Cloud and Grid Computing

Leen Blom

Centric Research & Development

June 21st, 2013

Cloud and Grid Computing

Part 1: Setting the scene• What is Cloud Computing and Grid Computing• Differences and similarities• Using grids and clouds

Part 2: What happens today• Grid activities in The Netherlands, Europe and Romania• Privacy and Security in the Cloud• The big cloud service providers• Our vision on Cloud

What is Cloud and Grid?

Power grid metaphor for both terms• Utility computing

• John McCarthy (AI, Lisp) opined in the 1960s that "computation may someday be organized as a public utility."[link]

• Grid definition used nowadays (2002)• “What is the Grid? A Three Point Checklist”,[link] Ian Foster lists

these primary attributes: • Computing resources are not administered centrally.• Open standards are used.• Nontrivial quality of service is achieved.

• Cloud definition based on NIST (2011)• NIST Definition of Cloud Computing

• NIST: National Institute of Standards and Technology• Part of the U.S. Department of Commerce

• Origin unclear: cloud was used as a metaphor for the Internet and a standardized cloud-like shape was used to denote a network

Grid Computing

Grid Computing is concerned with • “Coordinated resource sharing and problem solving in dynamic, multi-institutional

virtual organizations.” [Foster 2002]

Focus on• The problem• Coordination of resources shared• Multi organizations involved

Tasks are “bigger” than resources needed• One task to many resources• Task split-up in parallel sub-tasks

Classes of Grid Computing• High Performance Computing (HPC)

• Using message passing interfaces (MPI)• Expensive

• Embarrassingly Parallel Computing (EPC)• Or High Throughput Computing (HTC)• Often Volunteer Computing like Boinc

Cloud Computing by NIST

Service Models

PublicPrivate

CommunityHybrid

Deployment Models

source: http://www.nist.gov/itl/cloud/index.cfm

Cloud Service Models

Source: https://www.e-education.psu.edu/cloudGIS/cloud_introduction

Capex

Opex

Why Cloud (and Grid)?

On-demand self-service • A consumer can unilaterally provision computing capabilities without

requiring human interaction with each service provider.

Broad network access• Capabilities are available over the network and accessed through standard

mechanisms

Resource pooling• The provider’s computing resources are pooled to serve multiple consumers • Customer generally has no control or knowledge over the exact location of

the provided resources

Rapid elasticity• Capabilities can be elastically provisioned and released, in some cases

automatically, often appear to be unlimited and can be appropriated in any quantity at any time.

Measured service • Cloud systems automatically control and optimize resource use by

leveraging a meteringSee also: http://arxiv.org/pdf/0901.0131

Cloud and Grid: similar concepts

On-demand self-service • A consumer can unilaterally provision computing capabilities without

requiring human interaction with each service provider.

Broad network access• Capabilities are available over the network and accessed through standard

mechanisms

Resource pooling• The provider’s computing resources are pooled to serve multiple consumers • Customer generally has no control or knowledge over the exact location of

the provided resources

Rapid elasticity• Capabilities can be elastically provisioned and released, in some cases

automatically, often appear to be unlimited and can be appropriated in any quantity at any time.

Measured service • Cloud systems automatically control and optimize resource use by

leveraging a metering

Grid Cloud

Grid Cloud

Grid Cloud

Grid Cloud

Grid Cloud

Public cloud only

Batch oriented

Capabilities must match

Depends

Commercial grids

Public cloud only

See also: http://arxiv.org/pdf/0901.0131

Cloud and Grid: different concepts

Grid• Focus on the task at hand• Need resources to complete• Scientific research• Might use dedicated hardware components

• Like GPU or storage• Virtualization of jobs, not of programs

• Batch oriented

Cloud• Focus on servicing multiple users (multi-tenant)

• Information workers in many cases• Virtualization at several levels

• Infrastructure, Platform and Software• Userinterfaces like Windows

Cloud fundamentals

Virtualization

Up- and downscaling

From Capex to Opex• Prevents entry costs• Easier to downscale

Increasing utilization• Research of IBM/VMware

Cloud And Grid Computing

Cloud Computing is being integrated into Grids

Big Data, BigTable, Hadoop needs Grid techniques• Shift from userinterface to “batchjobs”

Rendering software in the Cloud• Offered as SaaS, • Autodesk 360• vSwarm (rendering for free)• Underlying principles are Grid

Weather services• Offered as SaaS to end users• Meteo organizations• Eurogrid.org

Financial services

New exciting cloud service provider!

Cloud: Privacy and Security

What is more secure: Private Cloud or Public Cloud?• Private

• Pro: easier to isolate• Con: need to keep knowledge up-to-date

• Public• Pro: may expect top security measures• Con: dependent on provider

What would respect privacy more: Private Cloud or Public Cloud?• Private

• Pro: no intrusion of people outside • Con: need to keep knowledge up-to-date

• Public• Pro: easier to uniform policies• Con: conflicting laws and jurisdiction in some cases

more than 77 million accounts affected, 12 million had

unencrypted credit card numbers

National security electronic surveillance program operated by

the United States National Security Agency (NSA) since 2007

Part 2: Cloud and Grid today

Grid examples

Old example: SETI@Home launched 1999• Goal is to detect intelligent life outside Earth• Uses Boinc: Open-source software for volunteer computing and grid computing

Life science Grid• Contains a limited number of bioinformatics applications • Free access for all

BiG Grid / SurfSara (NL)• National e-Infrastructure for research, SurfSara

• Huygens: the national supercomputer / Lisa: the national compute cluster / GPU cluster: a cluster equipped with GPUs / Grid: the Grid infrastructure / HPC Cloud: the cloud computing infrastructure / Hadoop: Big Data analytics framework / DIS: Data Ingest service / Collaboratorium: advanced visualization and presentation space / Visualization: high-end render cluster

European Grid Infrastructure• For scientists and researchers

Grid examples

Cartesius is the fastest supercomputer in the Netherlands and eventually will be able to perform

more than one quadrillion computations per second (1 Petaflop/s).

Grid examples

Rotterdam University of Applied Science (HR)• Donates their spare capacity of 6000 desktop to Erasmus Medical

Center. Usage• DNA-sequence analysis• Simulating spread of infectious deseases• MRI-scan analysis (Alzheimer)

• Technical• HTCondor™ middleware• Specialized workload management system for compute-intensive jobs• Open Source provides by University of Wisconsin Madison

• HR research by their students (Bachelor)• Fractal movie and rendering• Security (desk cracker)• Calculating prime numbers

Cloud examples

Numerous!

The Top Cloud Companies

1. Amazon.com• Inventor of IaaS

2. VMware• Used as cloud enabler• Provides: vCloud

3. Microsoft• Iaas, Paas: Azure• SaaS: enterprise products,

Office 365

4. Salesforce.com• SaaS pur sang

5. Google• SaaS: Apps, also for the

enterprise• PaaS: App Engine, Chrome

OS, IaaS: Storage (Drive)

6. Rackspace• IaaS pur sang, with NASA

inventers of OpenStack

7. IBM• All-in for OpenStack• Public: smart clouds , Private:

enterprise customers

8. Citrix• CloudStack as competitor for

OpenStack

Source: http://www.businessinsider.com/10-most-important-in-cloud-computing-2013-4?op=1

Questions?

Thank you very much for coming!

Leen Blom

Centric

leen.blom@centric.eu