An Introduction to Cloud-based Services

0
0
1895 days ago, 748 views
PowerPoint PPT Presentation

Presentation Transcript

Slide 1

An Introduction to Cloud-based Services Paul Watson Newcastle University, UK paul.watson@ncl.ac.uk

Slide 2

e.g. Amazon

Slide 3

Plan What is Cloud Computing? Potential Advantages Lessons from our own encounters Cloud Issues

Slide 4

What is Cloud Computing? ".. an expansive cluster of online administrations went for permitting clients to get an extensive variety of utilitarian abilities on a 'pay-as-you-go' premise that already required gigantic equipment/programming speculations and expert aptitudes to get." Irving Wladawsky Berger

Slide 5

What's New? deception of Infinite registering assets On Demand no in advance duty by clients Pay for utilization of assets on a fleeting premise as required (from "Over the Clouds: A Berkeley View of Cloud Computing")

Slide 6

Example – Amazon Web Services Based on Xen VMs run any OS & programming stack CPU: 1.0Ghz x86 occurrence @ $0.10/hour Blob Storage @ $0.12/GB month External Data Transfer @ $0.10/GB Also line, key store, piece store, scope of cases

Slide 7

Why is this Important (I): Internal IT Problems (slide by authorization of Arjuna Technologies) Silos = Inflexibility

Slide 8

Why is this Important (II)? Time to place Ideas without hesitation Research Have smart thought Write proposition Wait 6 months If effective.. Purchase Computers Install Computers Start Work Science Start-ups Have smart thought Write Business Plan Ask VCs to finance If effective.. Purchase PCs Install Computers Start Work

Slide 9

Why is this a Good thought: utilizing business mists Have smart thought Grab hubs as required from Cloud supplier Start Work Pay for what you utilized

Slide 10

Cloud Services Continuum (in view of Robert Anderson) http://et.cairene.net/2008/07/03/cloud-administrations continuum/Software ( SaaS ) Google Docs Salesforce.com Platform ( PaaS ) Flexibility Complexity Google AppEngine Microsoft Azure Infrastructure ( IaaS ) Amazon EC2 & S3

Slide 11

Example Lessons from CARMEN Project Design started in 2006 Commercial mists impossible Designed claim "private" cloud Experimenting with Commercial Cloud

Slide 12

UK EPSRC e-Science Pilot £4M (2006-10) 20 Investigators CARMEN Project Stirling St. Andrews Newcastle York Manchester Sheffield Leicester Cambridge Warwick Imperial Plymouth

Slide 13

Industry & Associates

Slide 14

Research Challenge Understanding the mind is the best informatics challenge Enormous ramifications for science: Medicine Biology Computer Science

Slide 15

Collecting the Evidence 100,000 neuroscientists produce tremendous amounts of information atomic (genomic/proteomic) neurophysiological (time-arrangement action) anatomical (spatial) behavioral

Slide 16

Epilepsy Exemplar Data examination guides specialist amid operation Further investigation gives confirm WARNING! The following 2 Slides demonstrate an uncovered human cerebrum

Slide 17

empowers sharing and collective abuse of information, examination code and aptitude that are not physically gathered CARMEN

Slide 18

CARMEN e-Science Requirements Store expansive amounts of information (100TB+) Analyze suite of neuroinformatics administrations bolster information escalated investigation Automate work process Share under client control

Slide 19

Background: North East Regional e-Science Center 25 Research Projects crosswise over numerous areas: Bioinformatics, Aging & Health, Neuroscience, Chemical Engineering, Transport, Geomatics, Video Archives, Artistic Performance Analysis, Computer Performance Analysis,.... Same key needs:

Slide 20

Result: e-Science Central Integrated Store-Analyze-Automate-Share foundation Generic CARMEN neuroinformatics & science as pilots

Slide 21

e-Science Central Web based Works anyplace e-Science Central Dynamic Resource Allocation Pay-as-you-Go* Controlled Sharing Collaboration Communities

Slide 22

Science Cloud Architecture Access over Internet (ordinarily by means of program ) Upload information & administrations Run examinations Data stockpiling and investigation

Slide 23

 Science Cloud Options U sers Science App 1 Science App n Service Developers .... Science Platform Science App 1 Science App n .... Cloud Infrastructure: Storage & Compute Cloud Infrastructure: Storage & Compute

Slide 24

.... Application App API e-Science Central Security Analysis Services Social Networking Science Cloud Platform Workflow Enactment Processing Cloud Infrastructure Storage

Slide 25

Editing and Running a Workflow on the Web

Slide 26

Workflow Result File Viewing the yield of Workflow Runs

Slide 27

Viewing comes about

Slide 28

Blogs and connections Communicating Results Linking to comes about & work processes

Slide 29

What we learnt: Moving into a Cloud Moving existing advancements into a cloud can be troublesome some can't keep running in a Cloud by any means

Slide 30

Raw Data Exploration with Signal Data Explorer

Slide 31

What we learnt : Scalability Clouds offer the potential for versatility snatch figure influence just when required Developers need to oversee adaptability for Infrastructure as a Service Clouds scale up and down

Slide 32

Adaptive Dynamic Deployment with Dynasoar Commercial "pay-as-you-go" mists would permit us to evade this utmost Adding Processors as you need them advances assets and spares cash in pay-as-you-go mists Ensure framework can likewise discharge undesirable hubs

Slide 33

Microsoft Azure Cloud for e-Science Demo Recent examinations with Microsoft Azure Cloud running Chemical investigations Silverlight App Thanks to: - Paul Appleby & Team at the Microsoft Technology Center, Reading - & MS External Research e-Science Group

Slide 35

Microsoft Azure Cloud Demo

Slide 36

When not to utilize Clouds? Huge information exchanges Time & Cost High Performance cpu/io/arrange data transmission/low idleness Predictable Performance Confidentiality High Availability? High Server Utilization? private mists better?

Slide 37

Create Private Cloud (slides by consent of Arjuna Technologies)

Slide 38

Private Cloud Examples Eucalyptus Amazon API Private Cloud arrangements of Microsoft Azure Arjuna Agility

Slide 39

Federating Private & Public Clouds Public Cloud Public Cloud e.g. Amazon App1 Service Agreement Arjuna Agility App1 & 2 Service Agreement Internal Cloud Dept A Dept B

Slide 40

Public Cloud e.g. Amazon App1 Public Cloud e.g. FlexiScale Arjuna Agility App1 & 2 Internal Cloud Dept A Dept B Arjuna

Slide 41

Summary Cloud registering can change e-science give manageable framework lessen time from thought to acknowledgment Don't think little of unpredictability building adaptable disseminated frameworks is still hard can Science Clouds help by bringing down the obstacles? e-Science Central Store-Analyze-Automate-Share e-science stage including content from a scope of spaces CARMEN is assessing it for neuroinformatics

SPONSORS