Universal laws and architecture 2:
Theoretical foundations for complex networks relevant to biology, medicine, and
neuroscience?
John Doyle
John G Braun Professor
Control and Dynamical Systems, EE, BE
Caltech
Thanks again
Outline
• Recap robustness, neuro motivation
• Bowties: metabolism, reward
• Layered architectures – Large, thin, nonconvex
– Clothing as layered architecture
– Operating systems and Internet
Lectures
1) Concrete motivation
2-3) Universal laws and architectures*
4) A teensy bit of math
*have you ever heard of anything more pretentious?
This paper aims to bridge progress in neuroscience involving
sophisticated quantitative analysis of behavior, including the use
of robust control, with other relevant conceptual and theoretical
frameworks from systems engineering, systems biology, and
mathematics.
Doyle and Csete, Proc Nat Acad Sci USA, JULY 25 2011
Very accessible No math
Lectures
1) Concrete motivation (recap)
2-3) Universal laws and architectures*
4) A teensy bit of math
*have you ever heard of anything more pretentious?
Robust Yet Fragile
Human complexity
Metabolism
Regeneration & repair
Immune/inflammation
Microbe symbionts
Neuro-endocrine
Complex societies
Advanced technologies
Risk “management”
Obesity, diabetes
Cancer
AutoImmune/Inflame
Parasites, infection
Addiction, psychosis,…
Epidemics, war,…
Disasters, global &!%$#
Obfuscate, amplify,…
Accident or necessity?
Robust Yet Fragile
Human complexity
Metabolism
Regeneration & repair
Immune/inflammation
Microbe symbionts
Neuro-endocrine
Complex societies
Advanced technologies
Risk “management”
Obesity, diabetes
Cancer
AutoImmune/Inflame
Parasites, infection
Addiction, psychosis,…
Epidemics, war,…
Disasters, global &!%$#
Obfuscate, amplify,…
Accident or necessity?
Fat accumulation
Insulin resistance
Proliferation
Inflammation
Robust Fragile Metabolism
Regeneration & repair
Healing wound /infect
Obesity, diabetes
Cancer
AutoImmune/Inflame Fat accumulation
Insulin resistance
Proliferation
Inflammation
• Fragility Hijacking, side effects, unintended…
• Of mechanisms evolved for robustness
• Complexity control, robust/fragile tradeoffs
• Math: robust/fragile constraints (“conservation laws”)
Accident or necessity?
Both
xor
Robust
Modular
Simple
Plastic
Evolvable
Fragile
Distributed
Complex
Frozen
Frozen
and
tradeoffs
Meta-layers
Fast,
Limited
scope
Slow,
Broad
scope
Which blue line is longer?
“Seeing is dreaming?”
“Seeing is believing?”
Fast
Inflexible
delay=
death Reflex
Vestibulo-
ocular
reflex
Act
Ac
Fast
Slow
Sense Act
Sense
Move
head
Move
hand
Slow
Fast?
Same actuators
Delay is limiting
sense
move Spine
Reflex
Reflect
sense
move Spine
Reflect
Reflex
move
Reflect
?
Physiology
Organs
Layered architectures
Cells
Ref
lect
move
sen
se
mo
ve
Spin
e
Ref
lect
Ref
lex
Prediction
Goals
Actions
errors
Actions
3D +time
Simulation
+ complex
models
(“priors”)
Seeing is dreaming
Conscious perception
Conscious perception
Zzzzzz…..
Which blue line is longer?
Which blue line is longer?
Which blue line is longer?
Which blue line is longer?
Chess experts
• can reconstruct entire
chessboard with < ~ 5s
inspection
• can recognize 1e5 distinct
patterns
• can play multiple games
blindfolded and simultaneous
• are no better on random
boards
(Simon and Gilmartin, de Groot)
www.psywww.com/intropsych/ch07_cognition/expertise_and_domain_specific_knowledge.html
Software
Hardware
Digital
Analog
Slow
Flexible
Fast
Inflexible
Delay is
most
important
For more
Evolution Facilitated variation Architecture = Constraints that deconstrain
• Weak linkage • Exploratory mechanisms • Compartmentalization
Kirschner and Gerhart
Unfortunately, not intelligent design
Ouch.
Making humans from fish parts.
wasteful
fragile
efficient
robust
Want to understand the space of systems/architectures
Want robust and efficient systems and architectures
Requirements on systems and architectures
efficient
robust simple
fragile
wasteful
complex
wasteful
fragile
efficient
robust
Want to understand the space of systems/architectures
Case studies?
Strategies?
Architectures? Want robust and efficient systems and architectures
Control, OR Comms
Compute Physics
Shannon Bode
Turing
Godel
Einstein
Heisenberg
Carnot
Boltzmann
Theory? Deep, but fragmented, incoherent, incomplete
Nash
Von Neumann
Kalman
Pontryagin
Control Comms
Compute Physics
Shannon Bode
Turing
Godel
Einstein
Heisenberg
Carnot
Boltzmann
wasteful?
fragile?
slow?
?
