If you are a traditional optical/DWDM vendor, then be prepared for a shock!
Because some of you believe that Optical/DWDM box can be the most closed platform in networking.
After all, an optical signal is analog (not digital) therefore it is difficult to understand it.
And since vendors use proprietary enhancements to carry signals; therefore, it is very difficult to have an open interface to manage them.
So, customers are obliged to buy from you everything including the management system, the ROADMs (optical switch) and above all your cash cow- the transponders.
However, you might be preparing for surprises sooner or later.
As the industry has, already, started talking about the optical white box (Open optical switch configurable with open APIs).
Therefore, in this blog, I discuss why the optical business model that revolves essentially around selling optical hardware could be at risk because of SDN and in particular, projects like ONOS.
And in the new scenario of open platforms, why it is essential for the optical vendors to look for other business models.
But before that
What is ONOS?
ONOS stands for Open Network Operating System.
According to ONOS website:
The Open Network Operating System (ONOS) is a software-defined networking (SDN) OS for service providers that has scalability, high availability, high performance and abstractions to make it easy to create apps and services.
ONOS targets service providers. In that sense, it is different than the other well-known but earlier SDN controller Opendaylight. It is designed from day one with the carrier needs i.e. reliability, redundancy, and scalability. ONOS is still in its infancy but its development is supported by well-known operators and vendors.
Before moving on to discuss how ONOS targets the business model of optical vendors, it is essential to understand the relevant use case of ONOS that is directly related to Optics
ONOS Use Case: Packet and Optics integration
One of the main use cases in ONOS is the integration of IP/Packet and Optical domains. In today’s networks, these two layers evolve independently with no relation to one another even though both form one transport core. This leads to the build-up of both IP and optical domains since the transport links are over provisioned.
Additionally there are extra maintenance costs since two teams are needed to manage two separate networks.Not to forget the delays in service provisioning owing to the involvement of multiple teams.
This results in high CAPEX and OPEX spending on acquiring and maintaining two networks.
ONOS will address all these problems by combining both layers under one controller through standard southbound APIs.
The controller intelligently optimizes resources and protection on both layers bringing down the usage of both layers ( this will slow down the expansion requirements) and with one team and one interface for provisioning, the time to provision new services will be much faster.
Interesting, isn’t it?
So far, everything is good- So where is the issue?
Well, the direction of ONOS is great!
But, there is a requirement for achieving the objective of the use cases such as Packet and Optical integration.
And that is making optical switches open and programmable. Yes, the optical white box.
The concept will become clear from the following two slides from a presentation, which I found on the wikiONOS website. (Presentation name: Control of packet over optical networks)
The first slide talks about open optical hardware. Which is programmable and acts like a White box. Such a box in future will be configurable by any third party supporting relevant open APIs.
The second slide talks about the shift from traditional ROADM to an open ROADM with the ability for the ROADM to follow open commands like Match and action for ADD, DROP, and FORWARDING using OpenFlow.
When ROADM becomes an open platform, a controller such as ONOS would be able to program it irrespective of which vendor is providing the box. And this would facilitate the Packet and Optical integration.
Where does this leave the optical industry?
Well, the concept is disruptive yet achievable, if vendors collaborate.
let’s face it.
The concept is still under trial phase. And it will take long time to have something in the production network when standards are finalized.
But the work has already started.
Many traditional optical vendors may not like the white box concept.
But the truth is that there are already small companies emerging that offer some sort of open white box solutions for DWDM. Take a look at Calient, Polatis, Lumen. More of such companies would emerge in future.
The use case like this one leads us to question the future business model of the traditional DWDM vendors.
What is your Optical Business Model?
Is your business model strictly selling your own optical DWDM system with your own proprietary management system?
Are you open to the idea of open APIs for managing your box?
In my view, commoditization will hit the optical industry as it has any other networking business.
It might take longer to happen given the analog nature of light and the very closed and proprietary nature of the current optical gear.
Yet, the ball has already started rolling.
In my personal opinion, this might lead to two business models in future.
1. Open your box but play in the strict commodity market and compete on price.
2. Change the business model and reduce dependency on the hardware. Move to the applications that run on top of the hardware. Bring a software story to your hardware gear.
Some of the hardware players have already sensed the direction of the wind:
The hard-core hardware vendor Ciena’s acquisition of software player Cyan and hardware centered ADVA’s acquisition of overture signal that these companies want to move to a more software-centric business model.
Some are sticking to their guns, yet.
Some are still defining on what would be the next right strategy.
What is your opinion?
How do you feel about the optical white box?
Do you feel that it is a genuine need for operators to have a more open platform thus solving some pain points?
As a vendor, what are the obstacles you see in implementing a more open optical platform?
I would love to hear your opinion in the comments section below.