g-social - enhancing e-science tools with social networking functionality

Post on 03-Jul-2015

501 Views

Category:

Education

0 Downloads

Preview:

Click to see full reader

DESCRIPTION

Presentation of "g-Social - Enhancing e-Science Tools with Social Networking Functionality" given at the Workshop on Analyzing and Improving Collaborative eScience with Social Networks, Chicago October 8th, 2012. Co-located with IEEE eScience 2012.

TRANSCRIPT

g-SocialEnhancing e-Science Tools with Social

Networking Functionality

Andriani Stylianou, Nicholas Loulloudes, Marios D. Dikaiakos

Overview

• Introduction

• Motivation

• Problem

• Current Solutions

• g-Social – Our Solution

• Abstractions

• Implementation

• Conclusion - Questions

2

• Thousand years ago science was empirical– describing natural phenomena

• Last few hundred years: theoretical branch– using models, generalizations

• Last few decades: a computational branch– simulating complex phenomena

• Today: data exploration (eScience)– unify theory, experiment, and simulation

– Data captured by instrumentsOr generated by simulator

– Processed by software

– Information/Knowledge stored in computer

– Scientist analyzes database / filesusing data management and statistics

– “Computational X” and “X-Informatics”

Fourth Paradigm of Scientific Exploration (J. Gray)Source: J. Gray, talk to NRC/CSTB, “eScience - A Transformed Scientific Method.” Mountain View CA, 11 January 2007.

2009

3

Jim Gray

Manager of Microsoft Research's eScience Group.

1998 ACM Turing Award

The disappearance of Tenacious (28/1/2007)

4

FarallonIslands

The search for Tenacious (28/1/07 - 16/2/07)

• Night of 28/1: the USCG launched an airborne and seaborne SAR operation for Tenacious

– The SAR lasted for nearly two weeks - no signs found

• 31/1: the scientific community mobilized to help the SAR mission using online tools

– Computer scientists, oceanographers, engineers, volunteers, and Silicon Valley power players [NASA’s JPL, Amazon, Microsoft, Oracle, US Navy, Monterey Bay Aquarium Research

Institute, SDSC, Cornell Theory Center, Purdue, UWisc, Singular, Canadian Space Agency, Digital Globe.]

• A blog was setup to coordinate efforts and share ideas.Main foci of the effort were:

– Map the trajectory that Tenacious might have followed, in case Jim Gray lost control of the boat - to help guide the SAR operation

– Discover clues about Tenacious presence at sea

– Map the trajectories of large vessels traveling in the area, that may have collided with Tenacious

US/CG scoured 132,000 sq. miles of ocean5

Drift modeling

6

The search for Tenacious: online version

An exemplary e-Science application scenario• A multidisciplinary virtual organization of people with a common goal

– Scientists, engineers, managers, officials, volunteers

• A variety of algorithms and software tools:

– Ocean-current models and simulators, image processing & recognition, cellphone signal tracking and triangulation, data-format transformation, data cleansing, satellite collection planning, data mining, image geo-referencing

• A deluge of data (hundreds of GBs) retrieved over the net from various sources, requiring processing and fusion to extract knowledge

– Satellite orbits, satellite imagery at different resolutions, multispectral datasets, Web Databases, radio buoy and airborne sensors, HF radars, data about offshore currents, Web cameras

• A federation of computing, networking and service infrastructures

– Grids, clusters, storage devices, crowd-sourcing services7

Computing Grids• e-Science motivated the development of Grid technologies and

Federated Computing Infrastructures during the last decade.

• The Grid vision by Foster, Kesselman, Tuecke [Grid 1.0]:

– Distributed computing infrastructures that enableflexible, secure, coordinated resource sharing among dynamic collections ofindividuals and institutions

– Enable communities ( “ Virtual Organizations ” ) to share geographicallydistributed resources as they pursue common goals, in the absence of:Homogeneity, Central location, Central control, Existing trust relationships

• The hype following the Grid:

– One of the sources of the impact of scientific and technological changes onthe economy and society [Jeremy Rifkin, “The European Dream,” Penguin2004]

– The Grid has been described as the Next Generation Internet, theimplementation of the Global Computer etc.

8

Grid Infrastructure development

‣ Nowadays, Grid infrastructures comprise an impressivecollection of computational and software resources‣ drawing an increasing number of users from various disciplines

9

MotivationData-Intensive Scientific Projects

Resources

Grid / Cloud Computing

Traditional Collaboration Tools

Scientists

10

Problem

• Collaboration is done externally to scientificsoftware environments(email, web, portals, IM, etc.).

• Manual effort for transferring informationfrom one tool to another.

• Error prone and time consuming.

Lack of a unified, user-friendly software and collaboration environment for scientists.

