Unlocking the Potential of HPC Cloud Computing

By Arpit Joshipura

July 5, 2011

Cloud computing promises a new era of flexibility and control in providing HPC data center resources. While many of today’s top HPC installations employ dedicated server clusters, cloud computing with virtualized resources is making high performance computing available to broader environments. 

Commercial enterprises can take advantage of HPC computations by deploying internal cloud capabilities or by sharing HPC resources in public clouds. In the cloud model, data center managers can mix and match computing, storage, and networking resources to provide an agile and highly flexible resource for customer applications.

To realize the full potential, this paradigm requires open, standardized interfaces between data center layers of compute resources, the network, and storage elements. While the industry has moved toward open computing and storage layers over the past few years, networking has remained largely proprietary.

Today, however, HPC data center owners can use open systems to unlock the network layer so they can get the most out of their data center architectures and in turn get the most out of their cloud deployments.

The Lagging Network Layer

The history of computing has been one of movement from proprietary to open standards.  From the mainframe days in which computing, network, and storage resources were all proprietary, the industry has moved inexorably toward open, standards-based layers in the data center. For example, server and software vendors have embraced Intel X86 architectures with standard operating systems such as Windows, Linux, VMWare, and others. In the same way, storage vendors have coalesced around the standards of Fibre Channel, FCoE, and iSCSI.

At the same time, the pace of innovation has accelerated. Mainframe development cycles were measured in multiple years because all layers of the data center stack came from the same vendor. Now, with a whole ecosystem of standards-based products in development, the pace of change is measured in months. For example, a network switch vendor that relies on its own ASICs is tied to a three-year development cycle for product revisions, while a switch vendor using merchant silicon can introduce new products much more quickly.

In this environment, customers are less willing to be tied down to the pace of innovation of just one vendor. Industry leaders from the world’s largest social networking, search engine, and HPC environments are advancing the state of networking at a phenomenal pace by combining open products in their own way, rather than expecting a vendor to tell them how to build a data center.

Ultimately, all layers of the data center stack – computing, storage, and networking – should be open to deliver the maximum customer choice and flexibility. Architectures, automation, and ecosystems should all be open to support the widely varied needs of data center owners.

This move toward openness has made it easier to build dynamic, virtualized data centers in which compute resources can be created or decommissioned on the fly – the essence of cloud computing. Unfortunately, the network layer has not kept pace in the drive toward openness. Each of the largest networking vendors has developed proprietary software, control plane, or interconnect technology that limits user choices when it comes to compute servers and storage units. By limiting choice, the closed network layer forces users to build data centers in a specific way that may not meet their needs. In contrast, an open ecosystem enables architects to build data centers flexibly and quickly using the most advanced technologies.

An open systems framework unlocks the network to provide greater flexibility, performance, and manageability to cloud and conventional data centers. Based on an expanding portfolio of hardware, system and automation software, and partner/customer ecosystem services, an open systems framework enables data center architectures that are driven by customer needs rather than vendor needs.

Open systems networking is based on three elements: open architectures, open automation, and open ecosystems.

Open Architectures

An open systems infrastructure relies on open, standards-based technology for interfaces, interconnect, control plane, and other aspects of network operations. The network thus supports any computing or storage solution that also supports open standards. This approach gives data center owners unrivaled flexibility in choosing solutions for their specific needs. No two data centers are exactly the same, so in providing this flexibility, an open systems approach meets varied data center needs as no other approach can.

Open systems architectures include switches for the core and the top-of-rack (ToR). In the core, data center owners should be able to choose either a traditional hierarchical architecture designed for traditional data centers or a next-generation distributed architecture optimized for fabric deployments. Both architectures must be standards-based and capable of advancing any existing environment with higher performance and lower cost structures. 

At the top of rack, switches nominally provide aggregation points for servers and storage, but it is now possible to combine more functionality in the switch. Next-generation ToR switches can combine networking, applications, and storage networking. This type of switch eliminates extra servers and appliances by combining flexible application bays with 10/40GbE ports, bringing a new level of convergence to the ToR. Additionally, any of the Ethernet ports can also be configured for native Fibre Channel or Fibre Channel over Ethernet (FCoE) in software.

The application bays can include blade servers that enable users to run load balancing, firewall, security, management, or other applications directly on the switch to reduce complexity and streamline operations in the data center.

Open Automation

One of the traditional objections to open systems is the inability to manage them as a single entity. Open automation provides standards-based automation for data center operations such as bare metal provisioning, configuration management, and monitoring. Data center managers have the ability to automate other control or monitoring functions with standard scripting, using Perl or Python. Automation is critical for data centers of any size as it allows operators to dynamically stitch together network, compute, and storage resources. For maximum choice, automation shouldn’t be something that forces users down a single-vendor path. Open automation allows users to automate with their choice of solutions.

With the adoption of virtualization, data centers have become more responsive and efficient but more complex. IT managers must now manage hundreds to thousands of virtual machines and their associated storage and networks. Data center infrastructure must be more responsive, quickly adapting to changes in application requirements. Additionally, server, storage and network infrastructure can no longer be managed as separate silos, but rather as a single, dynamic environment.

While large, dedicated HPC installations tend to use single image servers to maximize computations, virtualization is enabling HPC capabilities to be offered to broader audiences with less total expense.  Open automation addresses these management challenges using industry standards and common industry technology, allowing IT managers to deploy virtualized environments using best-of-breed technology.  Standards such as Edge Virtual Bridging (802.1Qbg) will be instrumental in providing data centers with complete manageability of virtualized resources.

Open Ecosystems

Open ecosystems bring together leading makers of standards-based go-to-market solutions and technology to offer unrivaled flexibility and choice in the selection of best-of-breed solutions for the data center. Open ecosystems are a critical and necessary element to unlock the full potential of data center deployments. Simply, put, there is safety in numbers – the more people trying to solve problems and innovate, the better. Ultimately, choice comes from having the broadest ecosystems.

Cloud computing is rapidly gaining acceptance as the primary model for delivering flexible compute resources with high agility and reduced management costs. However, the vision of cloud computing can’t be realized to its fullest potential unless the computing, networking, and storage layers of the data center are all derived from open, standardized interfaces and software. An open systems networking framework delivers the openness required at the network layer combined with flexible open automation software allowing HPC data center managers maximum choice with uncompromised price and performance.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire