Learning from Clouds Past: A Look Back at Magellan

By Tiffany Trader

February 1, 2012

In 2009, the US Department of Energy (DOE) launched a bold experiment, a $32 million program to assess the benefit of cloud computing to the scientific community. A distributed testbed infrastructure, named Magellan, was established at the Argonne Leadership Computing Facility (ALCF) and the National Energy Research Scientific Computing Center (NERSC) to provide a tool for computational science in a cloud environment. Magellan, with funding from the American Recovery and Reinvestment Act, was to help major research organizations answer the classic cloud question: is it better to rent or buy?

“What we’re exploring is the question of whether the DOE or other government agencies should be buying their own clusters … or whether those kinds of purchases should be done in a more consolidated way,” said NERSC Director Kathy Yelick in a previous article.

Despite high-hopes and community support, in late 2011, we learned that the Magellan project was being discontinued, leaving many wondering what happened. Now we have some answers in the form of a 169-page report, sponsored by the Department of Energy’s Office of Advanced Scientific Computing Research (ASCR), which funded the study to assess what Magellan tells us about the the role of cloud computing for scientific applications.

Since industry was already benefiting from the cloud model, from the economies of scale generated by a shared pool of network-accessible resources, the Magellan team members initially set out to determine if cloud would hold the same potential for science. As stated in the executive summary:

The goal of Magellan, a project funded through the U.S. Department of Energy (DOE) Office of Advanced Scientific Computing Research (ASCR), was to investigate the potential role of cloud computing in addressing the computing needs for the DOE Office of Science (SC), particularly related to serving the needs of mid-range computing and future data-intensive computing workloads. A set of research questions was formed to probe various aspects of cloud computing from performance, usability, and cost.

Specifically, Magellan was tasked with addressing the following questions:

  • Are the open source cloud software stacks ready for DOE HPC science?
  • Can DOE cyber security requirements be met within a cloud?
  • Are the new cloud programming models useful for scientific computing?
  • Can DOE HPC applications run efficiently in the cloud? What applications are suitable for clouds?
  • How usable are cloud environments for scientific applications?
  • When is it cost effective to run DOE HPC science in a cloud?

It should be noted that Magellan was not a typical commercial cloud, rather this “science cloud” was purpose-built for the special requirements of scientific computing. Magellan was based on the IBM iDataplex chassis using Intel processor cores for a theoretical peak performance of over 100 teraflop/s. Other components include:

  • High bandwidth, low-latency node interconnects (InfiniBand).
  • High-bin processors tuned for performance.
  • Preinstalled scientific applications, compilers, debuggers, math libraries and other tools.
  • High-bandwidth parallel file system.
  • High-capacity data archive.

During Magellan’s two-year run, the staff at NERSC and Argonne National Laboratory examined how different aspects of cloud computing infrastructure and technologies could be harnessed by various scientific applications. They evaluated cloud models such as Infrastructure as a Service (IaaS) and Platform as a Service (Paas), virtual software stacks, MapReduce and open-source implementation (Hadoop), as well as resource provider and user perspectives.

Using a wide-range of applications as benchmarks, the researchers compared the Magellan cloud with various other architectures, including a Cray XT4 supercomputer, a Dell cluster system, and Amazon’s EC2 commercial cloud offering. Despite the testbed moniker, a lot of important production science took place, contributing to advances in particle physics, climate research, quantum chemistry, plasma physics and astrophysics.

Science workloads, by their nature, tend to be cloud-challenged, although to varying degrees. The report outlines the three major classifications of computational models, beginning with large-scale tightly-coupled science codes, which require the power of traditional supercomputers and take a big penalty working in a virtualized cloud environment. Then, there are the mid-range tightly-coupled applications, which run at a smaller scale and tend to be good candidates for cloud, although there is some performance loss. The final category, high-throughput workloads, usually involve asynchronous, independent computations, and in the past relied on desktop and small clusters for processing. But due to an explosion in sensor data, cloud is a good fit, especially when you factor in the fact that these high-throughput and data-intensive workloads do not fit into current scheduling and allocation policies.

The two-year Magellan project led to these key findings:

  • Scientific applications have special requirements that require cloud solutions that are tailored to these needs.
  • Scientific applications with minimal communication and I/O are best suited for clouds.
  • Clouds require significant programming and system administration support.
  • Significant gaps and challenges exist in current open-source virtualized cloud software stacks for production science use.
  • Clouds expose a different risk model requiring different security practices and policies.
  • MapReduce shows promise in addressing scientific needs, but current implementations have gaps and challenges.
  • Public clouds can be more expensive than in-house large systems. Many of the cost benefits from clouds result from the increased consolidation and higher average utilization.
  • DOE supercomputing centers already achieve energy efficiency levels comparable to commercial cloud centers.
  • Cloud is a business model and can be applied at DOE supercomputing centers.

From this list, it is apparent that cloud was unable to measure up to a centralized supercomputer system in many ways, but the delivery model does have its place. According to the report, “users with applications that have more dynamic or interactive needs could benefit from on-demand, self-service environments and rapid elasticity through the use of virtualization technology, and the MapReduce programming model to manage loosely coupled application runs.”

In other words, cloud excels when it comes to flexibility and responsiveness. In fact, the report found that “for users who need the added flexibility offered by the cloud computing model, additional costs may be more than offset by the increased flexibility. Furthermore, in some cases the potential for more immediate access to compute resources could directly translate into cost savings.”

However, when it comes to the potential cost savings of using a public cloud versus the costs of hardware acquisition, the report makes the point that DOE procurement costs are often significantly discounted, which offsets some of the potential savings:

Existing DOE centers already achieve many of the benefits of cloud computing since these centers consolidate computing across multiple program offices, deploy at large scales, and continuously refine and improve operational efficiency. Cost analysis shows that DOE centers are cost competitive, typically 3-7x less expensive, when compared to commercial cloud providers. Because the commercial sector constantly innovates, DOE labs and centers should continue to benchmark their computing cost against public clouds to ensure they are providing a competitive service.

“Cloud computing is ultimately a business model,” state the authors. “But cloud models often provide additional capabilities and flexibility that are helpful to certain workloads. DOE labs and centers should consider adopting and integrating these features of cloud computing into their operations in order to support more diverse workloads and further enable scientific discovery, without sacrificing the productivity and effectiveness of computing platforms that have been optimized for science over decades of development and refinement.”

The authors further suggest that when an integrated approach is not sufficient, a private cloud solution should be considered based on its ability to provide many of the benefits of commercial clouds while avoiding some of the pitfalls, such as security, data management, and performance penalties.

To recap: cloud services are a good complement to centralized computing resources, but not a replacement. This should not come as a surprise to our community. This is HPC, high-performance computing, and whenever you add additional layers, i.e., virtualization, the application takes a performance hit. However, as the report makes clear, there are good use cases for cloud services, such as “scientific groups needing support for on-demand access to resources, sudden surges in resource needs, customized environments, periodic predictable resource needs (e.g., monthly processing of genome data, nightly processing of telescope data), or unpredictable events such as computing for disaster recovery.” The report goes on to note that “cloud services essentially provide a differentiated service model that can cater to these diverse needs, allowing users to get a virtual private cluster with a certain guaranteed level of service.”

Magellan was billed as an exploratory project, set to go for two years. In fact, the project was named Magellan in honor of the Portuguese explorer Fernão de Magalhães, the first person to lead an expedition across the Pacific. The original “clouds of Magellan” refers to two small galaxies in the southern sky. The current-day Magellan, as the first major scientific cloud testbed, also navigated uncharted waters and documented the journey for the benefit of future generations.

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