Transcript A Computer Science perspective on Vision and strategy for tomorrow's challenges CITRIS/INRIA joint workshop David E.

A Computer Science perspective
on
Vision and strategy for tomorrow's
challenges
CITRIS/INRIA joint workshop
David E. Culler
University of California, Berkeley
May 23, 2011
…at the dawn of a new Age
Grad
Window
2
… of integration across vast scale
Computers
Per Person
1:106
Mainframe
Mini
1:103
Workstation
PC
Laptop
1:1
PDA
Cell
103:1
years
Mote!
Bell’s Law: new computer class per 10 years
3
… in a changing World
ARPANet
2.0 B 1/26/11
WWW
1969 1974
HTTP 0.9
RFC 675 TCP/IP
Internet
1990
2010
4
A different “Graduation Window”
Global temperature change (relative to pre-industrial era)
0°
1°
2°
3°
4°
C
C
C
C
C
Food
5°
C
Crop yields fall
Water
Glaciers melt
Ecosystems
Reefs damaged
Water shortages
Rising seas
Species extinction
Weather
Storms, droughts, fires, heat waves
Feedback
Abrupt climate change
Today
5
Growth Paths
Arch
OS
Graph
Sci
Theory
Net
AI
HCI
70
80
DB
Sec
90
PL
00
10
20
6
Operating Systems
UNIX
OS/360
70
MCP TOPS
Multics
BSD
VM/370
RSX-11
MVS
X11
80
Linux
Posix
90
00
10
20
VAX/VMS
OSx
wince
Symbian
Windows
NT
MS-DOS
XP
MacOS
7
Security and Privacy
70
Multics
80
90
00
10
20
8
Database and Information Mgmt
 Driven by the data and $s
Real-time
Sensors, events
unstructured
documents
Integration
Technology
Structured
data
70
80
Sybase Oracle
Informix
90
00
10
20
SQL Server
9
System innovation perspective
 Pace and form of innovation driven by emergence of
computer classes
 70’s shared server
 80’s personal, networked, workstation, SMP & MPP
 90’s cluster, 00’s internet service, data center
 Hugely effective research community turned inward
toward highly competitive conferences
 So far has missed the personal mobile revolution
 If it looks like a mid-80’s PC “Unix will run on it” and always
did
 Industry led the Cloud / Analytics revolution, but
research community running fast to catch up
10
Cyber-Physical Systems: A
Cooperative Grid
• Availability
• Pricing
• Planning
• Forecasting
Source
IPS
• Tracking
• Market
energy
subnet
Intelligent
Power Switch
Load IPS
• Monitor, Model, Mitigate
• Deep instrumentation
• Waste elimination
• Efficient Operation
• Shifting, Scheduling, Adaptation
11
Research as “Time Travel”
- the secret formula
 Imagine a technologically plausible future
 Create an approximation of that vision using
technology that exists.
 Discover what is True in that world
 Empirical experience
 Bashing your head, stubbing your toe, reaching epiphany
 Quantitative measurement and analysis
 Analytics and Foundations
 Courage to ‘break trail’ and discipline to do the
hard science on problems that matter
12
Theory
Arch
OS
Graph
Sci
Theory
Net
AI
HCI
70
80
DB
Sec
90
PL
00
10
20
OS/360
13
Computational Lens on the Sciences
Costis Daskalakis
Constantinos Daskalakis, Paul W. Goldberg and Christos H.
Papadimitriou, The Complexity of Computing a Nash Equilibrium,
In the 38th ACM Symposium on Theory of Computing, STOC
2006
70
80
90
00
10
20
14
Computational Theory of Social Change
15
Theory of Computing - Ahead
 Massive Data
 Understanding why heuristics work so well
 transition between efficient algorithms and computational
intractability
 New Foundations of Economics
 Nash Equilibrium intractable (2008) =>
 algorithmic game theory and mechanism design, including pricing
digital goods
 Biology
 Computational role of sex in reproduction
 Post-quantum Cryptography
 Quantum techniques in classical complexity theory
 lattice cryptosystems, local Hamiltonians, constraint satisfaction
problems.
 Statistical Machine Learning a first-class citizen
16
17