This site uses cookies to improve your experience. To help us insure we adhere to various privacy regulations, please select your country/region of residence. If you do not select a country, we will assume you are from the United States. Select your Cookie Settings or view our Privacy Policy and Terms of Use.
Cookie Settings
Cookies and similar technologies are used on this website for proper function of the website, for tracking performance analytics and for marketing purposes. We and some of our third-party providers may use cookie data for various purposes. Please review the cookie settings below and choose your preference.
Used for the proper function of the website
Used for monitoring website traffic and interactions
Cookie Settings
Cookies and similar technologies are used on this website for proper function of the website, for tracking performance analytics and for marketing purposes. We and some of our third-party providers may use cookie data for various purposes. Please review the cookie settings below and choose your preference.
Strictly Necessary: Used for the proper function of the website
Performance/Analytics: Used for monitoring website traffic and interactions
The chipmaker has released a series of what it calls Enterprise Reference Architectures (Enterprise RA), which are blueprints to simplify the building of AI-oriented datacenters. Building an AI-oriented datacenter is no easy task, even by datacenter construction standards.
Datacenters this year will face several challenges as the demand for artificial intelligence introduces an evolution in AI hardware, on-premises and cloud-based strategies for training and inference, and innovations in power distributionsall while opposition to new datacenter developments continues to grow.
As datacenters evolve from traditional compute and storage facilities into AI powerhouses, the demand for qualified professionals continues to grow exponentially and salaries are high. The rise of AI, in particular, is dramatically reshaping the technology industry, and datacenters are at the epicenter of the changes.
Cisco is boosting network density support for its datacenter switch and router portfolio as it works to deliver the network infrastructure its customers need for cloud architecture, AI workloads and high-performance computing. Cisco’s Nexus 9000 datacenter switches are a core component of the vendor’s enterprise AI offerings.
There are numerous overall trends, experts say, including: AI everything: AI mania is everywhere and without high power hardware to run it, its just vapor. All the major players Nvidia, Supermicro, Google, Asus, Dell, Intel, HPE as well as smaller vendors are offering purpose-built AI hardware, according to a recent Network World article.
Cisco has unwrapped a new family of datacenter switches it says will help customers more securely support large workloads and facilitate AI development across the enterprise. The first major service these DPUs will perform on the switch will be Layer 4 stateful segmentation throughCiscos Hypershield security architecture.
Computex 2024 is taking place in Taiwan this week, which means lots of hardware news as the OEM and parts suppliers of the world gather to show off their latest wares. The MI325X uses AMD’s CDNA 3 architecture, which the MI300X also uses. And in 2026, the AMD Instinct MI400 series will arrive, based on the AMD CDNA “Next” architecture.
Edgecore Networks is taking the wraps off its latest datacenter networking hardware, the 400G-optimized DCS511 spine switch. This new hardware offering aims to address the increasing demands of modern computing infrastructures, particularly in the realms of cloud computing and artificial intelligence.
405B workload compared to the companys H200 NVL8, which is based on the Hopper architecture. In the latest round of benchmarks, the three year old architecture still managed to achieve a 60% performance boost over last year on the Llama 2 70B workload. McGregor described Dynamo as significant: Think of an OS for an entire datacenter.
Nokia is offering customers an alternative way to set up datacenter networks by expanding its datacenter fabric package to support the open-source Software for Open Networking in the Cloud ( SONiC ). The SONiC community has grown significantly since and now features some 4,000 contributors and 28 paying members.
Nvidia has partnered with leading cybersecurity firms to provide real-time security protection using its accelerator and networking hardware in combination with its AI software. BlueField data processing units (DPUs) are designed to offload and accelerate networking traffic and specific tasks from the CPU like security and storage.
The AI revolution is driving demand for massive computing power and creating a datacenter shortage, with datacenter operators planning to build more facilities. But it’s time for datacenters and other organizations with large compute needs to consider hardware replacement as another option, some experts say.
Nvidia has partnered with hardware infrastructure vendor Vertiv to provide liquid cooling designs for future datacenters designed to be AI factories. AI factories are specified datacenters emphasizing AI applications as opposed to traditional line of business applications like databases and ERP.
