computing for the masses 为人民计算 zhiwei xu 徐志伟 information science advisory committee,...

27
Computing for the Masses 为为为为为 Zhiwei Xu 为为为 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences [email protected] INSTITUTE OF COMPUTING TECHNOLOGY

Upload: shanna-reeves

Post on 23-Dec-2015

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Computing for the Masses为人民计算

Zhiwei Xu 徐志伟Information Science Advisory Committee, NSFC

Institute of Computing Technology (ICT)Chinese Academy of Sciences

[email protected]

INSTITUTE OF COMPUTING

TECHNOLOGY

Page 2: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Contents

• Background• Goals• Problems and Approaches

Page 3: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Demand: China Computer Market Grows

GDP US$Trillio

n

ComputerMarket

US$Billion

InternetUsers

(Million)

ClientDevices(Million)

1995 0.69 7.4

2000 1.08 25.9 22.5 8.9

2005 2.30 59.0 111(+80) 49.5

2010 3.00 115.6 233 106

2015 4.75 217.3 411 191

2020 7.07 403.9 662 308

Sources: China NBS, CCID, CNNIC, Goldman Sachs

2007 3.38 93.1 210 78

Still has big growth space

Internet Penetration (2007)

China 16%

USA 70%

World 19%

China in 10 years = USA today?

Page 4: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Supply: China computer industry is weak(Forbes 2000 for year 2007 US$ billion)

Rank Company Country Sales Profi ts Assets Market ValueProfi t/SalesP/EAppl i cati on Servi ces

709 Ebay USA 7. 67 0. 35 15. 37 35 4. 56% 100740 Amazon USA 14. 84 0. 48 6. 49 26. 87 3. 23% 561565 Expedi a USA 2. 67 0. 3 8. 3 6. 55 11. 24% 221863 Al ibaba China 0.17 0.03 0.23 12.25 17.65% 408

Software & Servi ces37 I BM USA 98. 79 10. 42 120. 43 157. 62 10. 55% 1563 Mi crosof t USA 57. 9 16. 96 67. 34 253. 15 29. 29% 15213 Googl e USA 16. 59 4. 2 25. 34 147. 66 25. 32% 35319 SAP Germany 14. 96 2. 81 14. 93 57. 77 18. 78% 211040 I nfosys I ndi a 3. 21 0. 89 3. 08 22. 09 27. 73% 251905 Tencent China 0.36 0.14 0.58 11.37 38.89% 81

Hardware53 HP USA 107. 67 7. 85 88. 57 122. 04 7. 29% 16178 Appl e USA 26. 5 4. 07 30. 04 109. 88 15. 36% 27192 Del l USA 61. 13 2. 95 27. 56 44. 6 4. 83% 151053 RI M Canada 2. 95 0. 61 3. 08 58. 73 20. 68% 961338 Lenovo China 14.53 0.16 5.35 6.47 1.10% 40

Page 5: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Challenges to Academia

• C4M supply is seriously lacking

• Lag behind demandLag behind industry

• Lag behind international peersToo much short-term “mission”

Page 6: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Computing for the Masses (C4M)

• Research and applications of computer science for mass adoption– Directly benefit the masses

• Billions of people

≠ scientific computing or business computing

– Including• Parallelism for the masses• Net computing for the masses• Social computing for the masses

Page 7: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Contents

• Background• Goals

– Mass Adoption billions of people• No dumbing down: Value = Ω(Adoption)

– Sustainability: Value↑, resource→ • Problems and Approaches

Page 8: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Sustainability

1960 1970 1980 1990 2000 2010 2020 2030 2040Time

Value &Resource Total IT Value

Resource consumptionEnvironment impact

Servers electricity bill: $1.9 billion (IDC, 2007 China )

Physics

Circuits

Systems

Energy/operation(2000-2007)

μJ

pJ

Page 9: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

C4M is not dumbing down

Personalized

Expertise

Ubiquity

Commodity

Basic

0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 1.2Adoption(Computer Users in China, Billion)

Value

2010

2020 2030 2040

Page 10: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Value-Augmenting Adoption

Personalized

Expertise

Ubiquity

Commodity

Basic

0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 1.2Adoption(Computer Users in China, Billion)

Value

2010

2020

2030

2040

PC: example of C4M• Reached more users• More value than mainframe/mini• More innovations with big ideas

– Frame buffer– GUI– OO programming– Ethernet

Page 11: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

What’s different now?

• The Net = Three Worlds• Man-Machine Symbiosis Man-Machine Society

Page 12: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Computing for the Masses

Business Infrastructure

IT Infrastructure

Business Value

IT Components

Business App & Svc

Servers, NetworksStorages, Data

Sensor Networks, CPS

Utility Devices

Computer users in China: 210 million (2007); 800 million (2020)

Godson CPULenovo PC/LaptopDawning Servers

BlueWhale StorageVega Grid

Page 13: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Contents

• Background• Goals• Problems and Approaches

– A Science of Three Worlds– Architectural Characteristics– Personal Net

Page 14: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

A Science of Net Computing

• Computation as a unifying theme in a new science of three worlds

• Enrich our beautiful algorithmic computing theory• Karp

– Computational lens

• Hurwicz– Mechanism design

• Science 2020

Page 15: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Google-Like Computing

Linux Servers and Other Resources Distributed in Wide Area

Google system softwareGoogle filesystem, resource mngt, fault tolerance, etc.

