The internet superhighway is accelerating! The deployment of 5G networks, artificial intelligence (AI) applications and Internet-of-Things (IoT) technology has increased the demand for higher bandwidth and data rates. As a result, there is increased anticipation and interest in 800G transceivers and networks. Conversations about the next speeds in the internet superhighway, 1.6T networks, aren’t just conjecture.
For those networking on the edge, 800G speed is beginning to replace 400G as the expected optimal data transmission rate. With 800G networks in place, demand for 800G transceivers, with their high data rate and capacity, will continue to increase.
Features of 800G Transceivers
Also called “dual 400G transceivers”, 800G transceivers and networks combine two different 400G light-paths into one optical channel of a single wavelength. Some features of 800G transceivers include:
- Use of PAM4 (Pulse Amplitude Modulation with four levels) to achieve maximum data rate over the same bandwidth.
- Include state-of-the-art forward error correction (FEC) algorithms to ensure reliable long-distance transmissions.
- Transfer data at high-capacity data rates of 800G per wavelength.
- Provide the size, power consumption and cost benefits of a pluggable (OSFP).
- Feature photonic integrated circuits (PIC) which use low power up to 20W in the 800G links.
- Support 16 QAM modulation scheme with 120 GBaud modulation rate.
GBaud stands for Giga Baud, wherein “Baud” is a common unit of measurement in telecommunications and electronics of symbol rate, which is one of the components that determine the speed of communication over a data channel.
- Use a fault-tolerant modular system that involves a combination of advanced coherent DSP, advanced PIC, efficient analog electronics, and robust packaging technology.
Applications of 800G Transceivers
- Cloud Computing: Hyper-scale data centers with cloud computing applications continuously transfer and store inordinate amounts of data streaming from global networks.
- Edge Data Centers: Due to reducing latency to support numerous applications as described, both wireline and wireless operators are moving data centers to the edge as an enterprise or macro cell site.
- Low Latency Industries and Applications: Industries and applications such as banking, diagnostic imaging, defense, navigation, stock trading, weather forecasting, collaboration, research, ticket sales, video broadcasting and online multi-player gaming that depend on almost “instantaneous” large data transfers with minimum latency.
- 5G Telecom: Massive data flow and continuous mobile network rollout of 5G network are extending data center capabilities.
- Virtual Reality (VR) and Augmented Reality (AR): Virtual machines and servers require constant, bidirectional data bandwidth with very low latency.
- Artificial Intelligence (AI) and Machine Learning (ML): AI applications require very high-speed data rates and bandwidth to quickly process large data sets and execute complex algorithms. This can include anything from autonomous vehicles, industrial automation, and robotic surgery to IT edge security.
- Video Streaming: High resolution 4K and 8K video streaming require high-bandwidth networks that can support ultra-high definition (UHD) displays. Applications run from sporting events to interactive, real-time online gaming.
- Work from Home (WFH): A slow trend before the COVID-19 pandemic, the WFH percentage is now significant and is here to stay. As more business and institutional professionals work remotely, the demand for speed and bandwidth will expand to include residential areas.
How Optics Evolved from 25G to 800G
The evolution of 25G to 800G, SFP to the QSFP double density 800 MSA, has taken place over many years with a combination and the advance of three key technologies:
- Increased Baud Rate: Expanding the amount of data that a single channel can carry from source to destination significantly improves the capacity of each channel. Result: more data is transmitted with reduced latency.
- Modulation Switch: 800G uses PAM4 modulation to enhance network performance and shift towards higher data rates. That means twice as much data can be transmitted per signal compared to traditional NRZ (Non-Return to Zero) modulation used in low-speed transceivers.
- Increased Lanes: Data rates are increased by either using parallel channels or by increasing the number of fibers in the cable.
400G to 800G Adoption: 2023 is the Year of Change
While 2025 is a predicted target for most 800G applications, hyper-scale data centers, industries and service providers that are using 400G links in their core networks, are already moving toward 800G deployments. Key to the transition from 400G to 800G optics is the agreement and adoption of specifications for two key 800G transceivers: OSFP and QSFP-DD.
- Octal Small Form-factor Pluggable (OSFP): Each lane of eight high-speed lanes is capable of transferring data at the rate of 100 Gbps per lane resulting in a total bandwidth of 800 Gbps. It is comparatively wider and deeper than the QSFP and meets the power requirements of 800G optics. 800G OSFP modules are used in high-speed optical communication networks.
- Quad Small Form Factor Pluggable Double Density (QSFP-DD): The 800G MSA will maintain the current design of 400G QSFP-DD optical modules as well as create a next- generation QSFP-DD optical module featuring 8 channels and a single-link capacity of 100 Gbps. The power consumption of the QSFP-DD 800 optical module is expected to be up to 24W with its cage and connectors to be compatible with existing QSFP-DD and QSFP modules.
Challenges When Implementing 800G
We all know that deploying new technology isn’t plug and play right away. (Remember the VCR to DVD leap?) For 800G fiber networks, compatibility, standards, application issues, integration with legacy equipment devices and high cost are all considerations for early adopters.
Some of the most significant implementation and performance challenges associated with 800G transceivers and fiber networks include:
- Current implementations of 800G use 8x lanes at 100Gbps per lane with double the PAM4 speeds from 50Gbps (previous generations) to 100Gbps. Already in development are 200Gbps per lane 800G transceivers…a huge challenge due to parallel development of higher order modulation and PAM4 data rate.
- The only available 800G standard is the Ethernet Technology Consortium’s (ETC) 800GBASE-R.
- Upgrading networks and components from 400G to 800G means doubling the spectrum, sampling speed and symbol rate. Minor 400G issues could become major problems that affect electrical performance at 800G.
- Early devices may not support both Auto-Negotiation (AN) and Link Training (LT) for performing electrical signal transmission. This compatibility issue between the ASICs could increase the risks of link flaps.
Auto-Negotiation enables a port on a switch, server, or router to communicate with the destination device to determine the connection speed and optimal duplex mode.
Link training refers to the process where the transmitter and receiver communicate with each other to synchronize their equalization settings on high-speed serial link.
ASIC stands for Application-Specific Integrated Circuit which contains multiple functions onto a single chip. It is much more reliable and cost-effective than discrete components.
- 800G optical transceivers dissipate large amounts of heat which can impact the performance of the components if not cooled properly, and increase utility costs.
- The number of high-speed optics on the market is very limited (for now).
- 800G testing can be costly.
Is Your Network Ready for 800G Transceivers?
Data-centered enterprises are pushing network providers and applications to provide them with higher bandwidth, reduced latency with reliable and seamless connectivity. 800G transceivers are the next logical deployment choice to provide that fast, secure, and dependable connectivity. However, network engineers, data center architects and other future planners must weigh the demand for a faster, bigger pipe against deployment challenges and migration planning of 800G. Providers with end users with the highest use and demand will be the first to implement and push standardization and product availability.
For anyone interested in 800G transceivers as part of a standalone or integrated with existing 200G and 400G networks and devices, talk to a Vitex fiber optical engineer. Or, watch our “Upgrade your network infrastructure with the new 800G QSFP-DD optical transceiver” episode in our Vitex Talks for Fiber Optic Pros series, or browse our 800G transceiver products. Vitex fiber application engineers are available to help with optical transceivers, cables, AOCs, and more! For all industries, we’ll work with you and your team to optimize a solution, develop the right product and resolve hardware challenges affecting your data network goals.