Energy efficiency, resource sharing, dynamic scaling enhance environmental awareness

In this time of increasing awareness of environmental sustainability, players within the media-tech industry are seeking innovative solutions to reduce their carbon footprint. In one approach, the media- and content-creation sector is moving toward remote production in the cloud, a process that was accelerated by the pandemic.

Cloud production, leveraging the vast capabilities of cloud service providers, introduces a paradigm shift in how media is created, managed, and distributed. One of the primary sustainability advantages lies in ensuring that cloud providers also rely on renewable energy sources and maximize energy efficiency within the infrastructure. Many cloud providers invest in technologies both hardware and software that operate with high energy efficiency, reducing operational costs while contributing to a smaller overall carbon footprint.

Some more-sustainable data centers have incorporated cooling systems that use outdoor air and direct evaporative cooling to save both energy and water.

Resource sharing is a compelling factor driving sustainability in cloud production. A multi-tenant model enables multiple clients to efficiently share the same infrastructure, optimizing resource utilization particularly in software-as-a-service applications. This collaborative approach minimizes the need for individual organizations to maintain excess capacity, leading to a more efficient use of computing resources.

Dynamic scaling, up or down, allows organizations to provision resources based on demand. The elasticity ensures that computing power is used only when necessary, reducing energy consumption during periods of lower demand, and operational efficiency is improved by avoiding unnecessary resource consumption.

The cloud production model can address the issue of hardware-lifecycle impact. Traditionally, organizations managing on-premises production face challenges related to frequent hardware upgrades and replacement, contributing to electronic waste. Cloud providers are responsible for infrastructure maintenance, including upgrades. By managing the lifespan of equipment, cloud providers help reduce electronic-waste generation.

Despite the advantages, cloud providers are in fact data centers, sometimes very large facilities with a very large physical footprint. Being on 24 hours per day, seven days per week, the facilities use a great deal of energy, and, according to recent estimates, cloud computing generates 1.55% of all global carbon emissions. Understanding this impact is critical in improving the situation. The energy used to run these facilities and the water used to cool them are being explored in an effort to reduce their environmental footprint and impact. Using renewable energy sources and innovative cooling technologies, cloud providers are actively moving toward more-sustainable operations.

Production in the cloud represents a growing sustainable option in the media- and content-creation industry. Energy efficiency, resource sharing, dynamic scaling, and a commitment to green initiatives by cloud providers offer a more environmentally conscious approach to content creation. This shift not only aligns with global sustainability goals but also positions the industry as a responsible steward of the planets resources. Though not yet perfect, as technology continues to advance, cloud production is reshaping the future of media in a more sustainable manner.

Note: The views expressed here are solely those of the author and do not represent positions of SVG.

See the original post here:
Cloud Production Offers an Option for Sustainability in Media and Content Creation - Sports Video Group

Chipmaker Ampere Computing LLC, which makes powerful, Arm-based computer chips for smartphones and data center servers, said today its partnering with Qualcomm Inc. on a new joint solution that will enable enhanced artificial intelligence inference at lower costs.

The new partnership was announced today during an annual update (below) that saw Ampere detail its 2024 product roadmap, which is focused on advancing high performance, power-efficient computing for cloud and AI workloads. Aside from the partnership with Qualcomm, one of the major highlights to look forward to is the launch of its new AmpereOne platform, with the first product being a 12-channel, 256-core central processing unit built on an advanced N3 process node.

Ampere is an agnostic, Arm-based chip developer that has forged strong partnerships with public cloud infrastructure giants such as Amazon Web Services Inc., Google Cloud and Microsoft Corp., helping them to design their own, proprietary data center processors. For instance, Ampere worked with AWS to design that companys Graviton chips.

The companys main backer is another cloud infrastructure provider, Oracle Corp., which has invested more than $400 million in the company. Oracle was one of the first cloud companies to adopt Amperes Altra central processing units in an effort to increase its competitiveness in the cloud computing industry.

Ampere has seen a lot of success in the cloud, in contrast to other Arm chip manufacturers such as Qualcomm, Marvell Technologies Inc., Advanced Micro Devices Inc. and Samsung Electronics Co. Ltd., which all have so far failed to make much impact in the market for Arm-based data center chips.

Amperes Altra CPUs are said to be customized to run real-time AI workloads such as chatbots, data analytics and video content analysis, offering rapid inference capabilities at a fraction of the cost of Nvidia Corp.s powerful graphics processing units, which power the majority of AI applications today.

