Cloud Computing Challenges and Opportunities 17.7. 22.7. 2011 - - PowerPoint PPT Presentation

▶

Jan 29, 2023 165 likes •406 views

DEUTSCH-FRANZSISCHE SOMMERUNIVERSITT UNIVERSIT D T FRANCO-ALLEMANDE FR NACHWUCHSWISSENSCHAFTLER 2011 POUR JEUNES CHERCHEURS 2011 CLOUD COMPUTING : CLOUD COMPUTING : DFIS ET OPPORTUNITS HERAUSFORDERUNGEN UND

SLIDE 1

DEUTSCH-FRANZÖSISCHE SOMMERUNIVERSITÄT FÜR NACHWUCHSWISSENSCHAFTLER 2011 UNIVERSITÉ DʼÉTÉ FRANCO-ALLEMANDE  POUR JEUNES CHERCHEURS 2011

CLOUD COMPUTING : DÉFIS ET OPPORTUNITÉS CLOUD COMPUTING : HERAUSFORDERUNGEN UND MÖGLICHKEITEN

17.7. ¡– ¡22.7. ¡2011

Cloud Computing Challenges and Opportunities

Ernst Biersack

EURECOM

Sophia Antipolis, France

SLIDE 2

21 July 2011

Cloud Computing: Historical Perspective



Mainframes

 Centralized computing resources, dumb terminals

 “I think there is a world market for maybe five computers” T.J.

Watson, 1943



Client server model

 PCs and servers



Peer to Peer

 Everything happens in the end-systems: Ex.: Skype



Cloud Computing

SLIDE 3

21 July 2011

Cloud Computing: Historical Perspective

 1960s: ”Computation may someday be organized as a public

utility.” by John McCarthy

 “The Challenge of the Computer Utility by David Parkhill,

Addison-Wesley Publishing Company, (1966)

 2006: Amazon launched Amazon Web Service (AWS) to provide

cloud computing to external customers

 Early 2008: Eucalyptus became the first open-source, AWS

API-compatible platform for deploying private clouds.

SLIDE 4

21 July 2011

Cloud Computing: Definition

 “Cloud computing is a model for enabling ubiquitous, convenient,

n-demand network access to a shared pool of configurable

computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction”. (NIST Special Publication 800-145, Jan 2011)

SLIDE 5

21 July 2011

Cloud Computing: Characteristics



On-demand self-service. A consumer can unilaterally provision computing capabilities, such as server time and storage, as needed

 à

àNo CAPEX



Resource pooling. The provider’s computing resources are pooled to serve multiple consumers, with different physical and virtual resources dynamically assigned according to consumer demand

  Better resource utilization, savings for HW and SW due to economies of scale

(can purchase at 1/7 to 1/5 of the price)



Rapid elasticity. Capabilities can be rapidly and elastically provisioned



 Scalability



Measured Service. Resource usage can be monitored, controlled, and reported, providing transparency for both the provider and consumer of the utilized service.

  Pay-what-you use

SLIDE 6

21 July 2011

Cloud Computing: Service Models



Cloud Infrastructure as a Service (IaaS). The capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software

 Ex.: Amazon EC2, Rackspace, …



Cloud Platform as a Service (PaaS). Is the delivery of a computing platform and solution stack as a service that allows to deploy consumer- created or acquired applications

 Ex.: Google AppEngine, Microsoft Azure, …



Cloud Software as a Service (SaaS). The capability provided to the consumer is to use the provider’s applications running on a cloud infrastructure.

 Ex.: Google Apps, Dropbox, Evernote

SLIDE 7

21 July 2011

Cloud Computing: Deployment Models

 Public cloud. The cloud infrastructure is made available to the

general public or a large industry group and is owned by an

rganization selling cloud services.

 Private cloud. The cloud infrastructure is operated solely for an

rganization. It may be managed by the organization or a third

party and may exist on premise or off premise.

