home About Us Podcast Episode 29

The Critical Lowdown Podcast Episode 29

Part 2: Unleashing the Power of 400G: Revolutionizing Data Center Connectivity


Missed Part 1? Listen back now.

In Part 1, we delved into the transformative impact that 400G ZR+ hardware technology is having on data center networking, highlighting the evolution from traditional network architecture to the integration of advanced 400G ZR+ optics, eliminating the need for separate transponder layers. This shift, enabled by technological advancements in digital signal processors and the standardization of optical specifications, promises enhanced data transmission efficiency and interoperability across different vendors' equipment.

In this episode, we'll explore the software side of implementing 400G technology in data centers. Victor Khen from IP Infusion shares insights into the role of network operating systems in managing and tuning 400G transceivers, emphasizing the importance of a common management interface for seamless interoperability, along with the comprehensive testing and qualification process for these high-performance components, as well as the anticipated advancements in network management through AI and machine learning.


Stay tuned to learn how these developments are set to revolutionize data center operations and network innovation.

Subscribe to The Critical Lowdown from EPS Global wherever you get your podcasts:

Apple
Spotify
Google
ximalaya
John K Lynch - EPS Global

John Lynch
Director Sales, EMEA, EPS Global

Gert Sarlet - Coherent

Gert Sarlet
Director of Product Management, Coherent

Nanda Ravindran - Edgecore

Nanda Ravindran
VP of Product Management, Edgecore Networks

Victor Khen - IP Infusion

Victor Khen
Marketing Manager for Partners, IP Infusion

If you have any questions about or need advice or tech support for your upcoming project, don’t hesitate to get in touch. Or check out our 400G ZR+ Hardware Bundle here.


Transcript of Podcast

John: Victor, turning to the software aspect, IP Infusion is set to release a new version of OcNOS soon. Could you discuss how OcNOS will improve a customer's capability to orchestrate and automate their networks, particularly in the context of a 400G environment?

Victor: Yes, John. To clarify the extent of IP Infusion's involvement in the 400G domain, let's discuss the current state of our 400G solutions. We are a leading vendor with a robust offering for both service provider and data center networks, addressing both sectors comprehensively. Presently, we boast the broadest support for white box hardware platforms, especially in the ZR+ area, where we have five platforms already qualified. We anticipate adding another three or four platforms at the start of the next year, enhancing our ZR+ offerings to span a wide array of applications.

Currently, our solutions range from high-capacity access network routers to large-capacity multi-terabit provider edge routers, as well as 400G data center switches. We provide an extensive suite of solutions in collaboration with our partners.

To ensure seamless operation and manageability of these solutions, we've integrated a full suite of network synchronization tools. We offer OpenAPI for integration with orchestration and management tools, which is a standard for network orchestrators. Our partners also have access to a CLI for managing ZR+, and we are consistently integrating various YANG modules and NetConf for expedited network configuration, particularly with the release of OcNOS 6.4. This version introduces new network resilience and load balancing tools that are configurable via NetConf.

Zero-touch provisioning has been a feature of OcNOS from its inception. With the latest 6.4 update, we've also implemented telemetry across both data center and service provider routers and switches, making telemetry available across all our products. Additionally, we've introduced a range of other features that enhance the efficiency and operability of our solutions.

 

John: Thank you, Victor. Let's discuss 400G technology. It's significant across various industry sectors, including data centers, interconnects, metro networks, high-performance computing clusters, and for service providers updating networks to meet today's high-bandwidth demands, such as video streaming. Can you share examples of real-world deployments where your products are utilized? If possible, mention a customer; if not, a general discussion on the technology would be valuable. Gert, could you provide insights into Coherent's deployments of 400G?

Gert: Certainly. Windstream is a customer we've collaborated with extensively, and they've been very open about our partnership through press releases and webinars. I can discuss their network specifically. Windstream was among the first to recognize the cost savings from removing the transponder-muxponder layer and integrating Coherent optics directly into their IP layer, including switches and routers. This not only reduces capital expenditure by decreasing the amount of equipment needed but also offers significant operational savings, particularly in power consumption. With global electricity costs rising in recent years, this is a substantial benefit.

