Technology that makes 100g economically feasible: OTN switching & grouped optical transmission (1)

In recent years, network operators have been heavily dependent on ROADM-based optical transmission equipment, using a fixed point-to-point WDN connection, using 10g wavelength in the entire metropolitan area network and wide area network convergence and transmission client services. If these networks are well planned, the existing spectra can be used reasonably and effectively. Typically, a tran0000der transceiver is used to map Client Services with a bandwidth equivalent to a WDM line rate (for example, 10GE or OC-192/STM-64 client business) to a 10g wavelength, the muxponder multiplexing forwarder is used to aggregate Sub-10G businesses (such as GE, OC-3/12/48, STM-1/4/16 clients) to 10g wavelengths for transmission. In 2013, 10 Gbit/s services accounted for over 50% of the ports connected to the optical network, and the average service rate was very close to the WDM line rate. Therefore, the existing optical network bandwidth can be effectively used [III].

Poor bandwidth utilization

Although the network bandwidth demand is predicted to surge, the overall rate of the client ports connected to the optical network cannot change significantly. Even by 2017, services with 10 Gbit/s or lower end-to-end management will still occupy over 95% of the access ports in the optical transmission network [IV]. The transition to 100 Gbit/s DWDM wavelength causes a large number of rate mismatch/discontinuous problems between Client Services and WDM uplink ports, resulting in a large amount of inefficient use of optical bandwidth.

Multiplexing repeater is a static solution. The client port is associated with a fixed WDM Wire port, that is, a specific wavelength. This solution works well when the business needs of all nodes on the network are balanced and fixed. However, in actual network deployment, services are constantly evolving, and business requirements from nodes to nodes are unpredictable in combination and distribution. For network operators, it is a difficult operation to avoid insufficient bandwidth utilization in their valuable DWDM networks.

For example, if a local node (CO) in a WDM or ROADM network needs to add 30 GB of new 10 Gb services, and all available client ports on the node are exhausted, telecom operators have only two options:

1. Deploy the back-to-back muxponder to reuse the repeater to reuse up to 70 Gbit/s of passthrough traffic and converge it with the new 30 Gbit/s client business to a Gbit/s wavelength, or

 

Figure 1: Back-to-Back multiplexing Repeater

2. Deploy a new 30% g wavelength with only utilization.

 

Figure2:G bandwidth not fully utilized

In the first case, the available Gbit/s optical bandwidth can be used better, but the device overhead is higher, which also introduces additional troubles, such as end-to-end connection latency. Considering that new cloud services have a dynamic network demand, it is extremely challenging to effectively plan and manage the solution.

In the second case, although it is easy to manage and deploy, the operator needs to add a new G wavelength, which consumes valuable spectral resources. If the available optical fiber pairs are exhausted on the node, this may increase the CAPEX of Optical Fiber facilities. Because the new bandwidth is only 30% used during deployment, this solution also causes a large amount of idle network bandwidth.

From the operational perspective, both solutions are not ideal. The client service is not only bound to a specific uplink wavelength, but also because the ROADM-based network can only be switched according to the whole wavelength, any network configuration at the business layer, if the client is re-routed or the network is re-optimized, the operator must dispatch technical personnel to the site for manual operation. The interconnection between cloud services and data centers is particularly dynamic in terms of bandwidth requirements, so it is very difficult to use point-to-point transmission facilities.

Because both optical capacity and CAPEX are of high value, many operators are looking for an appropriate network architecture to address the current limits.

Let G practical technology: OTN exchange

Similar to the previous generation of physical layer Optical Interconnection Technology Based on the SONET/SDH transfer protocol, the OTN exchange is also defined as the Cross-interconnection of digital transfer containers, also known as the optical transmission digital unit (oduk ), complies with the g.709 optical transmission network standard multiplexing level. Depending on the service load carried by these transfer containers, the speed varies from 100 Gbit/s (odu4) to 1 Gbit/s (odu0.

The difference between an OTN-based network and a traditional muxponder-based network is that it provides:

1. Support for processing sub-wavelength Client Traffic and
2. Physically isolate Client Services and WDM uplink ports on the network element.

Generally, OTN supports traffic processing and aggregation according to the minimum granularity (1 Gbit/s) specified in g.709, the intermediate nodes in the network can increase or remove client traffic at a granularity much lower than the WDM line rate.

In addition, this wavelet is implemented based on a centralized Electrical Crossover matrix. Any client port on any client line card can be routed to any available WDM line port in the system. Unlike the muxponder-based multiplexing forwarder solution, the client and the interfaces of the WDM line are physically separated, you can decouple the cost of deploying a new client port from the cost of deploying a more expensive WDM line port. The client business is no longer bound to a specific wavelength, which is particularly important in supporting evolving networks.

From the operator's perspective, the adoption of OTN switching in a G network brings the following significant benefits:

1. Improved CAPEX Efficiency
2. Reduced OPEX and faster service speed
3. Improved network Elasticity
4. Flexible "pay-as-you-go" Service

 

[I] "Cisco visual networking index: Forecast and methodology, 2012-2017.", ciscosystems inc., May 2013

[II] "Cisco visual networking index: Forecast and methodology, 2012-2017.", ciscosystems inc., May 2013

[III] "1g/10g/40g/100g networking ports (2014 edition).", infonetics, limit l 2014

[IV] Ibid.