 Hybrid cloud. The cloud infrastructure is a composition of two or

more clouds (private, community, or public) that remain unique entities

SLIDE 8

21 July 2011

Implications: Emergence of a New Paradigm

Thousand years ago – Experimental Science

 Description of natural phenomena

Last few hundred years – Theoretical Science

 Newton’s Laws, Maxwell’s Equations…

Last few decades – Computational Science

 Simulation of complex phenomena

Today – Data-Intensive Science

 Huge data sets from a variety of different sources

 Data captured by instruments, sensor networks  Data generated by simulations  Data generated by computational models

From: Roger Barga

SLIDE 9

21 July 2011

Implications: Data-Intensive Scalable Computing

 But, what is all this data good for?

 The 6000 Wal‐Mart stores worldwide record every purchase by every

shopper, totaling around 267 million transactions per day. They collect all this information in a 4‐petabyte (that’s 4 X 10^15 bytes) data warehouse

 The proposed Large Synoptic Survey Telescope (LSST) will scan the sky

from a mountaintop in Chile, with what can be considered the world’s largest digital camera, generating a 3200 megapixel image every 15 seconds, covering the total visible sky every 3 days. That will yield 30 terabytes (10^12 bytes) of image data every day.

 CERN …..

(From Data-Intensive Scalable Computing Harnessing the Power of Cloud Computing, by Randal E. Bryant of CMU, February, 2009)

SLIDE 10

21 July 2011

Implications: Examples

 Don’t buy SW license, use a service provided by the cloud

SLIDE 11

21 July 2011

Implications: TimesMachine

 The New York Times Archives + Amazon Web Services =

TimesMachine

 Using Amazon Web Services, Hadoop and our own code, we ingested

405,000 very large TIFF images, 3.3 million articles in SGML and 405,000 XML files mapping articles to rectangular regions in the TIFF’s. This data was converted to a more web-friendly 810,000 PNG images (thumbnails and full images) and 405,000 JavaScript files — all of it ready to be assembled into a TimesMachine. By leveraging the power of AWS and Hadoop, we were able to utilize hundreds of machines concurrently and process all the data in less than 36 hours.

From http://open.blogs.nytimes.com/2008/05/21/the-new-york-times- archives-amazon-web-services-timesmachine/

SLIDE 12

21 July 2011

Consequences: Creative destruction

 Creative destruction after J. Schumpeter, describes the

process of transformation that accompanies radical innovation.

 Innovative entry by entrepreneurs is the force that sustains

long-term economic growth, even as it destroys the value of established companies that enjoyed some degree

f monopoly power

 Ex.:

 IBM (mainframes)  DEC (microcomputers)  SUN (workstations)  Microsoft?  Nokia ?

SLIDE 13

21 July 2011

Cloud Computing: Risks

 Information security and users' privacy

Using a service of cloud computing to store data may expose the user to potential violation of privacy. Possession of a user's personal information is entrusted to a provider that can reside in a country

ther than the user's. In the case of a malicious behavior of the cloud

provider, it could access the data in order to perform market research and user profiling.

“We may disclose to parties outside Dropbox files stored in your

Dropbox and information about you that we collect when we have a good faith belief that disclosure is reasonably necessary to (a) comply with a

law, regulation or compulsory legal request; (b) protect the safety of any person from death or serious bodily injury; (c) prevent fraud or abuse of Dropbox or its users; or (d) to protect Dropbox’s property rights.“

SLIDE 14

21 July 2011

Cloud Computing: Risks

 Political and economic problems

 Crucial and intellectual productions and large amounts of personal

information are increasingly stored in private, centralized and partially accessible archives in the form of digital data. No guarantee is given to the users for a free future access. (Ex.: Scientific Publishing, Google’s digital library)

 Continuity of service

 Delegating their data-managing and processing to an external service,

users are severely limited when these services are not operating. A malfunction also affects a large number of users at once because these services are often shared on a large network.

SLIDE 15

21 July 2011

Cloud Computing: Risks

 Data migration problems (data lock-in)

 When a user wants to change his cloud provider. There is no defined

standard between the operators and such a change is extremely

complex. The case of bankruptcy of the company of the cloud

provider could be extremely dangerous for the users.

