Thursday, July 11, 2024

OTN-Transport or OTN-Cross connect, What to choose in modern times?

 

Fortune favors the brave

1       Introduction:


1.1    Background of the OTN – XC:


In the early 2010s where in most of the areas of the world there was an explosion in the traffic requirement due to rise of data consumption there needed to be a technology that can deal with a combination of IP and traditional TDM at the same time. It was difficult and almost commercially inviable to have parallel transmission network for TDM and IP at the same time. Two parallel networks meant

1.     Higher CAPEX

2.     More maintenance.

3.     More usage of fiber cores or Channels.

4.     More manpower.

 

Arresting all these issues there was a standard quite similar to the levels of SDH multiplexing but with higher line rates. This technology was termed as OTN, Optical Transport Network. OTN has a similar hierarchal structure like that of the SDH but then there is no use of pointers over here rather it uses overheads. The understanding and the basic operations of these devices were more SDH like and thus were very appealing to the people who are especially in the transmission section of the network and are used to working with legacy, SDH like networks.

 

The OTN device has one more very unique characteristic and that is, it can aggregate several signals like IP, TDM, FC, HDI, SDI etc. and can line out this from a higher bandwidth line. This line can also be a coherent DWDM compliant interface that can easily be put as an optical channel in a WSS or in a MUX-DEMUX system. Thus, with the availability of one box the problems of multiple signal rates from different sources can be managed and multiple devices are not required for different technology. With the use of the OTN technology the transmission capabilities of a network are made more homogeneous and accommodative towards many technologies. Whatever the client rate maybe, it can be put in a OTN device and these can be aggregated and segregated at the same time.

 

There is also this good ability to add and drop multiple clients of different characteristics and switch directions of these clients by the introduction of OTN cross connects or OTN switches.

1.1.1   Structure of OTN:



 The building block and the progression from the client signal to the final Optical transport unit is somewhat similar to what is done in SDH. The difference over here is the pointers which is not existent in the OTN framework. However, if we inspect in details then the flexibility of OTN is to a great degree than that of SDH with respect to carrying of traffic. This is because of the flexibility in the number of interfaces that can be carried out with respect to client is much better than SDH.



1.1.2   The OTN Cross connect Device:

The OTN cross connect device or in short OTN-XC is a device that has the capability to have several clients converted to the OTN frame and then switch them to multiple directions. Just like the SDH cross connect this also depends on several unit of cross connections which are in the form of ODU (Optical Data Unit). This is why it is often referred to as ODU cross connect as well. In SDH the cross connects used to happen in the level of VC-12/VC-3/VC-4 and in the OTN cross connect they happen in the level of ODU0, ODU1, ODU2, ODU4 etc.

 

The ODU-XC consists of line cards that are usually coherent modules of 100G/200G/400G. These can be in the form of cards or pluggable CFP2/QSFPDD on the line cards. In addition it contains the client modules that are grey interfaces of STM1/4/16/64 1G/10G/100G FC etc. There is a back-plane connectivity and a ODU-matrix. This matrix is usually ODU-k (where k can be 0,1,2,3 and 4) or this can also be a ODU-Flex matrix where we can have several combinations of ODU in the system.


Structure of the ODU-XC device


The figure shows the flexibility of the OTN cross connect device in a similar fashion as it used to be for the SDH devices. There is a concept to cross connects that can be translated into trails in the level of topology and network. There can be a protected segment just like the SNCP-I that was there in SDH. Thus in many ways the OTN-XC resembles, operationally, the workings of the SDH cross connect albeit for higher line rates and for multiple client rates.


1.1.3   How the OTN-XC is operationally easy to use

 For a person who is coming from the traditional SDH/TDM background the OTN XC is definitely a  gradual and a comfortable method of evolution. The OTN-XC does not bring in drastic complications for the person who is into planning, provisioning and maintenance of the transmission network, as it is a gradual upgrade of the previous SDH network that was prevalent. Essentially the complications of several protocols and different service types are avoided in the transmission layer keeping it as simple as it can be.

