标签归档:400G Technology

400G Optics in Hyperscale Data Centers

Since their advent, data centers have been striving hard to address the rising bandwidth requirements. A look at the stats reveals that 3.04 Exabytes of data is being generated on a daily basis. Whenever a hyperscale data center is taken into consideration, the bandwidth requirements are massive as the relevant applications require a preemptive approach due to their scalable nature. As the introduction of 400G data centers has taken the data transfer speed to a whole new level, it has brought significant convenience in addressing various areas of concern. In this article, we will dig a little deeper and try to answer the following questions:

  • What are the driving factors of 400G development?
  • What are the reasons behind the use of 400G optics in hyperscale data centers?
  • What are the trends in 400G devices in large-scale data centers?

What Are the Driving Factors For 400G Development?

The driving factors for 400G development are segregated into video streaming services and video conferencing services. These services require pretty high data transfer speeds in order to function smoothly across the globe.

Video Streaming Services

Video streaming services were already taking a toll on the bandwidth requirements. That, combined with the COVID-19 pandemic, forced a large population to stay and work from home. This automatically increased the usage of video streaming platforms. A look at the stats reveals that a medium-quality stream on Netflix consumes 0.8 GB per hour. See that in relation to over 209 million subscribers. As the traveling costs came down, the savings went to improved quality streams on Netflix like HD and 4K. What stood at 0.8 GB per hour rose to 3 and 7 GB per hour. This evolved the need for 400G development.

Video Conferencing Services

As COVID-19 made working from home the new norm, video conferencing services also saw a major boost. Till 2021, 20.56 million people have been reported to be working from home in the US alone. As video conferencing took center stage, Zoom, which consumes 500 MB per hour, saw a huge increase in its user base. This also puts great pressure on the data transfer needs.

What Makes 400G Optics the Ideal Choice For Hyperscale Data Centers?

Significant Decrease in Energy and Carbon Footprint

To put it simply, 400G raises the data transfer speed four times. 400G reduces the cost of 100G ports as breakouts when comparing a 4 x 100G solution to facilitate 400GbE with a single 400G solution to do the same. A single node at the output minimizes the risk of failures as well as lower the energy requirement. This brings down the ESG footprint that has become a KPI for the organizations going forward.

Reduced Operational Cost

As mentioned earlier, a 400G solution requires a single 400G port, whereas addressing the same requirement via a 100G solution requires four 100G ports. On a router, four ports cost way more than a single port that can facilitate rapid data transfer. The same is the case with power. Combined together, these two bring the operational cost down to a considerable extent.400G Optics

Trends of 400G Optics in Large-Scale Data Centers—Quick Adoption

The introduction of 400G solution in large-scale data centers has reshaped the entire sector. This is due to a humongous increase in the data transfer speeds. According to research, 400G is expected to replace 100G and 200G deployments way faster than its predecessors. Since its introduction, more and more vendors are upgrading to network devices that support 400G. The following image truly depicts the technology adoption rate.Trends of 400G Optics

Challenges Ahead

Lack of Advancement in the 400G Optical Transceivers sector

Although the shift towards such network devices is rapid, there are a number of implementation challenges. This is because it is not only the devices that need to be upgraded but also the infrastructure. Vendors are trying to upgrade them in order to stay ahead of the curve but the cost of the development and maturity of optical transceivers is not at the expected benchmark. The same is the case with their cost and reliability. As optical transceivers are a critical element, this comes as a major challenge in the deployment of 400G solutions.

Latency Measurement

In addition, the introduction of this solution has also made network testing and monitoring more important than ever. Latency measurement has always been a key indicator when evaluating performance. Data throughput combined with jitter and frame loss also comes as a major concern in this regard.

Investment in Network Layers

Lastly, the creation of a plug-and-play environment for this solution also needs to be more realistic. This will require a greater investment in the physical, higher level, and network-IP components layers.

Conclusion

Rapid technological advancements have led to concepts like the Internet of Things. These implementations require greater data transfer speeds. That, combined with the world going to remote work, has exponentially increased the traffic. Hyperscale data centers were already feeling the pressure and the introduction of 400G data centers is a step in the right direction. It is a preemptive approach to address the growing global population and the increasing number of internet users.

Article Source: 400G Optics in Hyperscale Data Centers

Related Articles:

How Many 400G Transceiver Types Are in the Market?

