Insights from Fyuz 2024
I am always pleased when I have the opportunity to travel to Dublin, Ireland. The city is both charming and vibrant, and I really enjoy a good Irish cider (I am not a Guinness drinker, sorry).
Two weeks ago, I had the pleasure of attending the Fyuz 2024 event in Dublin. This leading industry event, organized by Telecom Infra Project (TIP), brought together over 1,200 participants and 250 speakers from around the world. The theme for this year’s event was “Accelerate: Accelerate your deployments, business, and innovation,” which set the stage for three days of in-depth discussions and insights into the future of open and disaggregated networking technologies. Among the expert speakers was Harald Bock, Vice President, Network Architecture at Infinera, who participated in several Fyuz sessions.
I have mentioned in previous blogs TIP’s important role and relevance to the advancement of the open networking environment in general, and of optical networks in particular. At Fyuz, various aspects of open and disaggregated networking solutions were discussed, with a focus on the deployment, scaling, and utilization of the latest open networking technology to drive business growth and foster industry innovation.
Infinera’s Harald Bock, who joined the panel “Industrialization of open transport and IPoDWDM” on Fyuz’s main stage with other esteemed members of our industry, considers that while there is work to be done, open, disaggregated optical networks and IP over DWDM (IPoDWDM, where optical pluggables are directly installed into routers and connected over a DWDM optical line system, in multi-vendor deployments) are here to stay. These architectures can be used today, with the groundwork for scaling operations being established by standardization bodies and industry initiatives such as TIP. The primary challenge that remains is ensuring the maintenance of a functional ecosystem over time as networks evolve, including handling release upgrades and network extensions and deployments with the introduction of new vendors. On the other hand, although the increasing market interest in IPoDWDM is a significant driver for the adoption of open, disaggregated optical networks, both operators and vendors agree that IPoDWDM is not a universal solution. While many networking scenarios can benefit from the reduced space and power of pluggables equipped in router platforms, panelists agreed that transponder platforms continue to be relevant.
GX G42 Goes Silver!
Speaking of transponders, and of the growing maturity of open optical terminals…
At Fyuz 2024, TIP announced that the Infinera GX G42 platform received a “TIP Silver Badge” for complying with the requirements for open optical terminals defined by its Open Optical Packet Transport (OOPT) Mandatory Use Cases for SDN Transport (MUST) subgroup. The Infinera GX platform is our flagship multi-haul optical networking platform, of which the G42 is one of our latest compact modular chassis variants.
This award builds upon the previously received GX G42 TIP bronze badge and acknowledges the successful completion of tests validating the GX’s capability to integrate with open transport control and automation solutions through standard interfaces and common data models, specifically via the GX OpenConfig northbound interface (NBI). These tests were conducted earlier this year in the European laboratories of a major global Tier 1 operator. Infinera is honored to have received the first and only Silver Badge awarded by TIP to optical terminals to date. This recognition highlights that Infinera GX is a leading technology solution in the optical market and confirms our dedication to shaping the future of open networking.
AI, AI, and a Little More AI
Just like my colleague Jon commented on his MWC blog, artificial intelligence (AI) was also front and center at Fyuz. For us working in the telecommunication industry, AI is of interest from two complementary perspectives: networking for AI, and AI for networking.
Networking for AI
Discussions on networking for AI focused on how AI traffic might impact future network capacity needs. The million-dollar question is whether AI will replace recent capacity drivers like video streaming and cloud connectivity, sustaining current growth, or if it will significantly increase demand beyond the existing growth. This topic was also recently discussed at ECOC, with differing opinions presented. Notably, Meta stated that AI is already having a dramatic impact on backbone traffic, driving growth from 30% per year to 100% per year.
The impact on various network domains and the network architectures best suited to support AI are also important topics to be explored.
Although definitive answers to these questions are yet to be determined, operators and equipment vendors are closely monitoring market trends, adjusting their portfolios to ensure that their solutions align with future requirements, and developing business models to capitalize on the opportunities presented by AI advancements.
AI for Networking
Additionally, AI and machine learning (ML) are technologies that can be utilized to enhance both implementation and operational efficiency in open networking. At Fyuz, several applications of AI/ML to the optical domain were discussed in the OOPT project group breakout sessions. AI/ML engines equipped with context-aware models have the potential to greatly improve network performance, reliability, and scalability.
By continuously monitoring the network and leveraging the extensive data provided by modern network devices through streaming telemetry, AI/ML can produce actionable insights. These insights help determine where and when to add or remove capacity, recommend network upgrades and technology evolution, and predict degradation to prevent traffic impacts through preventive maintenance. AI/ML can also serve as valuable tools in our industry’s mission to reduce network power consumption. This can be achieved through power optimization strategies, carbon-aware planning, and the use of digital twins. A digital twin is a virtual representation of the network, continuously updated with real-time data to mirror its real-world counterpart. It can analyze performance, run simulations, and identify potential problems.
A generative AI-based operations assistant can help address the talent shortage in our industry. This assistant can guide network operations center personnel in their tasks, providing information on network status and assisting in completing complex processes by translating intent into actions. The operational assistant can orchestrate multiple (AI/ML) applications including new equipment onboarding, automated service planning and provisioning, and automated ticket resolution, unleashing the true power of automation.
Once again, Fyuz provided me with an excellent opportunity to meet, talk with, and learn from industry peers. I left Dublin very pleased with our Infinera GX Silver Badge and optimistic about the industry’s evolution, from open networking to AI/ML – and I raise my glass (of cider) to that!