In todays update, Ampere explained that it will work with Qualcomm to create a joint, cloud-based AI processor that combines its most advanced Ampere Altra CPUs with Qualcomms low-powered Cloud AI100Ultra AI inference cards for data center servers.

According to Ampere Chief Product Officer Jeff Wittich, the new offering is designed to enable large language model inference for some of the most powerful generative AI models currently available. Weve taken our Ampere CPUs and paired them with AI100 Ultra accelerators from Qualcomm, he told SiliconANGLE in an interview. They will be deployed via Supermicro servers initially and eventually come to others.

Wittich explained that AI is driving enormous demand for lower-cost data center processors, as companies are becoming increasingly concerned about the high costs of running workloads on GPUs, which are also becoming more difficult to procure.

For AI training, most companies have GPUs and thats fine, but theyre too power-hungry and expensive for inference workloads, Wittich explained. A lot of companies are saying they cannot continue to deploy lots of high-powered and expensive GPUs. We can help solve that problem.

Wittich said CPUs are ideal for many LLM models, with Amperes regular chips more than able to cater to the smaller ones of 7 billion parameters. And soon, theyll also be able to handle much bigger LLMs of up to 70-billion parameters. For those, thats where Qualcomms solutions come in, he said.

The combined Ampere/Qualcomm offering will be enhanced by the launch of Amperes most advanced AmpereOne CPU, which the company claims can deliver 40% greater performance than any existing CPU available today, even without any exotic platform designs. The 192-core, 12-channel memory chip will go into production soon and is expected to become available later this year.

Ampere said its efforts in the AI industry are being validated by strong customer adoption. One of its newest customers is Meta Platforms Inc., which is now running its Llama 3 LLM on Ampere CPUs in Oracles cloud platform. In an update, the company showed data that illustrates how Llama 3 running on a 128-core Ampere Altra chip with no GPU delivers the same performance as an Nvidia A10 GPU paired with an x86 CPU, while using two-thirds less energy.

Last year, Ampere announced that it had become a founding member of the new AI Platform Alliance, which sees a number of chipmakers combine their expertise to develop more sophisticated platforms for AI compute. The Alliances latest initiative will see Ampere leverage the open interface technology in its Ampere chips to incorporate rival chipmakers intellectual property into future CPUs. This suggests that the collaboration with Qualcomm might just be the first of many more in the coming years.

Ampere Chief Executive Renee James said the increasing power requirements and energy challenges associated with AI mean that the companys low-power Arm chip designs are becoming more relevant than ever.

We started down this path six years ago because it is clear it is the right path, she said. Low power used to be synonymous with low performance. But Ampere has proven that isnt true. We have pioneered the efficiency frontier of computing and delivered performance beyond legacy CPUs in an efficient computing envelope.

James said the use of power-hungry GPUs for running AI workloads is unsustainable in the long term, as the demands of AI models grow exponentially. Whats more, Nvidia is reportedly struggling to keep up with enterprise demand for its specialized processors. We believe that the future data center infrastructure has to consider how we retrofit existing air-cooled environments with upgraded compute, she said.

In addition, James stated her belief that the industry needs to build more sustainable data centers that do not put excessive strain on existing energy grids. That is what we enable at Ampere, she said.

With reporting from Robert Hof

THANK YOU

Read more:
Ampere to integrate its CPUs with Qualcomm's cloud-based AI chips for running large models - SiliconANGLE News

May 14, 2024

New board will replace Joint Authorization Board and govern FedRAMP provisional authorization

WASHINGTON The U.S. General Services Administration has announced a new board that will serve as the official governing body for the Federal Risk and Authorization Management Program (FedRAMP), replacing the Joint Authorization Board (JAB).

The FedRAMP programs commitment is to make it safe and easy for the U.S. government to take full advantage of cloud services to help agencies deliver on their missions.

The new board brings a wealth of technology, cybersecurity, and engineering expertise from federal agency executives who will work to champion the vision of FedRAMP and make that vision successful, said Eric Mill, GSAs executive director for cloud strategy. The board will approve and help guide FedRAMP policies, bring together the federal community to create a robust authorization ecosystem, and be a critical partner to the FedRAMP program in our shared goal of a more streamlined customer experience and stronger federal cybersecurity.

The Office of Management and Budget established the FedRAMP Board and appointed the inaugural board members under the FY23 National Defense Authorization Act. The FedRAMP Board intentionally comprises members from across government, bringing diverse perspectives from the frontlines of cyber and IT modernization efforts, stated Deputy Federal Chief Information Officer Drew Myklegard. By harnessing their collective expertise, the board will play a vital role in adapting the FedRAMP Program to address the evolving cyber landscape and enable the accelerated adoption of secure cloud technologies across the government.