Additionally, Windstream leases space in colocation facilities, so minimizing the equipment's physical footprint translates to considerable savings on their monthly lease expenses. Another less obvious benefit is improved network reliability. With fewer devices in the network, the likelihood of failure decreases, leading to lower maintenance costs. The total cost of ownership savings exceed 40%.

Initially, Windstream implemented 400G technology in their metro network. Now, they are adopting our new 100G transceiver, which brings the advantages of the 400G solutions to a smaller form factor, the QSFP28 at 100G. This allows them to extend these benefits to the network's edge, reinforcing the value they find in this approach.

John: Great. I think Victor and Nanda would be delighted to hear that I'll be asking them to call for your 100G optic once it becomes available. I can ask them here in public, which makes it harder to say no. Nanda, do you have an example of how this 400G technology is being deployed with Edgecore?

Nanda: Certainly, John. While I can't disclose the customer's name, I can mention that they're located in North America and are involved in a large data center deployment. They required support for over 400G and were looking for a unique architectural advantage. Gert highlighted several benefits, such as operational efficiency and power savings. In this instance, the customer also gained an architectural benefit from a single platform that could handle both the data center interconnect and the spine function within the data center.

For this customer, we qualified 400G ZR+ transceivers. Our platforms are open, as previously stated, and we successfully qualified not just one but two different 400G transceivers. The customer is quite satisfied with the deployment in their network. They are realizing the benefits we discussed, including operational and power efficiencies provided by these transceivers and solutions. Additionally, they appreciated the architectural advantage of seamlessly integrating a data center interconnect with the spine network, which proved to be extremely beneficial for them.

John: Super, thanks. Victor, are you observing any significant activity in the market regarding 400G technology?

Victor: Yes, John, there is significant activity around the 400G ZR+, particularly this year. We've seen a notable reference story with the London Internet Exchange (LINX), where they upgraded their infrastructure to our spine using OcNOS. The upgrade began with LINX and later extended to Manchester's Internet Exchange. As a result, OcNOS now manages the network for the data center segment, connecting approximately 980 networks across 80 countries.

What's particularly interesting is that, although we often discuss Greenfield deployments, LINX's project was a brownfield deployment. They started by upgrading their IP network and subsequently realized that with OcNOS management, the network became very streamlined, facilitating the addition of new components.

Currently, LINX operates using traditional transponders, but the new switches are enabling them to adopt ZR+. They have already commenced migration to a ZR+ infrastructure, which offers the advantage of running in parallel with the legacy transponders. This allows for a gradual transition to ZR+, presenting a significant benefit to their operations.

John: Excellent, thank you for that. Thanks for sharing the information.Gert, you've touched on some of this in your previous answers, but sustainability is a major theme now in product design, regardless of the sector. Could you discuss the role sustainability plays in Coherent's design process for new products, especially considering the significant power requirements of some devices?

Gert: It's all relative when discussing power requirements. Line side optics do consume more power than typical client side, shorter reach optics. However, when compared to legacy alternatives for line side, there's been a significant reduction in power usage. Our goal is to accommodate smaller form factors and ensure that edge core switches can handle these transceivers, which is why minimizing power consumption is a key focus for us. From a broader perspective, the greatest power savings in terms of network sustainability come from the ability to eliminate large amounts of equipment from the network. This is where we see the most substantial impact on sustainability.

John: Thank you, Gurt. Now, Victor, from a software perspective—and I apologize for the slight sales pitch on behalf of EPS—does our approach as a software company offer an advantage? Specifically, when EPS provides the market with an operating system bundled, pre-configured, pre-tested, and supported in a manner akin to traditional networking product vendors, do you perceive a benefit for your company in the market?

Victor: John, it's a pertinent question, and I must commend your team for their efforts. To realize the vision of Open Networking, merely having superior technology is insufficient. We recognize the necessity for solution partners who can amalgamate diverse solutions and present a unified offering to end-users. The value in this approach is undeniable.