These units support 400Gb Ethernet or InfiniBand networking and provide hardware acceleration for demanding storage and networking workloads. BlueField-3 accelerates networking traffic through hardware support for RoCE (RDMA over converged Ethernet), GPU direct storage and GPU initiated storage.
We are on the cusp of one of the most significant changes in the x86 architecture and ecosystem in decades – with a new level of customization, compatibility and scalability required to meet current and future customer demands,” said Intel CEO Pat Gelsinger in a statement.
AWS, Microsoft, and Google are going nuclear to build and operate mega datacenters better equipped to meet the increasingly hefty demands of generative AI. Earlier this year, AWS paid $650 million to purchase Talen Energy’s Cumulus Data Assets, a 960-megawatt nuclear-powered datacenter on site at Talen’s Susquehanna, Penn.,
The challenges of AI on WAN connectivity With the immense hardware and bandwidth requirements of AI, the challenges for AI connectivity across the WAN are numerous. Purkayastha emphasized the need for new standards and reference architectures to support the integration of GPUs into a wide range of devices, from phones to IoT. “We
One of the newer technologies gaining ground in datacenters today is the Data Processing Unit (DPU). As VMware has observed , “In simple terms, a DPU is a programable device with hardware acceleration as well as having an ARM CPU complex capable of processing data.
Secure access service edge (SASE) is a network architecture that rolls software-defined wide area networking (SD-WAN ) and security into a cloud service that promises simplified WAN deployment, improved efficiency and security, and to provide appropriate bandwidth per application. To read this article in full, please click here
Jointly designed by IBM Research and IBM Infrastructure, Spyre’s architecture is designed for more efficient AI computation. The Spyre Accelerator will contain 1TB of memory and 32 AI accelerator cores that will share a similar architecture to the AI accelerator integrated into the Telum II chip, according to IBM.
In estimating the cost of a large-scale VMware migration , Gartner cautions: VMwares server virtualization platform has become the point of integration for its customers across server, storage and network infrastructure in the datacenter. However, switching to HCI is a capital expense decision.
Cisco is updating its existing learning materials in the datacenter and security certification tracks to incorporate the AI-related information that IT professionals need at different levels of expertise. It begins with three courses on AI basics and a series on AI infrastructure requirements, Merat explained in the blog.
As a networking and security strategy, zero trust stands in stark contrast to traditional, network-centric, perimeter-based architectures built with firewalls and VPNs, which involve excessive permissions and increase cyber risk. The main point is this: you cannot do zero trust with firewall- and VPN-centric architectures.
Investors are buying up datacenters to create a Pony Express quantum signal to go coast-to-coast, he says. Quantum qubits are taking over traditional architectures for protein folding and mapping, he says. The machines are very affordable, says Holger Mueller, an analyst at Constellation Research Inc.
This includes acquisition of new software licenses and/or cloud expenses, hardware purchases (compute, storage), early termination costs related to the existing virtual environment, application testing/quality assurance and test equipment, the report reads. Add to all this personnel costs, and the expense might not be worth it.
How GPUs are different from CPUs GPUs and CPUs are both computer and server hardware components. But whereas GPUs are ideal for comprehensive, data-heavy tasks, CPUs act as a brain of sorts that manage computer systems. Tasks are parceled out into smaller, independent steps that are distributed across the GPUs architecture.
Secure Access Service Edge (SASE) is a network architecture that combines software-defined wide area networking (SD-WAN ) and security functionality into a unified cloud service that promises simplified WAN deployments, improved efficiency and security, and application-specific bandwidth policies. billion by 2025.
That worked when everything lived in the corporate datacenter. Paying a premium to backhaul traffic to a central datacenter made sense when that was where all applications lived. The Cafe-like Branch eliminates these inefficiencies while simplifying the network architecture. Ready to rethink your network?
Intel Wants To Re-Architect DataCenter Services. Intel Wants To Re-Architect DataCenter Services. Diane Bryant, senior vice president and general manager of the Datacenter and Connected Systems Group at Intel, discusses the company’s vision for the datacenter at an event Monday. Photo: Intel).