The inaugural board members are:

The initial focus of the board is to ensure a smooth transition from the JAB and any work in progress. The board is also focused on defining metrics for the program, engaging with agencies to perform joint-agency and single-agency authorizations, and working with FedRAMP to continuously monitor FedRAMP-authorized cloud products and services.

For more information about FedRAMP, please visit fedramp.gov.

###

About GSA: GSA provides centralized procurement and shared services for the federal government, managing a nationwide real estate portfolio of nearly 370 million rentable square feet, overseeing about $100 billion in products and services via federal contracts, and delivering technology services that serve millions of people across dozens of federal agencies. GSAs mission is to deliver the best customer experience and value in real estate, acquisition, and technology services to the government and the American people. For more information, visit GSA.gov and follow us at @USGSA.

Read more:
FedRAMP board launched to support safe, secure use of cloud services in government - GSA

Join us in returning to NYC on June 5th to collaborate with executive leaders in exploring comprehensive methods for auditing AI models regarding bias, performance, and ethical compliance across diverse organizations. Find out how you can attend here.

Today, Samsung Next, the venture capital arm of Samsung Electronics, announced it is making a strategic investment in Seattle-based Expanso, a startup looking to power distributed data processing.

The companies expect this move to help both of them, streamlining how Samsung handles the vast swathe of data coming from its globally distributed workloads and providing Expanso with more reach and growth.

Samsung is a global company with products and services that span every aspect of customers lives, David Aronchick, CEO of Expanso, said in a statement.Teaming up with Samsung Next is a no-brainer, and is core to our mission of tackling those pervasive data challenges. Our role is more about bridging gaps and enabling our customers to freely choose how and where to handle their varied workloads, all while ensuring they maintain full control over their data.

Today, enterprises are leaning heavily towards globally distributed workloads and generating data across various environments, including cross-regions, cross-cloud, IoT and edge devices and even on unreliable networks.

The AI Impact Tour: The AI Audit

Join us as we return to NYC on June 5th to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.

Usually, teams tend to extract insights from this data by moving all of it across networks through complex ETL pipelines and centralizing everything in a cloud data platform.

The approach works well (allowing for BI/AI applications) but also takes a lot of time and financial resources at the same time.

Even building custom peer-to-peer computing solutions doesnt work as it increases developer time, maintenance costs and forces teams to learn and reinvent 40 years of distributed computing knowledge.

Expanso takes a decentralized approach to this challenge by providing enterprises with Bacalhau an open-source distributed data processing platform.

The offering runs on the distributed systems organizations have already deployed (or plan to) and schedules data processing jobs against the data right where its generated, be it on the cloud or on the edge.

This enables teams to analyze data in situ, reducing the operational overhead of replicating data centers or managing data movement between clouds.

Not to mention, it also helps increase the speed of data jobs while also bolstering security at the same time.

Now, with the strategic backing of Samsung Next, Expansos Bacalhau is expected to support the global distributed workloads of the electronics major which has platforms spanning from mobile devices and computers all the way up to heavy industrial deployments.

This would also be very critical as the company and the industry at large continues to witness a surge in global data generation, with emerging technologies such as generative AI and 5G networks.

It would be hard to imagine a better fit for Expansos distributed computing architecture, in terms of providing faster answers, more reliable execution, and integrated security, even over spotty networks. There is no doubt that partnering with Samsung will be a wonderful catalyst for all of our use cases, Aronchick told VentureBeat.

While it remains unclear how exactly Samsung will leverage Bacalhau across its environments, Andy Duong, an investor at Samsung Next, did indicate some applications, including video processing on the smallest edge devices and sharing sensitive petabyte (or larger) datasets between enormous organizations.

Were excited to invest in Expanso for many reasons, including the potential for new applications driven by real-time data analysis, more efficient resource utilization than in traditional data centers, and the ability to empower users with greater data privacy control. We believe an approach that recognizes both data gravity, and the underused power of distributed devices while giving developers the same reliability and verifiability as running in a traditional data center will be part of every organizations IT infrastructure in the future, Duong told VentureBeat.

According to Aronchick, since the launch of Bacalhaus public demo last year, the platform has been used to run over 2 million jobs across use cases. The exact revenue stats remain unclear but some of the companys clients are heavyweights such as the U.S. Navy, University of Maryland, Prelinger Labs and WeatherXM.

