SDN controllers - (Open Networking Part 3-2)

Hello Friends, thank you for providing feedback on my earlier posts related to Open Networking. This post is a part of series of posts related to Open networking with a special focus on optical SDN architectures. As promised in my earlier post let's understand more about other SDN controller architecture available from other vendors in the industry.

In my previous post, we discussed the transport controller solution provided by Fujitsu, Infinera, and Sedona. This week let's review the controller solution available from Juniper, NetFlex LightRiver, and Cisco. 

Juniper's Northstar controller based on PCE architecture defined in RFC 5440, supports source packet routing in networks(SPRING) to manage and optimize network resources and data plane.

   

Netflix from LightRiver is a multi-vendor Optical domain SDN controller supporting the integration of devices provided by many industry-leading vendors like Adva, Ciena, Cisco, Fujitsu, Infinera, Nokia.

Cisco provides the commercial distribution of OpenDaylight platform providing solution most relevant use case line Device and Topology management, Traffic engineering, and Network Function virtualization. In the Reference section, I have provided a link to an interesting Open controller document providing details of various use cases and other domain controllers like WAE, Cable controller, and Cisco NFVI, worth read. 

Additionally, Cisco DEVNET provides a rich community for learning and sandbox environments as well for those trying to learn coding and testing apps. Again worth a try, it is FREE!

With this post, I would complete my research on various controller architectures available (or being worked upon) in the industry. Next week we will deep dive into one of the architectures and understand various use cases and components common to most of these architectures. 

One important aspect that I was not able to complete in this section was the adoption of these architectures to solve real-world use cases. If you know this information it will be of great value to me if you can leave it as comments.

" If you are not willing to risk the unusual, you will have to settle for the ordinary "

- Jim Rohn 

References:

Cisco Open SDN Controller

Juniper NorthStar controller

Netflex Optical SDN controller

My previous post- OPENROADM

My previous post - Open Optical Networking

Beyond 800G

Good morning, a detour this week from the original topic of this series Open networking( only for this week). 

I am amazed by the progress of the optical industry from 100G/wavelength deployed in the year 2011 to 400G/wavelength deployed in the year 2017 and now with 800G lambda/wavelength being deployed starting the year 2020 (Some of the recent 800G deployment and announcements available in the reference section).

So whats' NEXT?? 1.2T /wavelength(a.k.a lambda) 

There are two primary knobs that helped the industry to reach the 800G per wavelength capacity. First the use of coherent technology with higher modulation schemes and second increasing the baud rate(symbol/sec). However, physics does now allow us to play with these tools infinitely without paying a price :). 

With higher modulation formats while we gain higher spectral efficiency, the reach of such link is limited(compared to another design with the same baud rate and lower modulation format). Read this document for further details.

On the other hand with a higher baud rate, long-distance(long haul/submarine) can be realized but the overall spectral width(which is proportional to the baud rate) required to transmit the channel is higher, thus limiting the total number of channels that can be packed on a fiber. 

The industry is able to manage both these controls well and with improvement in FEC, DSP, mix modulation formats, and the real-world requirements, enabled combinations for the service providers to deal with the mind-boggling increase in bandwidth requirements. 

1.2T(1200 Gb/s) will not be much different. Acacia demonstrated its 1200Gb/s module in ECOC 2019 using 64QAM modulation with a ~100GHz spectrum requirement( reach not known). 

With 800G/wavelength in production now, it will be interesting to see the early leaders, developing and enabling network migration to 1.2T/wavelength.

Another interesting trend to watch: ill 1.2T/wavelength steal the thunder from 800G/wavelength (40GE vs 100GE)? 

Surprisingly I was not able to find many public announcements for the 1.2T/wavelength available in the public domain. If you know more please leave a comment and I will update the post. 

Have a good day 

Reference: 

Ciena 800G Announcements

Huawei and Turkcell 800G Trial

Infinera 800G Announcements

Acacia 1.2T single channel

Infinera Higher Baud rates

"Success is nothing more than a few simple disciplines, practiced every day"

- Jim Rohn

SDN controllers - (Open Networking Part 3-1)

Hello Friends, today, let's review different SDN controllers available in the industry and a high-level view of their design. I have tried my best to interpret the features published in the public domain and in case your interpretation is different please feel free to leave a comment and I will revise the post. 