Google utilities: MapReduce & BigTable

Google application software and data Sorting, machine learning, graph computing, etc.

Google Value (Visible)Search, AdWord, Map, Earth, News, Froogle, etc.

Hosting Environment: LAMP

DatacentersServersData, MetadataCode

2006 Data ($ billion)

Revenue:10.6

Profit: 3

Cost: 7

R&D: 1.2

Resource: 2.4

O=F(I) Value = F(Resource)

Page 16: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Enrich Algorithmic Computing • Traditional algorithmic computing

– Turing machine decision problem– Input, output, a procedure of mechanic steps– Time complexity, space complexity

• What are “algorithms” in the tri-world?– What is the “decision problem”?

• What is Web computable?• What is Wiki computable?

– How to quantify “value” and “resource”?– What is a “step”? What is termination?– What are the complexity metrics?

Page 17: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

System Characteristics

• C4M Workload Analytics– Time, space, information– Interaction, energy, effort

• Basic “Laws” revisited– Moore’s law– Network effect (Metcalf, Brown)– Viral market– Internet principles (E2E, REST)

• New Phenomena and Abstractions

Page 18: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Architecture Characterization

• Patterson & HennesseyPerformance = Program/Time

= 1/(#Instructions x CPI x CycleTime)• Need reexamination

– Program=C4M workloads– Time=?– Instruction=?

• Other important metricsTask/Energy = ?Task/Effort = ?

Page 19: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Architecture Characterization

Execution Single Multiple

Control

Decentralized

Centralized

Admin, Knowledge, Naming, Coding, Contribution

Salesforce.com

many web sites

GoogleAmazonTeragrid

WWWCloudsPNC environment

Virtual hosts

Virtual Machines

Number of Execution Sites (Datacenters, Machines)

Distributed Systems

DecentralizedSystems

Page 20: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Personal Net (PN)

• A PN for each member of the masses• A general-purpose, personal, net computing

platform – a dynamic, virtualized set of assets from the Net

(cyberinfrastructure, community, physical world)– appearing to be dedicated to a personal owner’s

use and control• People share the Net personally

Page 21: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

The Net Now• Offering

– Traditional network services: email, ftp, BBS, messaging– Consumer web: Amazon, eBay– Business web: salesforce.com– Community Web: Wiki, MySpace, Facebook– Grid services: Nanohub– Platform: Teragrid, Amazon S3 and EC2 (clouds)

• Characteristics– Institutional, not personal– Special-purpose solutions, not a general-purpose platform

• The Net now is like the Mainframe in 1960’s

Page 22: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Personal Net Computing

ASP

PG Platform Provider

A, DO, PC, N, S

A, DO, PC, N, S

……

PG Platform Provider

A, DO, PC, N, S

A, DO, PC, N, S

……

DSP CSP SSP NSP ASP……

PersonalGrids

PlatformProviders

ResourceProviders

AccessingDevices

Individuals

The Net

Assets A: applications D: data O: operating sys. P: policies C: computing N: networking S: storage

SP: service provider for resources

Resources are “raw” assets (capabilities)

DISC Grids Clouds

PocketWeb

Pocket Web: battery life > 2 weeks; assets on demand; the Net in you hand

Page 23: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Dual-objective optimizationFairness

Success Rate

Characterizing Emergence

1

2

QQ

Qn

nR

RR

2

1

Page 24: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Emergencesdo appear

• Workloads (over 1 million jobs simulated)– Synthetic workloads– Real workloads: www.cs.huji.ac.il/labs/parallel/workload/

• Value For consumers: for providers:

high job success rate fair revenue and utilization

Xiao et al, “Incentive-based Scheduling for Market-like Computational Grids”, IEEE Transactions on Parallel and Distributed Systems, 2008

Page 25: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Clash of the Computer and the Network Approaches

• Fetching 10-byte data from a blog server: 162 ms, 52 context switchesat server side

• How much is needed to host 100 million PG’s?– Response time < 0.25 s

• Sustained = 5% Peak?

1 Network Access

2 Inter Network

3 Transport

4 Application

?

GSMLBPEL

WSRFWSDLSOAP

HTML XMLHTTP

TCP/IP Stack Web/Web Service Stacks

HW systemVMM

OS

DatabaseApp

HW (core)Thread VMM

AppThread

Page 26: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Summary

• CS research lags behind demand• C4M is potentially a transformative

opportunity for CS research • C4M means sustainability and value-

augmenting mass adoption• C4M research agenda

– Establish a science of three worlds– Characterize net computing architectures– Create personal net

Page 27: Computing for the Masses 为人民计算 Zhiwei Xu 徐志伟 Information Science Advisory Committee, NSFC Institute of Computing Technology (ICT) Chinese Academy of Sciences

Thank you!

[email protected]