fiber specification for dwdm

[hashaari]

A question about G.652 fiber specification, can it support in both 1310nm and 1550 nm wavelength ?

Would like to standardize it to one wavelength only throughout the DWDM network, what would be best?

What would be the best physical topology for the fiber, star on ring?

[konigl]

G.652 fiber is standard single-mode fiber. It can support both 1310nm and 1550nm wavelengths. I use it regularly to build networks using both LX/LH and ZX GBICs. It is what's most commonly installed, and what's most often built, unless you ask for something else. Compare that with G.655, which is non-zero dispersion shifted single-mode fiber: not as common (usually must be specified or requested up front), but enhanced performance characteristics which make it more ideal for longer-haul metro and certain DWDM applications.

On the DWDM side of the network, you'll likely be standardizing the on 1550nm wavelength. Your client-side hand-offs may run from 850nm to 1310nm to 1550nm depending on what's being connected. But for transport over the DWDM network, the signal's being converted to one of the ITU-T G.692 channels, which can be found in the 1500nm to 1600nm wavelength range. Cable attenuation in the 1550nm region is lower per kilometer than it is at 1310nm; so you can drive the signal greater distances. (And the NZDSF extends that distance even farther.) Just be careful you don't overdrive the RX optics.

Best physical topology is influenced by several factors: location and number of sites to be connected to the DWDM network; whether you want or need physically diverse fiber optic cable path routing for redundancy or survivability of the network; and economics of what you can afford to pay for the fiber build or lease. Cost is usually the dominating factor. Sometimes stars are cheaper to build, other times rings are; occasionally the solution is a hybrid of both. It depends on the total number of kilometers of fiber, where it's installed (aerial or underground), and whether there are any obstacles in the way like river crossings or railroad tracks. Topology design sessions here usually go through multiple iterations until a best fit for each customer is arrived at: our goal is to meet as many network design requirements as possible, while keeping the cost affordable.

Hope this helps.

[affandi.indraji]

Hi,

I have a question about the SMF.

According to this URL:

http://www.cisco.com/en/US/products/hw/modules/ps5455/products_data_sheet09186a008014cb5e.html

For single mode fiber, it uses core size 9/10 micron (Table 1)

For 10G module:

http://www.cisco.com/en/US/products/hw/modules/ps5455/products_data_sheet09186a008007cd00.html

SMF, it uses core size based on G.652 Standard.

My question is if I have can i use the fibre cable that i used before (core size 9/10 micron) for 10G module?

Thank you.

[konigl]

Yes, you can use the same single-mode fiber with the 10G modules.

However, the distance you can push the optical signal at a particular wavelength may be different, because the optical power characteristics of the transceivers are different between 1-Gbps and 10-Gbps.

If you can, measure the existing fiber with an optical loss test set and a 1550nm source. Determine the optical loss of the cable from one end to the other (through all connectors and splices). Then choose the 10G XENPAK with minimum optical power budget that is greater than the dB loss you measured when testing the fiber.

An example of the different optical power levels between 1-Gbps and 10-Gbps equipment:

With LX/LH GBIC at 1310nm wavelength, you have a minimum optical power budget of 7.5dB.

With the XENPAK-10GB-LR or XENPAK-10GB-LW at 1310nm, you only have 6.2dB to work with.

If the fiber distance between the two transceivers is such that the measured dB loss is greater than 6.2 but less than 7.5dB, then what worked at 1-Gbps will not work at 10-Gbps. You would have to upgrade the 10G optics to something with more power, like the XENPAK-10GB-ER or XENPAK-10GB-ZR, to make it work over that distance.

Another example:

With ZX GBIC at 1550nm wavelength, you have a minimum optical power budget of 23.0dB (and a miminum required attenuation of at least 8.0dB).

With the XENPAK-10GB-ER at 1550nm, you only have 11.1dB to work with (and a minimum required attenuation of at least 5.0dB).

If the fiber distance between the two transceivers in this case is such that the measured dB loss is greater than 11.1 but less than 23.0dB, then what worked at 1-Gbps will not work at 10-Gbps. You would have to upgrade the 10G optics to something with more power, like the XENPAK-10GB-ZR, to make it work over that distance.

If you are replacing ZX GBIC with XENPAK-10GB-ZR, you don't have to worry about having enough optical power. The XENPAK-10GB-ZR at 1550nm has a minimum optical power budget of 24.0dB (and a miminum required attenuation of at least 11.0dB).

Also, when working with 1550nm optics, make sure you have at least the minimum required attenuation between the two transceivers. Otherwise you risk saturating the receiving optics with light, and may even damage them if the optical power is excessive.