Intelligent Interconnects
TeraSignal Introduces TSLink: Protocol-Agnostic Intelligent Interconnect for Plug-and-Play Linear Optics in AI Infrastructure
Innovative TSLink Architecture Enables Mass Adoption of Intelligent 800G LPO Modules with Automatic Link Training and Link Monitoring
Our Intelligent Re-Driver (IRD) Technology
TeraSignal’s Intelligent Re-Driver (IRD) technology combines advanced electronics with photonics to create a new generation of intelligent linear pluggable optical (LPO) modules. Unlike traditional LPO modules, which rely on DSP-based retimers, TSLink’s IRD technology provides automatic adaptation, digital link monitoring, and link training, overcoming the limitations of traditional solutions and simplifying integration and deployment in AI infrastructure.
50% Power Reduction
TeraSignal’s TS8401/02 modulator drivers cut power consumption by at least 50% compared to non-CMOS drivers. By leveraging existing microcontroller resources, TSLink eliminates the need for additional DSPs, making it an energy-efficient and sustainable choice for large-scale data centers, reducing operating costs and environmental impact.
Near-Zero Latency
By removing DSPs and using a continuous-time signal path, TeraSignal’s re-drivers significantly reduce latency. TSLink’s streamlined architecture eliminates the delays caused by DSP processing, which is essential for high-performance AI applications, ensuring faster data transmission and real-time processing in data centers.
Real-Time Link Diagnostics
TSLink integrates digital eye monitoring and link training to provide real-time insights and automatic configuration. With its advanced monitoring capabilities, TSLink enhances signal integrity, ensures optimal link performance, and allows for adaptive optimization, identifying and addressing link issues in real time, simplifying deployment in complex environments.
Intelligent 800G Connectivity
The new TS8401/02 intelligent 400G (4x100G) PAM-4 modulator driver is the industry’s first CMOS solution with link diagnostics for 800G linear pluggable optical (LPO) modules. This intelligent re-driver innovation is poised to set new optical interconnect standards in power efficiency, low latency and reliability for self-optimizing intelligent components for AI deployments in data centers.
Automatic Link Training
TSLink automates link training and real-time performance monitoring, eliminating the need for additional DSPs. It uses CMIS to automatically optimize the connection between the host and module, while advanced digital eye monitoring and BER reporting ensure continuous, peak link performance.
DSP-Free Efficiency
TSLink leverages existing microcontroller resources to automate link training and monitoring, eliminating the need for DSPs. This advanced solution reduces power consumption and latency, optimizing data transmission and ensuring reliable link integrity for AI and high-performance computing environments.
Applications
By leveraging TeraSignal’s Intelligent Re-Driver technology, these application areas benefit from reduced power consumption, minimized latency, enhanced reliability, and a quicker time to market. Our technology redefines the infrastructure capabilities necessary for next-gen AI and compute environments.
AI and Accelerated Computing
TeraSignal enhances PCIe and CXL connectivity for GPUs (JBoG), accelerators, and FPGAs, maximizing link uptime. This ensures reliable, high-speed connections essential for advanced compute tasks in AI, providing the backbone for innovation.
Server Infrastructure
Our technology ensures high-bandwidth, low-latency PCIe connectivity for compute, storage, networking, and accelerator resources, crucial for server efficiency and data processing capabilities.
Expanding Memory Solutions
We deliver memory connectivity that not only expands bandwidth but also increases capacity, meeting the demands of modern data-driven applications and facilitating faster access to large datasets.
Next-Gen Networking
Active Ethernet (100/50/25G per Lane) along with PCIe/CXL offers both intra-rack and inter-rack connectivity, bridging the gap between devices with unprecedented speed and reliability.
Storage Connectivity
Our solutions support massive data storage demands, particularly in PCIe JBoF applications, ensuring data is stored and accessed efficiently to keep pace with growing digital storage needs.
AI and Accelerated Computing
Experience low latency and reliable PCIe and CXL connectivity for GPUs, accelerators, and FPGAs, ensuring maximum efficiency for AI and compute tasks.
Server Infrastructure
Achieve unmatched bandwidth and low-latency with PCIe for enhanced compute, storage, and network performance.
Memory Solutions
Expand bandwidth and capacity to meet the needs of data-intensive applications with advanced memory connectivity.
Networking
Leverage Active Ethernet and PCIe/CXL for superior intra-rack and inter-rack connectivity, supporting speeds up to 100G per lane.