Peter Rutten, research VP, performance intensive computing, and worldwide infrastructure research at IDC, says the key takeaway from the DeepSeek results is the current approach to AI training that AI can only improve with bigger, more, and faster architecture is not justified.
AI networking is specific to the network itself, covering domains including multi-cloud software, wired and wireless LAN, datacenter switching, SD-WAN and managed network services (MNS). It’s critically important to select the right architecture for your enterprise to ensure the best outcomes.
Juniper Networks is advancing the software for its AI-Native Networking Platform to help enterprise customers better manage and support AI in their datacenters. The HPE acquisition target is also offering a new validated design for enterprise AI clusters and has opened a lab to certify enterprise AI datacenter projects.
We have invested in the areas of security and private 5G with two recent acquisitions that expand our edge-to-cloud portfolio to meet the needs of organizations as they increasingly migrate from traditional centralized datacenters to distributed “centers of data.” The new service packs will be orderable later in 2023.
At a high level, NaaS requires a scalable cloud-native architecture thats flexible, incorporates a high degree of automation, and makes great use of AI and machine learning (ML) to facilitate self-healing, streamline management, and boost observability. Managed service: Subscription-based hardware plus a managed service to operate it.
EMC Refreshes Data Protection Portfolio. At a Backup to the Future event this week EMC announced a broad array of new hardware and software products that enable customers to deploy new Protection Storage Architectures. About John Rath John Rath is a veteran IT professional and regular contributor at DataCenter Knowledge.
But modernization projects are pushing ahead: In the same PWC survey, 81% of CIOs said they prioritized cloud-based architecture as a positive and tangible step forward to improve readiness to handle future challenges. The question that remains is, can this be done with the funding available in 2025?
Intel IPUs are hardware accelerators that offload a number of tasks such as packet processing, traffic shaping, and virtual switching from the server CPU. Developers can deploy their AI-infused application via a lightweight container in a datacenter, in the cloud, or at the edge on a variety of hardwarearchitectures,” he wrote.
Alice & Bob devise cat qubits Also in January, quantum computing startup Alice & Bob announced their new quantum error correction architecture. Quantinuum provided the quantum hardware and Microsoft handled the error correction. The improvements come from both hardware and software advances.
All this has a tremendous impact on the digital value chain and the semiconductor hardware market that cannot be overlooked. The apps and tools have to gather, process and deliver back data to the consumer with minimal latency. Hardware innovations become imperative to sustain this revolution.
“The CPU becomes a data-moving machine, and unfortunately, CPUs were built to compute, not to move a ton of data back and forth. There are a number of reasons why the CPU architecture basically becomes a bottleneck.” The appliance takes over for the CPU, greatly enhancing the throughput and scalability of the GPUs.
For decades, my involvement with enterprises has shown that datacenter networking drives enterprise networking, and datacenter technology drives datacenter networking. Cisco has servers (UCS), but they dont have a significant position in the enterprise datacenter market.
DreamWorks animators use HP Z800 and Z820 workstations with Intel Xeon E5 processors, a HP FlexNetwork architecture, and HP networking to simplify collaboration among artists. About John Rath John Rath is a veteran IT professional and regular contributor at DataCenter Knowledge. Sign up for the DataCenter Knowledge Newsletter.
Open RAN (O-RAN) O-RAN is a wireless-industry initiative for designing and building 5G radio access networks using software-defined technology and general-purpose, vendor-neutral hardware. Power usage effectiveness (PUE) Power usage effectiveness (PUE) is metric that measures the energy efficiency of a datacenter.
Datacenter developer Digital Realty Trust has purchased a 5.3 megawatt datacenter, the company said today. The 15,900 square meter development at De President, Hoofddorp, Haarlemmermeer will feature six data halls, each capable of supporting 1.92 Each data hall will be designed using Digital Realty’s POD 3.0
We organize all of the trending information in your field so you don't have to. Join 83,000+ users and stay up to date on the latest articles your peers are reading.
You know about us, now we want to get to know you!
Let's personalize your content
Let's get even more personalized
We recognize your account from another site in our network, please click 'Send Email' below to continue with verifying your account and setting a password.
Let's personalize your content