EPS Global has been our reliable partner for years, with a global presence and remarkable flexibility in providing Open Networking solutions. While some end-users opt to purchase individual network components and assemble them on their own, the majority prefer pre-integrated solutions, which aligns with the traditional market operations.

The true value of EPS lies in its adaptability to cater to all types of customers, positioning it as one of the most mature, if not the most mature, value-added distributors for Open Networking solutions. This is a significant factor in making the procurement and servicing of Open Networking as straightforward as proprietary solutions, which is paramount for our customers and for us.

Furthermore, EPS offers industry-level pre-sales and post-sales support services, which adds immense value to the services they provide.

John: Thank you, Victor, for the insightful contribution. As we approach the end of our discussion, let's consider the future. AI and machine learning are hot topics. Are there applications for these technologies in network management? Could AI or machine learning influence the areas we've discussed today? Nanda, from Edgecore's standpoint, where do you see AI fitting into the future of network management?

Nanda: Certainly, John. Edgecore has been heavily involved in AI technology, particularly in the development of switches for AI and ML use cases. We've focused on software integration, tail latency, and conducting POCs with customers to measure job completion times. Our goal is to establish Ethernet as the go-to networking technology for AI infrastructure and related applications.

Regarding the role of AI in network management systems, we've observed a trend towards increasingly intelligent systems. Network management systems (NMS) are evolving to autonomously handle events, such as responding to alarms without operator intervention. This progression suggests that NMS will become even more sophisticated, incorporating algorithms that learn network behavior and operator management strategies. Eventually, these systems will begin to autonomously manage and troubleshoot the network.

The integration of AI into network operations is not a distant reality; it's an imminent change. As AI continues to learn and adapt, it will increasingly take over network operations. This shift is just a matter of time.

Given the breadth of AI as a topic, it certainly warrants its own dedicated podcast discussion.

John: Yes, I agree. I'm aware that all of you have upcoming products that are pertinent to the AI world. In a few months, we'll reconvene to have another podcast focused specifically on AI, as per your suggestion, Nanda.

I want to express my sincere gratitude for your time today. Having all of you together to discuss the interplay of your products has been incredibly valuable. Our usual podcasts feature individual vendors discussing their products in isolation. However, this joint discussion has been particularly enlightening, not only highlighting each product's standalone capabilities but also their synergy.

I also want to acknowledge the effort you've all put into the qualification process. Your investment in ensuring these parts work cohesively allows us, as facilitators, to confidently present these bundles to our customers. The pre-qualification gives us a significant advantage in marketing and selling these combined solutions, and for that, I am truly thankful.

A special thanks to Coherent for joining our podcast series for the first time. Gert, your participation is greatly appreciated.

Gert: It's been an honor. I value the collaboration with EPS, IP Infusion, and Edgecore. As you mentioned, offering an integrated solution to network operators presents tremendous value.

Thank you. Finally, I would like to remind our listeners that if you have any questions stemming from our discussion today, we encourage you to contact us. Our sales teams are available worldwide, so no matter where you are, you should be able to locate an EPS salesperson in your vicinity. All the necessary contact information is available on our website at www.epsglobal.com. We look forward to assisting you. Thank you very much.


Glossary of Terms

  • 400G ZR+ Optics: Advanced optical transceivers designed for high-speed data transmission over optical fiber, supporting data rates of 400 gigabits per second with enhanced reach and performance features beyond the standard 400ZR specifications.
  • Coherent Optical Transceivers: A type of optical transceiver that uses coherent detection and digital signal processing (DSP) techniques to modulate and demodulate light signals, enabling higher data transmission rates and longer distances compared to traditional direct-detection methods.
  • QSFP-DD Transceivers: Quad Small Form Factor Pluggable Double Density transceivers. A high-density, high-speed pluggable module used in data communications for 400G and beyond, featuring an eight-lane electrical interface.
  • Digital Signal Processing (DSP): A method of processing signals digitally to improve signal quality, bandwidth efficiency, and error correction

 

Need Help?

We have local language and currency support in each of our 28 locations, ensuring you always have access to friendly customer support to deliver your hardware solutions regardless of your location.