Global Optical Transceiver Market: Striding to High-Speed 400G Transceivers

NRZ vs. PAM4 Modulation Techniques

The leading trends such as cloud computing and big data drive the exponential traffic growth and the rise of 400G Ethernet. Data center networks are facing a larger bandwidth demand, and innovative technologies are required for infrastructure to meet shifting demands. Currently, there are two different signal modulation techniques examined for next-generation Ethernet: non-return to zero (NRZ), and pulse-amplitude modulation 4-level (PAM4). This article will take you through these two modulation techniques and compare them to find the optimal choice for 400G Ethernet.

NRZ and PAM4 Basics

NRZ is a modulation technique using two signal levels to represent the 1/0 information of a digital logic signal. Logic 0 is a negative voltage, and Logic 1 is a positive voltage. One bit of logic information can be transmitted or received within each clock period. The baud rate, or the speed at which a symbol can change, equals the bit rate for NRZ signals.

NRZ
NRZ

PAM4 is a technology that uses four different signal levels for signal transmission and each symbol period represents 2 bits of logic information (0, 1, 2, 3). To achieve that, the waveform has 4 different levels, carrying 2 bits: 00, 01, 10 or 11, as shown below. With two bits per symbol, the baud rate is half the bit rate.

PAM4
PAM4

Comparison of NRZ vs. PAM4

Bit Rate

A transmission with NRZ mechanism will have the same baud rate and bitrate because one symbol can carry one bit. 28Gbps (gigabit per second) bitrate is equivalent to 28GBdps (gigabaud per second) baud rate. While, because PAM4 carries 2 bits per symbol, 56Gbps PAM4 will have a line transmission at 28GBdps. Therefore, PAM4 doubles the bit rate for a given baud rate over NRZ, bringing higher efficiency for high-speed optical transmission such as 400G. To be more specific, a 400 Gbps Ethernet interface can be realized with eight lanes at 50Gbps or four lanes at 100Gbps using PAM4 modulation.

Signal Loss

PAM4 allows twice as much information to be transmitted per symbol cycle as NRZ. Therefore, at the same bitrate, PAM4 only has half the baud rate, also called symbol rate, of the NRZ signal, so the signal loss caused by the transmission channel in PAM4 signaling is greatly reduced. This key advantage of PAM4 allows the use of existing channels and interconnects at higher bit rates without doubling the baud rate and increasing the channel loss.

Signal-to-noise Ratio (SNR) and Bit Error Rate (BER)

According to the following figure, the eye height for PAM4 is 1/3 of that for NRZ, causing the PAM4 to increase SNR (Signal-Noise Ratio) by -9.54 dB (Link Budget Penalty), which impacts the signal quality and introduces additional constraints in high-speed signaling. The 33% smaller vertical eye opening makes PAM4 signaling more sensitive to noise, resulting in a higher bit error rate. However, PAM4 was made possible because of forward-error correction (FEC) that can help link system to achieve the desired BER.

NRZ vs. PAM4
NRZ vs. PAM4

Power Consumption

Reducing BER in a PAM4 channel requires equalization at the Rx end and pre-compensation at the Tx end, which both consume extra power than the NRZ link for a given clock rate. This means PAM4 transceivers generate more heat at each end of the link. However, the new state-of-the-art silicon photonics (SiPh) platform can effectively reduce energy consumption and can be used in 400G transceivers. For example, FS silicon photonics 400G transceiver combines SiPh chips and PAM4 signaling, making it a cost-effective and lower power consumption solution for 400G data center.

Shift from NRZ to PAM4 for 400G Ethernet

With massive data transmitted across the globe, many organizations pose their quest for migration towards 400G. Initially, 16× 25G baud rate NRZ is used for 400G Ethernet, such as 400G-SR16, but the link loss and size of the scheme can not meet the demands of 400G Ethernet. Whereas PAM4 enables higher bit rates at half the baud rate, designers can continue to use existing channels at potential 400G Ethernet data rates. As a result, PAM4 has overtaken NRZ as the preferred modulation method for electrical or optical signal transmission in 400G optical modules.

Article Source: NRZ vs. PAM4 Modulation Techniques

Related Articles:
400G Data Center Deployment Challenges and Solutions
400G ZR vs. Open ROADM vs. ZR+
400G Multimode Fiber: 400G SR4.2 vs 400G SR8