Entries filed under “Cloud HPC”

Over at Computing Now, Art Sedighi writes that while cloud, grid, and HPC remain as distinct approaches, the twist in recent years has been the ability to coordinate and integrate these seemingly different environments. To illustrate, he describes a new peer-reviewed paper that introduces the concept of a meta-scheduler that can move workloads across all three environments.

Over the last decades, the cooperation amongst different resources that belong to various environments has been arisen as one of the most important research topic. This is mainly because of the different requirements, in terms of jobs’ preferences that have been posed by different resource providers as the most efficient way to coordinate large scale settings like grids and clouds. However, the commonality of the complexity of the architectures (e.g. in heterogeneity issues) and the targets that each paradigm aims to achieve (e.g. flexibility) remains the same. This is to efficiently orchestrate resources and user demands in a distributed computing fashion by bridging the gap among local and remote participants. At a first glance, this is directly related with the scheduling concept, which is one of the most important issues for designing a cooperative resource management system, especially in large scale settings. In addition, the term meta-computing, hence meta-scheduling, offers additional functionalities in the area of interoperable resource management because of its great proficiency to handle sudden variations and dynamic situations in user demands by bridging the gap among local and remote participants. This work presents a review on scheduling in high performance, grid and cloud computing infrastructures. We conclude by analysing most important characteristics towards inter-cooperated infrastructures.

Read the Full Story or Download the paper (PDF).

Over at Admin HPC, Dell’s Jeff Layton writes that a pair of recent use cases have helped changed his mind about the validity of using the Cloud for HPC.

At first, it was fairly easy to dismiss cloud computing for traditional HPC workloads. The “HP,” after all, stands for “high performance,” and doing anything to reduce performance is counterproductive. You are paying more and getting less. However, new workloads are being added to HPC all of the time that might be very different from the classic MPI applications in HPC and have different characteristics. The amount of computation in these new workloads is increasing at an alarming rate – so much so, that I think HPC is giving way to RC (research computing).

Read the Full Story.

Over at GigaOm, David Meyer writes that European IaaS provider CloudSigma has abandoned magnetic disks for solid-state storage. After a pilot test of SolidFire’s all-SSD storage system, CloudSigma now feels confident enough to offer a service-level agreement for performance, as well as uptime.

According to CloudSigma COO Bernino Lind, the shift to SSD is a major help when it comes to handling HPC workloads, such as those of Helix Nebula users CERN, the European Space Agency (ESA) and the European Molecular Biology Laboratory (EMBL):

They want to go to opex instead of capex, but the problem is there is no-one really who does public infrastructure-as-a-service which works well enough for HPC. There is contention — variable performance on compute power and, even worse, really variable performance on IOPS [Input/Output Operations Per Second]. When you have a lot of I/O operations, then you get all over the spectrum from having a couple of hundred to having 1,000 and it just goes up and down. It means that, once you run a large big data setup, you get iowaits and your entire stack normally just stops and waits.” Lind pointed out that, while aggregated spinning-disk setups will only allow up to 10,000 IOPS, one SSD will allow 100,000-1.5 million IOPS. That mitigates that particular contention problem. “There should be a law that public IaaS shouldn’t run on magnetic disks,” he said. “The customer buys something that works sometimes and doesn’t work other times – it shouldn’t be possible to sell something that has that as a quality.”

Read the Full Story.

In an April 1 blog post, Jason Stowe writes that his company Cycle Computing has launched new intelligent orchestration software called “ZeroCompute.”

ZeroCompute solves access to remote high performance cloud computing environments with a patent pending technology designed to accelerate access to HPC and BigData systems. This approach can simultaneously manage science that might normally require brontobyte datasets, and five hundred billion concurrent jobs. After years building Utility HPC & Utility Supercomputing software, Cycle’s engineers have discovered that the fastest way to execute any algorithm is to just not run it at all. Although seemingly obvious, by simply not running the science on any cores, we remove the computational challenges and data transfer bottlenecks of today’s BigCompute and BigData workloads. Because ZeroCompute “completes” the floating point in the nanosecond it takes the software to decide not to run anything, accordingly, the system exceeds a peak floating-point performance of over one billion petaflops, at a cost of $0.00 per flop.

Read the Full Story and be sure to enjoy the rest of April Fool’s Day.