Resource provisioning for e-Science environments
Abstract
Recent works have proposed a number of models and tools to address the
growing needs and expectations in the field of e-Science. In
particular, they have shown the advantages and the feasibility of
modeling e-Science environments and infrastructures according to the
Service-Oriented Architecture (SOA). At the same time, the
availability and models of use of networked computing resources needed
by e-Science are rapidly changing and see the coexistence of many
disparate paradigms: high performance computing, grid and recently
cloud, which brings very promising expectations due to its high
flexibility. Unfortunately, none of these paradigms is recognized as
the ultimate solution, and a convergence of all of them should be
pursued.
In this project we wish to test a model to promote the convergence and the
integration of different computing paradigms and infrastructures for
the dynamic on-demand provisioning of the resources needed by
e-Science environments, especially those developed according to
SOA. In addition, such a model aims at endorsing a flexible, modular,
workflow-based computing model for e-Science.
A working implementation used to validate the proposed approach will
be developed and tested using FutureGrid resources.
Intellectual Merit
Promote the convergence and the
integration of different computing paradigms and infrastructures for
the dynamic on-demand provisioning of the resources needed by
e-Science environments, especially those developed according to
SOA. Endorse a flexible, modular, workflow-based computing model for
e-Science.
Broader Impact
Promote the usage of shared research and education infrastructure,
including facilities and science and technology centers and
engineering research centers.
Contribute to computational models exploiting advanced computing
resources and standardized open software tools.
Develop activities that ensure that multi-user facilities are sites of
research and mentoring for large numbers of science and engineering
students.
Use of FutureGrid
A working implementation used to validate the proposed project will
be developed and tested using FutureGrid resources.
In particular, IaaS cloud resources will be needed to run "custom" VM
using Eucalyptus, Nimbus and OpenNebula/OpenStack (when available).
Scale Of Use
A few VMs (< 10) for a number of experiments, few hours of expected run time for
each VM.
Publications
- [fg-2603] Bosin, A., "Resource Provisioning for e-Science Environments", IGI Global, 31/2013
- [fg-2307] Bosin, A., "A SOA-based model for the integrated provisioning of cloud and grid resources",
- [fg-1923] Bosin, A., M. Dessalvi, G. M. Mereu, and G. Serra, "Enhancing Eucalyptus Community Cloud", 2012
Results
0. Environment inspection
Eucalyptus environment tested, identified a working set of
(image, kernel, ramdisk), inspection of a running VM instance
to extrapolate underlying configuration (virtual devices, kernel
and kernel modules) for subsequent custom image setup
1. VM setup
Setup and deployed custom VM images
- VM for publishing a Java web service: JRE and web service are
dynamically downloaded and executed immediately after boot; a
start-up script is in charge of downloading the web service
configuration from a public URL
- VM for publishing an Apache ODE workflow engine (deployed inside an
Apache Tomcat container) and a supervisor web service; a
start-up script is in charge of downloading the workflows (BPEL
processes) to be deployed into the engine; the supervisor service
is in charge of downloading workflow input from a public URL and
enacting one or more workflow instances
2. VM test
Deployed VM images have been manually instantiated to verify the correct
behavior of start-up scripts
3. Programmatic VM interaction
Programmatic VM instance creation and termination has been
successfully achieved through the EC2 APIs by means of the jclouds
library
Futuregrid is a resource provider for