To Grid or to Evergrid, That Is the Question

By By Derrick Harris, Editor

June 25, 2007

Since acquiring the base technology from Virginia Tech in the summer of 2004, Evergrid has been working toward the release of its initial software solutions, and when this day finally came on June 11, the IT world was reintroduced to its old friend the resource manager — but this time with some seriously unique capabilities.

Designed to function as a “datacenter resource manager,” Evergrid’s Cluster Availability Management Suite (CAMS) currently consists of Availability Services (AvS) and Resource Manager (RM-Batch), both of which are designed — as the name indicates — for environments running critical batch applications. Sitting as an abstraction layer between the operating system and the application library, the Evergrid software plays the role of a vertically integrated management system, able to handle a wide variety of functions from powering machines on and off as necessary to workload scheduling and load balancing.

The reason for all of this functionality, according to Evergrid CEO David Anderson, is to make possible the company’s vision of making today’s increasingly complex datacenters as easy-to-manage as possible by providing customers with policy-based, autonomic management, guaranteed service-level agreements (SLA) and stateful failover capabilities. The first two parts of the vision fall under the realm of resource manager, currently only RM-Batch, which features closed-loop control (including preemptive and priority scheduling) to ensure that jobs always have available to them the mandated resources. In addition, jobs can be migrated within the resource pool without losing the progress that already has been made. In the near future, Evergrid will release its Datacenter Resource Manager, which steps up the policy-based, autonomic management and SLA features while targeting virtualized enterprise datacenters running transactional and online applications.

Apart from the uniquely policy-based and fine-grained control, though, what really allows CAMS to separate from the crowd is the Evergrid AvS, which also plays a role in the aforementioned migration capability. Essentially, AvS captures the collective state, including I/O and message-passing data, from parallel applications using its checkpoint/resume feature. These captures can happen periodically on user-set intervals so that failover never is an issue. Should a job — especially a long-running, compute-intensive batch job — be interrupted due to hardware or software failure or a high-priority job that requires the same resources, the original job can either wait until the high-priority job finishes and restart from the last checkpoint or migrate to other available resources.

Anderson believes this capability is especially important in the HPC space, where long-running applications — and, as a result, businesses — often suffer due to jobs not running to completion and needing to be restarted from the top. “If … [you are] running a chip simulation for a new product release, and that chip simulation runs for two months but fails six weeks into that, you’ve just slipped the delivery of your product by six months,” hypothesized Anderson. “I’m not going to calculate the return in that case, but, trust me, the return’s a heck of a lot higher than the cost of the machines.”

Interestingly, perhaps, although Evergrid believes its resource manager has a deeper capability set than any potential competitors, the company already is integrating its AvS with these same competitors’ solutions. Platform LSF is the first for which AvS has been certified, but Anderson says the software can work with others, too, such as Altair’s PBS Professional. These companies, Anderson said, have a lot of batch customers clamoring for this feature, and “I don’t expect their customers to change to a different resource management system, so it makes complete sense for us to integrate with their product and allow the customer to continue using it the way they use it today, but with some additional capabilities.”

Customers of the presently available batch products — which, not surprisingly, are being marketed to companies (i.e., those in the manufacturing, financial services, pharmaceutical, and oil and gas markets) running computationally intensive applications in dedicated HPC environments — can expect to see between 10 and 40 percent resource utilization when utilizing the preemptive scheduling and checkpoint/resume capabilities, as well as a guarantee that jobs will complete within the expected time. In fact, one current customer, which Anderson can describe only as a “top-five Wall Street firm,” was “wowed” by the 85 percent TCO reduction it saw since adopting the Evergrid suite of products. Citing some impressive statistics, Anderson said the firm saw a 4X improvement in server utilization while reducing server administration costs by 50 percent and cutting power costs by 65 percent thanks to its Evergrid platform automatically powering off unused servers. The shutting down of these servers is facilitated by the software’s migration capabilities, which, aside from allowing jobs to continue running, can push jobs to machines with available resources, thus maximizing utilization on some servers while allowing others to rest.

CAMS also can be utilized across the globe thanks to its global meta-scheduler. According to Anderson, the company already has one customer planning to implement the software across six datacenters.

However, while Evergrid’s current and beta customers (including the aforementioned Wall Street firm, the Supercomputing Center for Education & Research (OSCER) at the University of Oklahoma and British Telecom) tend to have pretty large HPC environments, Anderson says the company’s products will be ideal for a much broader class of customer, especially with the release of the Datacenter Resource Manager, which he foresees as being attractive to any organization running a datacenter of 16-plus servers. While the solution can be effective while running over a collection of as few as eight nodes, Anderson says the sweet spot is into the hundreds of nodes, in part because this is where traditional, non-policy-driven VM environments begin to fail. “In the long term,” he said, “we’re going after the datacenter.”

In the opinion of Forrester Research vice president Jean-Pierre Garbani, though, while Evergrid’s software set is “very interesting,” he has seen this type of functionality attempted via various approaches in the past. In fact, he described Evergrid’s platform as “the first real approach to what was thought of [a few] years ago by IBM and HP as autonomic computing.” While the idea of dynamically reconfiguring resources to the needs of applications is nothing new, he says that Evergrid’s use of virtualization to meet this end is “a very clever way to solve that issue.”

As “clever” as the approach might be, what is more important, as far as Garbani is concerned, is Evergrid’s focus on the workload. “What was unrealistic in autonomic computing or utility computing,” he said, “ … is that there was absolutely no decision support,” so while users could allocate resources dynamically, there was nothing telling them when or why they should re-allocate them. Continued Garbani: “They gave you the means to do something without giving you the means to reach the decision.” With Evergrid, he says, users are given the information necessary to take action along this front. He also compares Evergrid’s solution to a distributed version of a mainframe when it comes to its aims in terms of operability.

Assuming today’s (and tomorrow’s) organizations would like to see mainframe operability from distributed systems, this should be good news for companies like Evergrid, which Garbani believes can succeed if they are able to educate the market and ride the momentum created by more and more vendors getting into the space. “Innovation by itself requires many friends in order to get adopted,” he said, and because there are some similar solutions in the market and others potentially are on the way (he believes companies like XenSource and VMware might be looking at this kind of technology as a next step), Evergrid could soon find itself part of a family of products that collectively would educate the market. “For the time being,” he said, “what I expect from Evergrid is they’re going to get a certain number of customers — the leading edge, the people who see the advantages now — but the global adoption is maybe a couple of years in the future.” This type of mindshare, he noted, could lead to acquisitions from big-time vendors like IBM, HP and Sun, who would then proceed to push it into datacenters at a significantly faster pace.

In the end, Garbani added, Evergrid’s approach to resource management is a “terrific idea” very much needed by the current market, where people are just jumping on the virtualization bandwagon without any idea how to use it.

And while Garbani’s forecast is hardly a doomsday scenario for Evergrid, CEO Anderson sees the future in an optimistic light that is withheld for company representatives contemplating a product’s success. “We think that, in the long run, customers are all going to want to consolidate their server environments,” he said, “and if they can then dynamically allocate servers to workloads in a way that manages utilization and availability seamlessly, everybody’s going to want to do it.”

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