11

Current Solutions

General-Purpose OSN

Pros• Professional Networking• Minimal Collaboration FunctionalityCons• External to existing scientific software

environments – Web Based• Do not support resource* sharing

Scientific OSN

Pros• More immersive collaboration environment

than Generic OSN.• Resource sharing and ability to run

experiments.Cons• Application Domain Specific.• Proprietary infrastructures – High

maintenance.• Introduce additional information sources ->

User Information overload 13

Our Solution

g-Eclipse (www.eclipse.org/geclipse)• Integrated workbench framework• Build on-top of Eclipse (Extensible and community support)• Toolset for users, operators & developers of Grid/Cloud infrastructures

(gLite, GRIA, Amazon AWS) – Middleware agnostic• Rich functionality:

• Development & Deployment• Benchmarking & Testing• Workflow Programming

Online Social Networks• Easy establishment and management of groups• Automatic dissemination of notifications• Professional Networking• High Availability

14

15

Information View

Grid Project View

Authentication View JSDL Editor View

Workbench

g-Eclipse

g-Social

Build on-top of the g-Eclipse FrameworkAims to enable collaboration among scientists that are/will utilize g-Eclipse

Features• Social Abstractions (Resources, Meta-data, Authentication).

• Definition of structured and standardized social meta-data

• Enrich social meta-data with links to project related resources.

• Access resources easily .

• Share project data and meta-data.

• Retrieve shared information.

• Seamless interaction with OSN.

• Facebook

• Twitter

• Extensible for other OSNs

g-Social Work Cycle 16

g-Social Abstractions

Enable seamless sharing and retrieval (via an OSN) of all particulars of theresearch work performed in the context of a real scientific project.

Abstract a Scientific Collaborative Environment which utilize Online SocialNetworks.

17

Abstractions - Resources

Any file(s) related to the execution ofa Grid task specific to a scientificproject

• Input / Output Dataset

• Executable

• Source Code

• Documentation

• Publications

• …

18

Abstractions – Social Meta-data

Descriptive meta-data that provide tothe OSN and its users informationabout purpose and function of eachshared particular

• Name

• Function

• Purpose

• Version

• Tags

• License

• ….

19

Abstractions – Authentication Manager

Enforces security and privacy controlof users while interacting with theOSN

• Authorization / Authenticationagainst an OSN

• Monitor life-cycle of authenticationtokens

20

Abstractions – Resource Manager

Resource sharing• Interact with Authentication Manager• Social meta-data• Encapsulate the above in a form

acceptable by and OSN

Resource Retrieval• Extraction of published meta-data• g-Eclipse Authentication Manager

invocation• Resource access via g-Eclipse file

system• Resource import in g-Eclipse workspace

21

Abstractions – OSN Interface

• OSN are by design web-basedsystems

• OSN-gEclipse interface serves as anintermediate between the web-browser and g-Eclipse.

• Invoking g-Eclipse when user clickson an g-Social link inside an OSN.

22

g-Social Implementation

• The g-Eclipse Grid Project.

• A placeholder for the organization of files/information related to the execution of Grid/Cloud tasks

• Executables (local file system)

• Input / Output dataset (g-Lite, AWS)

• Documentation

• Publication (IEEE, ACM, Elsevier)

• Infrastructure Configurations

23

Implementation (Social Meta-Data Editor)

• Multi-Page GUI Editor• Easy Insertion of social

meta-data• Specify Location of

Resources

• XML content meta-data• Extend Job Submission Definition

Language (JSDL) schema to includesocial meta-data specification.

24

Share Job

View Job Details

OSN AuthenticationSearch for Shared JobsCollaborators

List of Shared Jobs

25

g-Social View

Authorization• Authenticate / Authorize

against OSN• Check auth of the underlying

storage infrastructure whenlinking or retrieving aresource

• Manage auth tokens life-cycle

26

Implementation (g-Social View)

Share Job to OSN• Share job details as defined

in meta-data editor• Ask user to which OSN

details should be posted• Parse social meta-data• Encapsulate them in OSN

specific post formats.

27

Implementation (g-Social View)

Implementation (g-Social View)

View Share Job Details• Social Meta-data

• Name• Description• Version

• Resource Handles• Download Resource

28

Conclusions & Future Work

Future Work• Standardize social meta-data definition• Support additional OSNs• Recommendation System• Release g-Social to Eclipse

29

Conclusionsg-Social enhances integrated e-Science Tools (g-Eclipse) withSocial Networking functionality. Specifically it:• Enables the definition of social meta-data for sharing and

retrieval of information among scientists.• Enriches meta-data with resource handles which might be

scattered in heterogeneous storage infrastructures.• Provides mechanisms for sharing and retrieving scientific

information with just a few clicks.

Questions – Contact Information

Andriani Stylianou (andriani.stylianou@epfl.ch)

Nicholas Loulloudes (loulloudes.n@cs.ucy.ac.cy)

Marios D. Dikaiakos (mdd@cs.ucy.ac.cy)

http://grid.ucy.ac.cy

30

top related