 

On top of that the OTN-XC can do the same kind of provisioning which is closely aligned with the processes of the erstwhile TDM architecture. This is extremely important when a planning team and an operational team in the system is generally evolving from the SDH architecture. 


2       Challenges with the OTN – XC


2.1    Bulky:


2.2    Heavy consumption of Power:



2.3    Not efficient for point-to-point traffic


Generally, when we talk about wholesale bandwidth providers in the industry most of the nature of their traffic is point to point. The traffic is either between two data center points or specific leas of bandwidth, protected or unprotected across two different endpoints. Realizing an OTN-XC under such a condition is extremely expensive option.

Let us assume that there is a requirement of 4x100G across two points that are 500kms apart. The most efficient way would be to have a Muxponder with 4x100G groomed to a 400G QPSK line.


OTN on Card for Point to Point traffic


 

As we can see in the figure the Muxponder card provides a localized OTN – XC on a  card function and the traffic can be groomed in one unit and provide the aggregate traffic to the two different points.

The same system can also be used for 800G wavelengths and for 1.2Tb/s wavelengths as well. Here the main thing is that we are saving a lot of Opex when it comes to service delivery and cost by introducing the concept of Muxponder.


2.4    Not suitable for a full packet environment


As we have seen from 2017 there has been a consistent decline in the traditional TDM traffic that used to ride on SDH. Even international connectivity is now no more on SDH but on pure ethernet or IP peering. If we analyze the data from 2017 to 2023 then we will see that today almost 95% of the traffic that is there in the network is IP or Ethernet.

The utility of the OTN-XC was when we had a mixture of clients that were from the ethernet zone and the traditional SDH zone with some elements of fiber channel as well. However, with most of the traffic migrating to IP/Ethernet this mixture is becoming far from homogenous. The ethernet share of the traffic is increasing to a great extent.

This has further led to a thought process which mandates the coherent interfaces of 100G/200G/400G and even 800G to be used as a pluggable in the router itself which totally eliminates the need for the OTN-XC and to some extent even transponders and muxponders. The coherent interfaces can directly interact with the WSS or the Optical line system.

In the later section as we study the IPoDWDM system this is explained better.


3       OTN – Transport system



3.1    Understanding the OTN-TX components:

In order to delve more into the OTN – TX we need to first understand the components of the OTN-Transport. There are many elementary components that comprise of the OTN-Transport and this is what we are exploring in the sections below. 


3.1.1   Transponder:


The structure of a Transponder

As we can see in the picture this is a module where the input is essentially a signal of 100G / 400G and the output is a colored signal of OTU4/OTUC4.

The concept of transponder is slowly getting faded away with the arrival of IPoDWDM devices, which we will discuss later. However, transponder have a very good case where we need to connect third party grey interfaces to an optical line system of the same line rate. Eg. Say all the routers in the data center have 400G line out and this needs to be put on different wavelengths of 400G in the optical line system. Here the routers do not have colored interfaces and we need a device to convert the 400G to a proper coherent OTUC4 that can ride on the optical line system. This is where we take the help of a transponder.


 


3.1.2   Muxponder:







Structure of the ADM on a Card (AoC)

The picture shows an extension of the concept of Muxponder to a system that can be providing an east-west kind of a topology. The AoC concept is extremely beneficial to use if we know the proper drop points and plan the traffic in a way that it will properly adhere to the traffic matrix with a growth margin. The AoC in a way simulates the multi directional attribute of the OTN-XC in a very compact manner and gets accommodated in one of the slots of the photonic shelf. The AoC can seamlessly interoperate with a OTN-XC on a mesh side and a Muxponder on a terminal site and this can provide a comprehensive way of dealing with networks that have multiple layers. In the later sections it will be more apparent to understand how the traffic can be groomed across different kinds of the networks. 

