-
High latency in industrial switches
High latency can disrupt real-time communication, causing delays in critical processes like machine control, robotics, or assembly lines. Network slicing can isolate. Latency refers to the time delay between the sending and receiving of data packets in a network. Latency is not a defined value for Ethernet as specified by the IEEE 802. While a slight delay might be imperceptible in everyday internet browsing, for industrial applications, even milliseconds of latency can have significant. Avnu presented a liaison requesting guidance regarding the problem of accumulated latency in industrial networks. Subsequently, there have been a number of the contributions on the subject.
-
ABB High Voltage Motor Relay Protection
Compact protection and control for asynchronous and synchronous motors REM620 is a dedicated motor management relay for the protection, control, measurement and supervision of medium- and large-sized asynchronous and synchronous motors in the manufacturing and process industry. Numerical relays are based on the use of microprocessors. A big difference between conventional electromechanical and static relays is how the relays are wired. Also external conditions when connecting to the power grid or during use have to be detected and abnormal conditions must be prevented. Additionally, the protection relay prevents the. To prevent this from happening more than one million protection and control relays from ABB supervise electricity distribution networks in over 100 countries, enabling the safe and reliable distribution of electric power. The main purpose of protection relays is to observe the electrical grid and. ABB Relays-Online makes finding, selecting, ordering, and tracking of your next digital substation product order quick and easy. 2, with corresponding formu-las. In these formulas the propagation of speed is included as a variable.
[PDF Version]
-
A crucial step in high computing power the optical module
CPO optical modules put optical and electronic parts together. This can cut power use by up to half. CPO technology lets more data fit in. Fiber-optic cables are creeping closer to processors in high-performance computers, replacing copper connections with glass. Technology companies hope to speed up AI and lower its energy cost by moving optical connections from outside the server onto the motherboard and then having them sidle up. NVIDIA's networking innovations, including Spectrum-X Ethernet and NVIDIA Quantum InfiniBand, are designed to handle the high-bandwidth and low-latency demands of modern AI training and inferencing at scale. The adoption of co-packaged optics (CPO) in NVIDIA's latest platforms, such as NVIDIA. This article provides a comprehensive overview of CPO optical modules, exploring their technology, benefits, challenges, and the pivotal role they play in future data centers and AI infrastructure. At the same time, startup Micas Networks, announced that it is in volume production with a CPO switch based on Broadcom's technology. As AI model training and inference scale to thousands of GPUs, traditional network architectures are being pushed to their limits.
[PDF Version]
-
AI Server Growth Data
A comprehensive report by Global Market Insights Inc. The market is expected to grow from USD 167. 56 trillion in 2034, at a CAGR of 28. Market Leader: Nvidia Corporation led with over 31%. North American CSPs' continued investments in AI infrastructure are expected to increase global AI server shipments by more than 28% YoY in 2026, according to the latest market research from TrendForce. The rapid growth of AI inference services is boosting demand for general-purpose servers. Size, Share, & Trends Analysis Report By Processor (GPU-based Servers, FPGA-based Servers), By Cooling Technology (Air Cooling, Liquid Cooling), By Form Factor, By End Use (BFSI, Automotive), By Region, And Segment Forecasts The global AI server market size was valued at USD 131. 73% during the forecast period. The North America AI server market accounted. AI Server Market (By Servers: AI Data Server, AI Training Server, AI Inference Server, Others; By Hardware: GPU, ASIC, FPGA, CPU, Others; By End-user: IT and Telecommunication, Transportation and Automotive, BFSI, Retail and E-commerce, Healthcare and Pharmaceutical, Industrial Automation, Others).
[PDF Version]
-
Which is better a mining rig or an AI server
Simply put, mining data centers focus heavily on the lowest power cost per watt. They are willing to give up backup systems for this goal. But are AI computing centers and crypto mining data centers really the same thing? Why do both industries use the word “Token,” while AI tokens and blockchain tokens follow completely different economic rules? This blog uses simple industry logic to break down the physical limits of these two types. AI does not make Bitcoin mining faster. ASICs still handle hashing, while AI improves timing, energy use, and uptime. Post-halving pressure and energy scarcity pushed miners. By mid-2025, a surprising transformation is well underway: dozens of former Bitcoin mining firms have begun to repurpose their infrastructure into AI data centers, turning their GPU-rich, power-intensive setups into rentable compute farms for training, inference, and high-performance computing. While mining rigs could be optimized with basic hardware and minimal power management, AI systems demand robust CPUs, ample memory, and high-speed connectivity to maximize GPU performance.
[PDF Version]