Should You Implement Optical Open Line Systems and Disaggregation?
Open line systems and disaggregation allow network operators to prevent vendor lock-in, upgrade technology cycles at a rapid pace, increase network capacity/speed and reduce CapEx and OpEx. This empowers operators to deploy the most cost-effective, highly adaptive and scalable network possible.
A long time ago, in an optical transport network far, far away…
Episode 299790: Optical Open Line Systems (OLS) and Disaggregation
In this article, we're going to break down what open line systems are, what makes them tick and why they have become such an attractive solution for network operators today. It looks as if optical transport systems have adopted some of the same strategies as data center applications, with their separation of storage and compute or in IT with the decoupling of applications, operating systems and hardware.
Pretty much anywhere you look, the entire networking landscape seems to end up with the same message: “disaggregation and openness.” So what are open line systems, and how does disaggregation play into this model? Well, before we can appreciate the open nature and disaggregation of line systems today, we have to get that old DeLorean up to 88 mph and set it back to the year 2005.
Traditional optical transport systems
In the early days of optical networking, line transport systems and other layers that leverage it for transport needs were, for the most part, a contained system from end to end, with their own native network management system. Please refer to Figure 1.
As we can see, there is an optical line system. This was typically a WDM system that housed filters, amplifiers and ROADM devices to add and drop wavelengths to and from their destinations through the optical line system footprint. This optical line system used its own vendor-specific network management system (NMS) to provision and monitor all services.
In most cases, these same optical line systems had the ability to house transponder and muxponder cards within a chassis solution, which could also leverage that same NMS. However, if the operator should choose a different vendor's transponder or muxponder, this would need to be managed by a separate NMS.
Moreover, as we start to stack in additional layers, each has its own native management system. This is shown in the OTN switching, alien wavelengths and routing layers. Even though we deliver the same services through multiple layers, there is a clear distinction between those devices and the systems used to manage them.
As expected, most standard ITU grid compliant line systems can accept alien wavelengths. Moreover, those ITU grid compliant systems, for the most part, can accept wavelengths from most any system as long as it meets the ITU and power specifications required. So in a sense, even the line systems of the old were somewhat open and disaggregated, meaning that if we were to use an ITU grid optic in a router, we could indeed inject that signal into an optical line system.
The same goes for OTN and alien wavelengths. However, the biggest difference between the past and today's transport systems is the network management system (NMS) and the ability to support open APIs for programmability across all platforms.
Open line systems of today
Open line systems (OLS) and disaggregation allow the operator to choose the best, most relevant solutions for each portion of the network they are addressing. As seen in Figure 2, regardless of the block of the network an operator is looking to address, we can leverage open APIs to have a single network management system (NMS) to provision and monitor the system and services.
This opens up the playing field of potential solutions and, more importantly, how the operator maintains those solutions. It is widely known that vendor technologies' cycles, direction, research and development, and survivability can be wildly different. If a network operator initially decides to go with a single vendor's solution and network management system (NMS) because, at that time, the direction and message are in-line with organizational goals, but later finds out that the future direction or research and development has shifted, then there is a potential risk of misalignment.
A single vendor can also be acquired by another entity or simply go belly up. In situations like this, it is never in the best interest of an organization or its shareholders to stick to a single vendor solution. Through open line systems and disaggregation, the operator now has the power to choose which solutions work best for them and can quickly shift when needed.
Packet and optical advancements
No true conversation around OLS and disaggregation is complete without mentioning the advancements being made across the industry regarding the convergence of packet and optical. Please refer to Figure 3.
Within recent years, there have been vast enhancements made in Layer 3 routing to allow for higher capacity WDM pluggable interfaces inserted directly into line cards and chassis of routers. These pluggable interfaces allow all of the magic done through traditional transponders and muxponders; those being advanced forward error correction (G.709) on a wavelength housed directly in the router and high-performance wavelengths across the ITU grid spectrum.
These optics are already reaching 400Gb of capacity and have advanced performance matching and in some cases are exceeding what can be done within transponders and muxponders, further consolidating the layers of the optical transport layer.
So why is this important? First and foremost, when blocks of the operators' network are being consolidated into other layers, there is a natural cost-saving in rack, space, power and cooling. Second, when integrating the functionality of a transponder or muxponder within the routing layer, we can drop a lot of the OpEx needed to manage and maintain that hardware.
In most cases, an operator is already leveraging a routing layer, and it is simply one less device or system to manage when the interface coming directly out of the Layer 3 device is injected directly into the OLS. Third, the survivability and failover of services through a routed platform are second to none. Layer 3 devices, in conjunction with MPLS/SR, provide real-time fast reroute protection through a feature called Topology Independent Loop-Free Alternate, or TI-LFA.
These protocols are vastly superior to truly Layer 1 solutions and offer substantial benefits to operators' growing infrastructure. Finally, with the consolidation of these layers, the network is simplified. By leveraging open APIs, an operator can seamlessly manage all of these services within the disaggregated open line system (OLS) construct. Recently, two of my co-workers (John Chiarini and Mike DiVincenzo) wrote a great article covering the convergence of packet and optical.
