Abstract:Commercial clouds bring a great opportunity to the scientific computing area. Scientific applications usually require significant resources, however not all scientists have access to sufficient high-end computing systems. Cloud computing has gained the attention of scientists as a competitive resource to run HPC applications at a potentially lower cost. But as a different infrastructure, it is unclear whether clouds are capable of running scientific applications with a reasonable performance per money spent. This work provides a comprehensive evaluation of EC2 cloud in different aspects. We first analyze the potentials of the cloud by evaluating the raw performance of different services of AWS such as compute, memory, network and I/O. Based on the findings on the raw performance, we then evaluate the performance of the scientific applications running in the cloud. Finally, we compare the performance of AWS with a private cloud, in order to find the root cause of its limitations while running scientific applications. This paper aims to assess the ability of the cloud to perform well, as well as to evaluate the cost of the cloud in terms of both raw performance and scientific applications performance. Furthermore, we evaluate other services including S3, EBS and DynamoDB among many AWS services in order to assess the abilities of those to be used by scientific applications and frameworks. We also evaluate a real scientific computing application through the Swift parallel scripting system at scale. Armed with both detailed benchmarks to gauge expected performance and a detailed monetary cost analysis, we expect this paper will be a recipe cookbook for scientists to help them decide where to deploy and run their scientific applications between public clouds, private clouds, or hybrid clouds.

Cloud Versus In-House Cluster

10 Years Later: Cloud Computing is Closing the Performance Gap

A Cloud Computing Infrastructure on Heterogeneous Computing Resources.

A Study on Today’s Cloud Environments for HPC Applications

One Optimized I/O Configuration Per HPC Application

Design and Implementation of a High Availability Cluster Based on Cloud Computing

Providing a Cloud Infrastructure on Clusters

About the Suitability of Clouds in High-Performance Computing

Understanding the Performance and Potential of Cloud Computing for Scientific Applications

Building Semi-Elastic Virtual Clusters for Cost-Effective HPC Cloud Resource Provisioning

High Performance and Security in Cloud Computing

Energy-Efficient Data Processing in Cloud Computing Centers

Priority-Based Consolidation of Parallel Workloads in the Cloud

High-Performance Processing of Large-Scale Parallel Applications in Heterogeneous Cloud Computing Data Centers

Managing High-Performance Computing Applications As an On-Demand Service on Federated Clouds

Exploring Cloud Computing for Large-Scale Scientific Applications

HPC Cloud for Scientific and Business Applications: Taxonomy, Vision, and Research Challenges

Monte Carlo Performance Study of Virtual High Performance Computing Cluster over Cloud

An Adaptive Approach to Better Load Balancing in a Consumer-Centric Cloud Environment

Cloud Performance Modeling with Benchmark Evaluation of Elastic Scaling Strategies

Design and Implementation of Multi-Tenant Oriented Lightweight Virtual HPC Cluster Based on Supercomputing Environment