3.1.4   Transponder/Muxponder:


Layout of a Transponder/Muxponder 


As we can see in the picture here, we have an example of a transponder/Muxponder. Another thing to note down is the 4x100GE realization on a single port. Here we use a concept that is called the fan-out. The fanout takes the advantage of grooming lanes of 100G on a 400G DR4/LR4 mapping. Generally, the QSFP slot of a 400GE takes the LR4/DR4 of 400G. This essentially means 4 lanes of 100G combining to make the 400Ge interface. Now this can be further distributed using a MPO cable and a passive fanout device to 4x100GE separate inputs. However, here the thing needs to be noted that the router should have the 100GE LR1 instead of the conventional 100GE LR4.


3.2    How the protection works?


It may be a common argument from the operator side that the realization of a simple level protection is difficult in the OTN-TX scenario as compared to a full-Fledged OTN cross connect device. There may be arguments that the OTN-XC has redundant mechanisms like the ASON in the electrical level and the SNCP when it comes to 1+1.  As these arguments are true, a point needs to be considered that with the inclusion of more an more electrical protection layers the cost increases drastically. This is because the most expensive component in a optical scenario is the Optical line transceiver. While providing multiple transceiver for the line is a solution to enhance protection but then it increases the cost to exorbitant levels. This is why alternative protection schemes should be envisaged looking at the MTBF of the line transceivers.

3.2.1   OLP Level Protection:


Below is an example of an OLP based configuration on one side of the network. The similar configuration is replicated on the other side of the network also. Please note that since this is using the collector and the ROADM network the OLP mechanism can also be implemented in a network that consists of a WSS Mesh and is not limited only to a network that is traditionally point to point.

 

With the introduction of OLP level protection we can provide


1.     1+1 protection for multiple number of client services with less cost.

2.     Optimize the number of line coherent interfaces.

3.     Reduce a lot of cost on the footprint by providing the same amount of redundancy on the line.

4.     Provide business services in a converged meshy network with desired amount of protection.


 

OLP level protection setup with dual collector



 


3.2.2  WSON (Wavelength Switched Optical Networks)


There would be a lot of demand that is based on multiple restoration if more paths are available in the network. A network that is complex and meshy demands multiple switching paths of the traffic with respect to the availability of paths after a failure event occurs on multiple sections. This also takes into account the availability of resources along the path. A classical way to achieving this in the OTN based network is ASON (Automatic switched optical network) that is based on the GMPLS protocol. A similar extension of this in the layer-0 is WSON. In the case of ASON it is the ODU that is the main unit of switching and there needs to be several electrical options and line interface availability for the switching of the network. This, however, is eliminated in the WSON network.

 

The WSON enabled network can do multiple level restoration and switching at layer-0 but switching wavelengths. This however needs at least a colorless-directionless configuration in the nodes or best a CDC configuration that makes the switching of wavelength easy for the traffic.

 

For networks that are metro level and connecting to several of data centers and traffic aggregation points over a region say spanning around 400-500 kms, WSON is a very good solution to provide cost effective services with multiple restoration. Here the requirement of line interfaces are less as there is no O-E-O switching and the cost of the infrastructure is one time.


Example of a WSON Network


 
In the picture above we see an example of a WSON network that is built out of the CDC ROADMs. There are multiple paths for the restoration of the main path and here three options of the alternate paths are shown. A Thing to note over here is that the WSON restoration paths do not need any additional line interface hardware to be constructed or generated. On the basis of availability, the system will create optical cross connects at the OCh level in the system. These OCH level cross connects will lead to the creation of new paths.

So to summarize when we do the network using OTN-Transport and WSON the following are the advantages of WSON

1.   

1.     1. Economical to implement in the network.

2.    2.  Does the restoration at the OCH level so there is no need for additional hardware.

3.    3. In the event more services are added extra line cards need not be added in the transit as the network is self-healing in the optical mode.

4.     4. GMPLS infrastructure has to be developed only once and not again and again. 


