Download presentation
Presentation is loading. Please wait.
Published byRaymond O’Neal’ Modified over 8 years ago
1
VIII Encuentro Grupos Investigación Sep 2010 Grupo oneGrid Cluster Computing Grid Computing Cloud Computing Vicerrectoría de Ciencia, Tecnología e Innovación Universidad Antonio Nariño
2
Content Introduction (Clusters, Grids, Clouds?)EHEP ATLAS Research Group (Clustering)Grid Colombia (Grid Computing) Sistema de Gestión de la Investigación (Cloud Computing) Bimler Cephalometric Analysis (SOA)
3
Introduction Research Question Research Question Since the inception of computers (~1950’) the research community has always been characterized by requiring a high computational capacity, having at the same time hard budget constraints.
4
Introduction Research Question Research Question How to provide such a computational capacity at minimum cost ? How to achieve this transparently to the researcher ?
5
Introduction Related Issues Related Issues How to leverage collaborative (interdisciplinary) work between geographically dispersed researchers ? What about Security ? Confidentiality, Integrity, Availability
6
Introduction 1 st Option: Supercomputers 1 st Option: Supercomputers 5 th Fastest computer June 2010 High cost (IBM BlueGene/L: us$100m+) Complex admin & maintenance Limited scalability Rapid obsolescence
7
Introduction 2 nd Option: Computing Clusters 2 nd Option: Computing Clusters Homogeneous (HW/SW), high-performance servers, interconnected by a high speed local network (Fiber Optics, 10 Gb Ethernet): UNDER ONE ADMINISTRATION DOMAIN
8
Introduction 3 rd Option: Grid Computing 3 rd Option: Grid Computing Heterogeneous (HW/SW), low/medium-performance workstations (commodity / shared), interconnected by a low speed local/wide area network (Internet): ACROSS MULTIPLE ADMINISTRATION DOMAINS
9
Introduction 4 th (new) Option: Cloud Computing 4 th (new) Option: Cloud Computing Software as a Service (SaaS) + Hardware as a Service (IaaS) = Platform as a Service (PaaS) e.g., Google Mail running on Google Cluster, Google App Engine
10
EHEP ATLAS Research Group Research on Experimental High Energy Physics in collaboration with CERN Research on Experimental High Energy Physics in collaboration with CERN Requirements:. Local storage capacity Each dataset is ~30 MB. Local computing capacity Each analysis requires processing many datasets Solution:. Computational Cluster (Tier-3)
11
UAN Tier-3 Cluster Manufacturer: Dell Processors: 16 AMD Opteron QC 2.3 GHz RAM: 128 GB Storage: ~4 TB SPECint2006: ~144
12
Grid Colombia Research Project Construction of a countrywide computing grid using RENATA (~Internet2) Requirements:. Use resources (storage & computational) spread across several universities 11 universities at the beginning, 20+ now Solution:. Computational Grid
13
Grid Colombia Research Project Knowledge GridInformation GridData GridComputing Grid
14
Sistema de Gestión de la Investigación (SGI) Information system to manage research groups, projects and products Requirements:. No budget for hardware. Service Orientation. Scalability Solution:. Cloud Computing (PaaS) Amazon EC, Google App Engine
15
Bimler Cephalometric Analysis A technique devised by Dr. Hans Peter Bimler used in orthodontics for the treatment of malocclusion Requirements:. Interactive service. Online/Offline service. Zero installation Solution:. Service Oriented Architecture
16
Bimler Cephalometric Analysis
17
VIII Encuentro Grupos Investigación Sep 2010 Grupo oneGrid Gracias
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.