mean if 3 people on the PON demand 1G of downstream traffic you are in trouble if they all happen to be on the same lambda. On a 10G PON you are fine. Same goes for multi-cast services, if 4 people all want to watch the same show, but they are different lambdas, the OLT needs to make 4 copies and burn the bandwidth on all 4 lambdas.
There is also the additional operational complexity of making sure the ONTs have the proper optics to match the appropriate lambda.
Of course if the price is right, it could all be OK. Or perhaps we will see WDM PON for business applications where bandwidth demand is more constant in theory and the operation complexities are outweighed by the higher revenues?
All in all it will be interesting to see if the PON folks move to 10Gbps (~300Mbps / ONT), how will the copper folks (cable and AT&T) compete with that?
> how the vendors are doing 10Gig.
The article gives the link to the 10GEPON draft.
http://grouper.ieee.org/groups...
It is nearly completed (I think).