Let's get started.....

As summarized in my previous post many service providers are transforming their Infrastructure to multi-vendor solutions. This migration enables the need to thoroughly validate hardware interoperation and a new requirement to build multi-vendor and multi-domain control and management plane software.

RFC8453 provides a framework for Abstraction and Control of TE  Networks (ACTN) to support virtual network services and connectivity services

Generalized Multi-vendor, Multi-domain architecture

There are many controller solutions available in the market and this week my attempt is to provide a brief overview of controller solutions from Fujitsu, Infinera, and Sedona. 

Next week I will post details for NetFlex LightRiver, Cisco, and Juniper solutions.

Again my attempt is to summarize and provide a starting point and you can make use of the reference link provided to explore further. 

I will be thankful if you can share(as comments) anything interesting you find during your research to augment my learning. 

References:

Good Read - ACTN

Fujitsu Virtuora NC

Infinera Transcend Controller Solution

Sedona NetFusion

"Start from wherever you are and with whatever you’ve got" - Jim Rohn

Open Networking Optical- Part2 - OpenROADM

Welcome back! this post is part of an Open Networking blog series,  exploring different Open Networking architectures and their current state(trials, demos, and deployment). You can refer to my previous post for an overview of all the major industry initiatives.

Today, let's take a closer look into OpenROADM progress made so far. 

OpenROADM is a major initiative within the optical industry primarily driven by AT&T( a service provider) to build scalable, cost-effective, open, and flexible networks.

Primary Use-Case for OpenROADM Trials

AT&T along with Vendor partners completed its first field test in December 2016, summarized below. 

For 400GbE trials, AT&T announced 3 phases for the testing. 

Phase 1: Will use optical gear from Coriant to carry a true 400GbE service across a long-distance span of AT&T global backbone from New York to Washington, demonstrating that AT&T’s nationwide software-centric network is 400G-ready.

Phase 2: Will trial a 400GbE on a single 400G wavelength across AT&T’s OpenROADM metro network. We’ll use optical gear from Ciena, a developer of next-generation coherent optical solutions, to show the network is ready to transport 400GbE to serve our customers in a metro area.

Phase 3: Will test the first instance of a 400GbE open router platform. The “disaggregated router” platform uses merchant silicon and open-source software – another industry first.

Finally, in November 2019 AT&T announced the completion of trials and enabled the first 400GbE production traffic on its network. Interestingly the traffic was terminated on another white box router. 

Definitely, a lot is accomplished between the OpenROADM MSA agreement and finally 400GbE production traffic using OpenROADM Interface definition. Definitely, the solution has transitioned from concept to production, and programmable, open optical networks should soon become a norm. 

Feel free to leave a comment, you can find below some important reference resources for additional information. See you soon, have a good day!

References:

AT&T Launches Industry-First 400-Gigabit

AT&T Phased Trials

AT&T Phase-1 Trial

400G Ciena-AT&T OpenROADM

AT&T ECOMP Architecture

" Formal education will make you a living; self-education will make you a fortune"

- Jim Rohn

Open Networking Optical - Part1

Major goals for Open Networking have been to take the best components, hardware or software or both and build a best-in-class network architecture and network. Hardware and Software disaggregation is playing a key role in this journey and in the year 2021 accelerated deployment of such architectures is on the horizon.

Many of the below initiatives impact Optical networking as well and this blog is one of many in this series that will follow. The Optical Networking Digital symposium was organized last week(Feb 16 & 18)  and some of the keynotes touched upon building programmable disaggregated networks.

       Major Open Networking project

There are overlaps and contradictions in some of these initiatives and in the next post, my attempt will be to summarize how much work has been done on some of these projects. You can find references to all these projects below and feel free to share your thoughts and comments. Have a good day!

Reference Links:

OpenROADM

TIP-OOPT

ODTN

ONOS

OpenDaylight

ONAP

OpenConfig

"Take care of your body. It's the only place you have to live  " - Jim Rohn