From distributed ETL, to log processing, to Edge ML (including video processing), we have seen customers take our platform and use it in tons of ways we could never have imagined. For each significant case, we have been able to build out solutions and open-source the project so that other folks can do the exact same thing. This means that whether you are in retail, manufacturing, transportation, or high-traffic web/mobile applications, you can use Bacalhau from Expanso quickly and easily, with a lot of the basics taken care of for you, he added.

Moving ahead, the company plans to focus on adapting its platform to all the ways people are already using it with easier frameworks, integration into more services, and better getting-started examples. This, the CEO hopes, will also shake off many of the reservations about building distributed computing architectures.

VB Daily

Stay in the know! Get the latest news in your inbox daily

By subscribing, you agree to VentureBeat's Terms of Service.

Thanks for subscribing. Check out more VB newsletters here.

An error occured.

Here is the original post:
The end of centralized data? Samsung teams with Expanso on distributed processing - VentureBeat

Platform as a Service (PaaS) provider Rafay has extended its Kubernetes management platform to better support enterprise AI and ML workloads, with a focus on GPU resource management, democratizing access to ML pipelines, and assisting with model testing and selection.

The new capabilities make compute resources for AI instantly consumable by developers and data scientists with enterprise-grade guardrails, said Haseeb Budhani, co-founder and CEO of Rafay Systems.

Rafay is a Kubernetes company that helps customers manage their environments, including Kubernetes, CI/CD pipelines and deployment platforms.

The company noticed customers deploying AI workloads on Kubernetes using Rafay’s product, and identified three gaps they could address, Budhani told The New Stack. The first gap is efficiently consuming and sharing expensive GPU resources. Rafay extended its existing PaaS to provide GPU resources to internal customers, with features like time limits and cost management.

“What we saw happen was our customers were deploying AI workloads on Kubernetes, and using our product to do it, unbeknownst to us,” Budhani said.

The second gap is democratizing access to machine learning (ML) pipelines beyond just data scientists. Rafay introduced an AI/ML workbench on top of their platform to make consuming these pipelines easier for everyone in an enterprise.

The third gap is testing and selecting the best ML models. Rafay added an “LLM playground” layer between the PaaS and ML workbench to allow users to quickly test and select the best models for their needs, Budhani said.

Rafay’s newly added support for GPU workloads helps enterprises and managed service providers power a new GPU-as-a-service experience for internal developers and customers.

Rafay’s new AI Suite provides standards-based pipelines for machine learning operations (MLOps) and large language model operations (LLMOps) to quicken the development and deployment of AI applications.

Moreover, as the global GPU-as-a-service market is expected to reach $17.2 billion by 2030, organizations are seeking scalable solutions to connect their data scientists and developers to accelerated computing infrastructure.

Rafay’s PaaS now addresses issues like environment standardization, self-service consumption of compute, secure use of multitenant environments, cost optimization, and auditability for GPU-based workloads.

“GPU-accelerated workloads are a growing part of enterprise portfolios and organizations need scalable tools to manage them,” said Justin Warren, founder and principal analyst at PivotNine. Customers also want to maintain tight control over the sovereignty of sensitive data, a challenge that is only growing in complexity. It’s good to see Rafay providing enterprises with options beyond the narrow vision of a few major cloud providers.”

The new features for GPU workloads include developer and data scientist self-service, AI-optimized user workspaces, GPU matchmaking, and GPU virtualization.

“Beyond the multicluster matchmaking capabilities and other powerful PaaS features that deliver a self-service compute consumption experience for developers and data scientists, platform teams can also make users more productive with turnkey MLOps and LLMOps capabilities available on the Rafay platform,” Budhani said in a statement. “This announcement makes Rafay a must-have partner for enterprises, as well as GPU and sovereign cloud operators, looking to speed up modern application delivery.”

NTT DATA has been an early user of Rafay’s new AI capabilities and has collaborated with the Rafay team to help deliver its new GPU support and AI Suite to market.

“Rafay’s approach satisfies users responsible for application development and management, making it easy to cross-collaborate within enterprises’ security and budget boundaries,” said Mike Jones, vice president of partners and alliances at NTT DATA, in a statement.

YOUTUBE.COM/THENEWSTACK

Tech moves fast, don't miss an episode. Subscribe to our YouTube channel to stream all our podcasts, interviews, demos, and more.

SUBSCRIBE

The rest is here:
Rafay's PaaS Now Supports GPU Workloads for AI/ML in the Cloud - The New Stack