Below we can see a few use cases for how an operator can deploy an OLS. It is important to note that there are varying levels of how disaggregated an operator can choose to implement their OLS solution. OLS is not a “one size fits all,” and it depends on the operators' environment for how disaggregated they would like to be. As we step through each example's progression, the OLS continues to become more disaggregated until we reach example 5, where the entire system is fully disaggregated.
By looking at example 1, we can see that the operator chooses to use a single vendor solution for the entire OLS. It is still possible to deploy an open line system and utilize a single vendor. Some vendor management solutions are capable of maintaining and managing multiple vendors by allowing APIs and SDN orchestration.
In example 2, the operator chooses to utilize multiple vendors' transponders and muxponders in a bookended fashion, while utilizing a single vendor's mux/demux solution. With example 2, the operator can use a single NMS to provide, maintain and manage multiple vendors' transponders/muxponders and the single-vendor filters.
Example 3 is very much like example 2, except this operator chooses to utilize multiple vendors' mux/demux filters and line systems. Again, the NMS is capable of maintaining and managing the entire solution.
Example 4 is displaying an operator choosing separate transponders and muxponders on each end of the wavelength. This adds a layer of interoperability into the solution, and the operator needs to be sure that the transponders and muxponders functionality are tested.
In the final example 5 solution, the operator chooses to disaggregate just about every aspect of the solution. They are using separate vendors on each end of the transponder/muxponder wavelengths. Additionally, they are using separate mux/demux filters.
In example 5, the entire solution is capable of leveraging APIs for complete provisioning, monitoring and managing. However, the majority of the interoperability burden falls on the operators' shoulders. This solution is incredibly disaggregated, and the operator needs to be sure to test and vet each solution to fully ensure proper functionality.
When deploying an OLS, the operator has the flexibility to be selective in which vendor they choose. These could be for various reasons, but the underlying benefit of preventing vendor lock-in is a key strength. The operator now has the leverage to choose and drive down costs.
With a new speed and feed coming out, which seems like every quarter, the operator can leverage the latest and greatest innovation regardless of where it comes from. In some cases, an operator may be looking for the highest bandwidth possible within a transponder to take advantage of wavelength services truly. In other cases, they may be looking for the highest data rate within a Layer 3 solution that offers a converged packet and optical product. Regardless of the reason, the operator now has the ability to be picky when fulfilling their bandwidth needs.
Reach and capacity
With a disaggregated OLS, we can choose separate vendor solutions that increase the system's reach and capacity. Whether it is a new ultra-long-haul reach 400G high-performance optic or a new amplifier with specifications that exceed the past solutions, the operator now has the choice.
Through the use of open APIs, the operator can now get over the hump of managing and maintaining multiple network management systems. This enables the selection of the right products to match organizational goals.
High availability (HA)
Whether it is through a more disaggregated solution to allow for failover of multiple devices or utilizing the convergence of packet and optical to leverage MPLS or SR for service reassurance, the operator can instill HA in nearly all aspects of the network.
By consuming the most recent products, there will be inherent cost savings by lowering power, space and cooling costs. As with most new products, there are also benefits towards consolidating while adding new features. Those new features are typically speed or bandwidth. Thus, the operator can pack more data into a single communication channel. Ultimately by implementing new products, there are typically enhancements and cost reductions throughout the infrastructure.
Throughout most of this article, we are championing the benefits of a disaggregated OLS. However, these don’t come without sacrifice. With traditional line systems and all of the other portions of the network, the operator had a huge weight lifted off of their shoulders, and that involves reassurance. Vendors spend an incredible amount of research and development dollars assuring that their products will work within their contained system, if that means between the WDM line interfaces or the integration of transponders and muxponders within a chassis-based solution.
Regardless of how far a vendor's solution reached, they always controlled that environment and would assure functionality. With disaggregation and OLS, this somewhat goes out the window. A vendor can comply with standards and agreements of openness, but the testing and functionality burden will fall mostly on the operator to ensure the systems operate as expected.
For those who are used to doing things from a traditional transport system perspective, Disaggregated OLS may be a more complex task. In traditional systems, there is a simplistic method in which provisioning and monitoring happen. Again, OEMs spend a lot of R&D on being certain this works, and it is as easy to use as possible for their customers. There is a certain amount of open API “know-how” and experience that will be needed.
We all know that some of the benefits of an OLS are the operational savings from a space, power and cooling perspective. However, depending on the in-house ability to support such a solution in configuring APIs and SDN flows, OpEx may be a wash or even a loss.
It is important to understand that OLS and disaggregation is not a one size fits all. There are many varying sizes and techniques that can still benefit the operator while not over-complicating the solution or exceeding existing skill sets. Those benefits being pricing competition, growing bandwidth, reach and capacity, single NMS, high availability and lowing costs.
It is also important to note that adopting a disaggregated OLS doesn’t come without sacrifice. Typically network operators are accustomed to vendors taking on the responsibility of functionality between network elements. While adopting disaggregation and OLS, this ownership will most likely fall on the operator. Additionally, in some cases, network management complexity can increase, and OpEx may actually be a wash or loss depending on the existing skill set.
All in all, industry adoption of OLS and disaggregation is happening, and it is up to the operator to decide which pace they would like to adopt it.