3.2.3   OLP + WSON


As we know that GMPLS restoration has its challenges or restoration times that can exceed the limit of 50ms. This is the reason why in ASON also we combine it with the protection of SNCP where there is a pre-provisioned protection path so that we have 50ms switching always.  In the case of WSON also we can combine this with OLP protection, and this can provide the 50ms switching always.

In the ROADM-CDC sites we need to ensure dual collector and ensure that the OLP connectivity is done properly along the lines. In case we use the CDC-collector, this will be better as we will have more flexibility in the wavelength re-usage in the whole network.

 

In the OLP+WSON mechanism we have complete end to end protection mechanism that is carrier grade in nature.


4       The trend to choose OTN- transport over OTN-XC


4.1    Comparison over cost and power



1.     The OTN-XC for a 400G lambda is 3 times as costly than the OTN – Transport.
2.     In case we are introducing ASON in the OTN-XC then the cost jumps 6 times.
3.     Power consumption is 4 times in the OTN-XC than the OTN-Transport.
4.     With ASON the power consumption is around 10 times more for the OTN-XC.

5.   As we increase the lambdas from 400G to 800G the cost even goes up.

 

CAPEX and Power consumption is extremely high as we go more towards the OTN-XC. Here this is the case because there will be expansion on the line side and the client side separately. On the other side there will be a good amount of increase on the line side when ASON is added. OTN – XC expansion will create more number of cards and line modules. In addition, it will increase the load on the matrix and the cross connect. These more modules and more number of processes increase the power consumption to a great deal.


4.2    Comparison on Space and Density:



Blending the OTN-Transport and the OTN-XC making an efficient solution


The figure above represents a comprehensive device for the OTN-XC and the OTN-Transport blending. For the present case of services which the customers have, especially in the region of wholesale bandwidth and providing bandwidth as a service, this model can be extremely efficient based on services.

5       IPoDWDM (IP over DWDM)


How the combination of transport and IP is in most of the cases today


The figure above was pertinent if there is a proper mix of the TDM clients and the IP clients. TDM aggregation coming from the legacy TDM mux while the IP clients coming from the IP interfaces. Both of them merging to the OTN-XC and then further the OTN-Wrapping is done to take this over the optical line system. Here we have a complete segregation of network and services. There is a layer of TDM, IP, OTN and Optical separately.

Today, however, the situation is different. The number of TDM-Mux in the network has reduced substantially and there is a good amount of increase in the IP-Traffic in the network. Most of the services are IP oriented so the relevance of the TDM device has reduced considerably. On top of that the bandwidth requirements have increased to a great extent. The need is thus of a router that can have line interfaces as coherent line.


How the IPoDWDM is setup

 

In the IPoDWDM structure we are able to groom the services of IP on a line that is coherent in nature. This line can be C-band or L-Band and thus can be integrated directly to the Optical Line system. Here, the OTN-XC and the OTN-Tranport both are eliminated from the network and the network becomes even more economical.

 

The carriage of the legacy bandwidth of TDM can be done over CES, Circuit emulated service. The CES can carry E-1, STM1/4/16 over encapsulation when needed and the IP services go as it is.

The line interfaces can be 100G/200G/400G coherent. The protection can also be based out of LDP-FRR and other IP protection mechanisms.


6      Conclusion



1.     OTN-Transport solutions involving transponders, muxponders, compact-modular serve to be much economical and efficient ways of delivering bandwidth especially in a wholesale bandwidth scenario. 


2.     OTN-Transport solutions have much higher densities, especially when we go to the options of compact modular to delivery high amounts of bandwidth with better efficiency. 


3.     The protection mechanisms and the reliability of the network is hardly compromised with options like OLP and WSON also in the offering. 


4.   The deployment ease and upgrade in an OTN transport network is much higher than the OTN-XC.


5.     IPoDWDM devices can be a better option considering the futuristic growth of the network. 


And Finally:

      So friends, fairly long article and the article is put up after a lot of research and observation. Hope all of you will like it. Please comment, share and like.... 


Cheers, 

Kalyan.....