SLIDE 16

21 July 2011

Cloud Computing How: Data Center



Unprecedented economies of scale Approximate costs for a medium size center (1000 servers) and large, 50K server center.

Technology ¡ Cost ¡in ¡ Medium-‑sized ¡ Data ¡Center ¡ Cost ¡in ¡Very ¡Large ¡ Data ¡Center ¡ Ratio ¡

Network ¡

$95 ¡per ¡Mbps/ ¡ month ¡ $13 ¡per ¡Mbps/ ¡ month ¡

¡ ¡ ¡7:1 ¡ Storage ¡

$2.20 ¡per ¡GB/ ¡ month ¡ $0.40 ¡per ¡GB/ ¡ month ¡ ¡ ¡ ¡5:7 ¡

Administration ¡

~140 ¡servers/ ¡ Administrator ¡ >1000 ¡Servers/ ¡ Administrator ¡

¡ ¡ ¡7:1 ¡

James ¡Hamilton, ¡LADIS ¡‘08

¡ ¡

From: Roger Barga

SLIDE 17

21 July 2011

Cloud Computing How: Warehouse

scale computer

From “The Datacenter as a Computer: An Introduction to the Design

f Warehouse-Scale Machines”, Hoelzle & Barroso, 2009

 These new large datacenters are quite different from traditional

hosting facilities of earlier times and cannot be viewed simply as a collection of co-located servers.

 Large portions of the hardware and software resources in these

facilities must work in concert to efficiently deliver good levels

f Internet service performance, something that can only be

achieved by a holistic approach to their design and deployment. In other words, we must treat the datacenter itself as one massive warehouse-scale computer (WSC).

SLIDE 18

21 July 2011

Cloud Computing: How

 Hardware

 Multicore CPUs  Communication network  Virtualization

 Storage (data model)

 Block Storage  Key value store  Database

 SQL DB (MS Azure)

 Compute Model for parallel processing

 Map/Reduce

SLIDE 19

21 July 2011

Cloud Computing: Conclusion

 Many benefits  Many open issues

SLIDE 20

21 July 2011

Program

Monday, July 18th

Morning (9-12h): 1h: E. Biersack, G. Carle: Introduction to the seminar + Presentation of Speakers 2h: J. Jackson: Introduction to cloud computing : Motivation, driving forces, data center vs HPC, Cloud SW models, IAAS: Eucalyptus Afternoon (14-18h): 4 h: P. Senellart: Hadoop Part I (Distributed computation): Introduction, HDFS, Map reduce, Pig, Open issues Night (20-21h): Improvisational theatre: TATwort Munich

Tuesday, July 19th

Morning (9-12h): 3h: W.-T. Balke: Hadoop part II (Distributed storage): DHTs, Dynamo, HBASE or Bigtable, Open issues Afternoon (14-18h): Leisure trip: Visite of chateau "Herrenchiemsee" Evening: Poster session

Wednesday, July 20th

Morning (9-12h): 2h: J. Jackson: Microsoft Azure as a PAAS, Case Study (AzureBlast), Relationship to Hadoop, Open issues 1h: Poster session Afternoon (14-18h): 4h: P. Michiardi: Practical exercises on Hadoop: Map reduce and more Night (20-21h30): Zizal - Klezmer Music Concert

SLIDE 21

21 July 2011

Thursday, July 21th

Morning (9-12h): 3h: V. Markl: Beyond Map Reduce Afternoon (14-18h): 2h: Roundtable (W.-T. Balke , J. Jackson, V. Markl, P. Senellart): Research issues in cloud computing 2h: T. Fetzer: Cloud computing and privacy laws Evening: Sport and Co.

