• Starsurge is a reliable provider of IT network system integration and solutions for global enterprise customers.

Solutions

sageconnect is a global technology company focused on high-speed network system solutions, providing product, solution design and test, global warehouse and more.

Home » Solutions »

Cisco leverages NVIDIA BlueField-3 DPU to enhance workload security and operational efficiency

As cyber attacks become increasingly complex, organizations must continuously adopt advanced solutions to protect their critical assets. One of the solutions is Cisco Secure Workload, a comprehensive security solution designed to protect application workloads across different infrastructure, locations, and form factors. Cisco recently released Cisco Secure Workload version 3.9, which takes enterprise security and operational efficiency to a new level. It provides new features to mitigate threats and vulnerabilities, and offers greater flexibility for deploying microsegmentation. Now, it has also been extended to the NVIDIA BlueField-3 data processor, whose dedicated Arm core can accelerate hardware tasks and isolate specific operations, ensuring efficient data processing and strong security, thus creating a more streamlined and secure infrastructure。 The main functions of Cisco…

More

Unlock GPU accelerated RDMA with NVIDIA DOCA GPUNetIO

NVIDIA DOCA GPUNetIO is a library in the NVIDIA DOCA SDK specifically designed for real-time inline GPU packet processing. It combines technologies such as GPUDirect RDMA and GPUDirect Asynchronous to create GPU centric applications, where the CUDA core can directly communicate with network interface cards (NICs) for sending and receiving data packets, bypassing the CPU and excluding it from the critical path. The core principles and applications of DOCA GPUNetIO have been discussed in previous articles such as “Inline GPU Packet Processing with NVIDIA DOCA GPUNetIO” and “Realizing the Power of Real Time Network Processing with NVIDIA DOCA GPUNetIO”, as well as in the DOCA GPUNetIO programming guide. Previously, DOCA GPUNetIO, along with DOCA Ethernet and DOCA Flow, was limited…

More

Integrating NVIDIA Bluefield-3 DPU with DDN storage to accelerate AI infrastructure

As AI becomes an indispensable part of organizational innovation and competitive advantage, the demand for efficient and scalable infrastructure is more urgent than ever before. The collaboration between NVIDIA and DDN Storage is setting new standards in this field. By integrating NVIDIA BlueField data processors into DDN EXAScaler and DDN Infinia and using them in innovative ways, DDN Storage is transforming data centric workloads. The synergy between NVIDIA’s advanced data processing and DDN’s powerful storage solutions can optimize AI workflows, enhance operations, improve resource utilization, and support multi tenancy in AI environments. In this article, we will focus on introducing the integration advantages of Infinia。 Integrated DPU storage solution DDN Infinia is a software defined data platform that utilizes the…

More

NVIDIA Spectrum-X accelerates large-scale AI workload optimization

Today’s rapidly developing technological landscape, maintaining a leading position is not just a goal – it is a necessary condition. The wave of innovation, especially in the field of AI, is driving a huge transformation of the entire technology stack. One area that has witnessed profound changes is Ethernet networks, which are the cornerstone of digital communication and have been the foundation of enterprise and data center environments for decades. Nowadays, every data center is accelerating to support modern AI workloads, thereby increasing the demand for infrastructure to support these workloads. Many enterprises are already very familiar with Ethernet as a trusted network standard. However, they lack a solution to fully support the characteristics of AI workloads using Ethernet protocols….

More

NVIDIA Contributes NVIDIA GB200 NVL72 Designs to Open Compute Project

During the 2024 OCP Global Summit, NVIDIA announced that it has contributed the NVIDIA GB200 NVL72 rack and compute and switch tray liquid cooled designs to the Open Compute Project (OCP). This post provides details about this contribution and explains how it increases the utility of current design standards to meet the high compute density demands of modern data centers. It also explores how the ecosystem is building on top of the GB200 designs, reducing cost and implementation time for new AI data centers. NVIDIA open-source initiatives NVIDIA has a rich history of open-source initiatives. NVIDIA engineers have released over 900 software projects on GitHub and have open-sourced essential components of the AI software stack. The NVIDIA Triton Inference Server,…

More

NVIDIA BlueField Bringing Performance and Cost Optimization to the JBOF Architecture

The development of AI has driven exponential growth in computing power, with network speeds doubling every few years. Less well-known is that it also imposes new demands on storage. Training new models typically requires high-bandwidth network access to petabyte-scale data, while inference using novel retrieval-augmented generation techniques (RAG) requires low-latency access to hundreds of terabytes of storage. New models can also train, index, and retrieve rich image and video data. Many new AI inference indexing and search tools rely on vector databases and must retain large amounts of metadata about all embedded content. Traditional file storage is often not suitable for these new workloads, and traditional storage hardware is not the most efficient or performance-oriented way to support AI. Usually,…

More

Building a Retrieval-Augmented Generation (RAG) based Question-Answering Large Language Model Workflow in NVIDIA

Leveraging Retrieval-Augmented Generation (RAG) technology to rapidly develop solutions aids in enhancing the question-answering LLM workflow, giving rise to new system architectures. NVIDIA has applied AI to its internal operations, making several significant discoveries in the balance between system functionality and user expectations. Research Findings: Regardless of the scope or use case, users typically expect to perform non-RAG tasks, such as document translation, editing emails, or even writing code. A typical RAG application might implement this by executing a retrieval pipeline for each message, leading to overuse of tokens and unnecessary latency due to the inclusion of irrelevant results. Even when applications are designed for accessing internal private data, users greatly appreciate access to web search and summarization capabilities. We…

More

NVIDIA DOCA-OFED makes the network stack future-ready

The NVIDIA DOCA software platform unleashes the potential of the NVIDIA BlueField networking platform and provides all the necessary host drivers for NVIDIA BlueField and ConnectX devices. DOCA is optimized for peak performance, enabling users to meet the demands of increasingly complex workloads. Its modular architecture offers the flexibility required to adapt to emerging technologies and higher data throughput. In the latest released network software version, NVIDIA has transitioned from the MLNX_OFED drivers and tools suite (for InfiniBand and Ethernet solutions) to DOCA-OFED, marking the end of independent MLNX_OFED releases. This shift reflects a move towards a more unified, scalable, and programmable network stack integrated within the NVIDIA DOCA framework. The final version of the lightweight subset of MLNX_OFED, MLNX_EN,…

More

Leveraging accelerated Ethernet networking and network storage to expand enterprise RAG (Robotics, Automation, and AI).

Leveraging accelerated Ethernet networking and network storage to expand enterprise RAG (Robotics, Automation, and AI). In the realm of generative AI, machines can not only learn from data but also generate human-like text, images, videos, and more. Retrieval-Augmented Generation (RAG) is a groundbreaking approach in this field. The RAG workflow is based on Large Language Models (LLMs), which can understand queries and generate responses. However, LLMs have limitations, including training complexity and a lack of up-to-date (and sometimes proprietary) information. Additionally, when not trained on specific data to answer prompts, they tend to produce hallucinations and synthesize factually incorrect information. RAG enhances queries by providing LLMs with enterprise-specific information, helping to overcome these limitations. In this article, we discuss how…

More

Copyright © 2024 starsurge mellanox network. Powered by starsurgemellanox network.