Friday, July 22th

Morning (9-12h): 3h: H. Heller, N. O. vor dem gentschen Felde: Cloud computing and HPC: Security, Virtualization, Energy efficiency Afternoon (14-18h): 4h: H. Kosch, D. Coquil: The German-French Doctoral College on Multimedia, Distributed, Pervasive and Secure Systems (MDPS): scientific issues and structure of a bi-national PhD-program

Program

SLIDE 22

DEUTSCH-FRANZÖSISCHE SOMMERUNIVERSITÄT FÜR NACHWUCHSWISSENSCHAFTLER 2011 UNIVERSITÉ DʼÉTÉ FRANCO-ALLEMANDE  POUR JEUNES CHERCHEURS 2011

17.7. ¡– ¡22.7. ¡2011

Cloud Computing Challenges and Opportunities

Ernst Biersack

EURECOM

Sophia Antipolis, France

Cloud Computing: Historical Perspective

Mainframes

Watson, 1943

Client server model

Peer to Peer

Cloud Computing

Cloud Computing: Historical Perspective

utility.” by John McCarthy

Addison-Wesley Publishing Company, (1966)

cloud computing to external customers

API-compatible platform for deploying private clouds.

Cloud Computing: Definition

computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction”. (NIST Special Publication 800-145, Jan 2011)

Cloud Computing: Characteristics

On-demand self-service. A consumer can unilaterally provision computing capabilities, such as server time and storage, as needed

Resource pooling. The provider’s computing resources are pooled to serve multiple consumers, with different physical and virtual resources dynamically assigned according to consumer demand

Rapid elasticity. Capabilities can be rapidly and elastically provisioned

Measured Service. Resource usage can be monitored, controlled, and reported, providing transparency for both the provider and consumer of the utilized service.

Cloud Computing: Service Models

Cloud Infrastructure as a Service (IaaS). The capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software

Cloud Platform as a Service (PaaS). Is the delivery of a computing platform and solution stack as a service that allows to deploy consumer- created or acquired applications

Cloud Software as a Service (SaaS). The capability provided to the consumer is to use the provider’s applications running on a cloud infrastructure.

Cloud Computing: Deployment Models

general public or a large industry group and is owned by an

party and may exist on premise or off premise.

more clouds (private, community, or public) that remain unique entities

Implications: Emergence of a New Paradigm

Today – Data-Intensive Science

Implications: Data-Intensive Scalable Computing

shopper, totaling around 267 million transactions per day. They collect all this information in a 4‐petabyte (that’s 4 X 10^15 bytes) data warehouse

from a mountaintop in Chile, with what can be considered the world’s largest digital camera, generating a 3200 megapixel image every 15 seconds, covering the total visible sky every 3 days. That will yield 30 terabytes (10^12 bytes) of image data every day.

Implications: Examples

Implications: TimesMachine

TimesMachine

Consequences: Creative destruction

process of transformation that accompanies radical innovation.

long-term economic growth, even as it destroys the value of established companies that enjoyed some degree

Cloud Computing: Risks

Using a service of cloud computing to store data may expose the user to potential violation of privacy. Possession of a user's personal information is entrusted to a provider that can reside in a country

provider, it could access the data in order to perform market research and user profiling.

Dropbox and information about you that we collect when we have a good faith belief that disclosure is reasonably necessary to (a) comply with a

Cloud Computing: Risks

information are increasingly stored in private, centralized and partially accessible archives in the form of digital data. No guarantee is given to the users for a free future access. (Ex.: Scientific Publishing, Google’s digital library)

users are severely limited when these services are not operating. A malfunction also affects a large number of users at once because these services are often shared on a large network.

Cloud Computing: Risks

standard between the operators and such a change is extremely

provider could be extremely dangerous for the users.

Cloud Computing How: Data Center

Unprecedented economies of scale Approximate costs for a medium size center (1000 servers) and large, 50K server center.

Cloud Computing How: Warehouse

scale computer

hosting facilities of earlier times and cannot be viewed simply as a collection of co-located servers.

facilities must work in concert to efficiently deliver good levels

achieved by a holistic approach to their design and deployment. In other words, we must treat the datacenter itself as one massive warehouse-scale computer (WSC).

Cloud Computing: How

Cloud Computing: Conclusion

Program

Program

17.7. ¡– ¡22.7. ¡2011

Thanks