• Each theory one dimension
• Tradeoffs across dimensions
• Assume architectures a priori
• Progress is encouraging, but…
• Stovepipes are an obstacle…
Chandra, Buzi, and Doyle
UG biochem, math, control theory
Most important paper so far.
simple enzyme
Fragility
Enzyme amount
complex enzyme
lnz p
z p
2 2
0
1ln ln
z z pS j d
z z p
Theorem!
z and p functions of enzyme complexity
and amount
Savageaumics
How general is this picture? Very! Constraints! i.e. hard limits and architecture
simple tech
complex tech
wasteful
fragile
efficient
robust
Cat
abo
lism
AA
Ribosome
RNA
RNAp
transl.
Proteins
xRNA transc.
Pre
curs
ors
DNA
DNAp
Repl. Gene
ATP
ATP
Enzymes
Building Blocks
Control, OR Communications
Compute Physics
Shannon Bode
Turing
Einstein
Heisenberg
Carnot
Boltzmann
Delay is most
important
Delay is least
important
Control, OR Communications
Compute Physics
Shannon Bode
Turing
Einstein
Heisenberg
Carnot
Boltzmann
Delay is most
important
Delay is least
important
Compute
Turing (1912-1954)
• Turing 100th birthday in 2012
• Turing
− machine (math, CS)
− test (AI, neuroscience)
− pattern (biology)
• Arguably greatest*
− all time math/engineering combination
− WW2 hero
− “invented” software
*Also world-class runner.
Turing’s 3 step research: 0. Virtual (TM) machines 1. hard limits, (un)decidability
using standard model (TM) 2. Universal architecture
achieving hard limits (UTM) 3. Practical implementation in
digital electronics (biology?)
Essentials: 0. Model 1. Universal laws 2. Universal architecture 3. Practical implementation
Software
Hardware
Digital
Analog
Turing as “new”
starting point?
indirect
hard
relevance
easy
direct
Essentials: 0. Model 1. Universal laws
2. Universal architecture 3. Practical implementation
Brain
Useful case
studies?
3. Universal architecture 4. Practical implementation
indirect
hard
relevance
easy
direct
Brain
Tech nets
Cells
? ?
Bewildering
Frivolous
OS
Deconstrained
(Hardware)
Deconstrained
(Applications)
Layered architectures
Constrained Control, share, virtualize,
and manage resources
Processing
Memory
I/O
Few global variables
Don’t cross layers
Essentials
Cat
abo
lism
AA
Ribosome
RNA
RNAp
transl.
Proteins
xRNA transc.
Pre
curs
ors
DNA
DNAp
Repl. Gene
ATP
ATP
Enzymes
Building Blocks
Shared
protocols
Deconstrained
(diverse)
Environments
Deconstrained (diverse)
Genomes
Bacterial
biosphere
Architecture
=
Constraints
that
Deconstrain
Layered architectures
TCP IP
Physical
MAC
Switch
MAC MAC
Pt to Pt Pt to Pt
Diverse applications
Layered architectures
Proceedings of the IEEE, Jan 2007
Chang, Low, Calderbank, Doyle
OR
optimization
A rant
food intake
Glucose
Oxygen
Amino acids
Fatty acids
Organs
Tissues
Cells
Molecules
Universal metabolic system
Blood
Peter Sterling and Allostasis
foo
d
inta
ke
Glucose Oxygen
Amino acids Fatty acids
Organs Tissues Cells Molecules
Blood
food intake
Organs
Tissues
Cells
Molecules
Robust
Highly
variable
supply
Highly
variable
demand
Evolvable evolving
diet
evolving
function
Efficient
Organs
Tissues
Cells
Molecules
Highly
variable
demand
food intake
Highly
variable
supply
evolving
diet evolving
function
Glucose
Oxygen
Blood
Conserved
core
building
blocks
VTA Prefrontal cortex Accumbens
dopamine
Universal reward systems
sports music dance crafts art toolmaking sex food Dopamine,
Ghrelin,
Leptin,…
Ridiculous oversimplification
Universal reward systems
sports music dance crafts art toolmaking sex food
PFC
CG
OFC
NAcc
Amyg
STR
TH PIT
HIP
SN
VTA dopamine
Reward
Drive
Control
Memory
Robust and evolvable
VTA dopamine
Universal reward systems sports music dance crafts art toolmaking sex
food
Glucose
Oxygen
Organs
Tissues
Cells
Molecules
Universal metabolic system
Blood
Constraints
that
deconstrain
Reward Drive Control Memory
dopamine
Blood
Glucose
Oxygen
Constraints
Modularity 2.0
sports music dance crafts art toolmaking sex
food
Organs
Tissues
Cells
Molecules
Reward Drive Control Memory
that
deconstrain
Modularity 2.0
Layered architectures
“bow-tie”
diverse diverse
Layered architectures
“hourglass”
diverse
diverse
dopamine
work family community nature
Universal reward/metabolic systems
Robust and adaptive, yet …
food sex toolmaking sports music dance crafts art
Blood Organs
Tissues
Cells
Molecules
Reward Drive Control Memory
work family community nature
food sex toolmaking sports music dance crafts art
Organs
Tissues
Cells
Molecules
Reward Drive Control Memory
Modularity 1.0
“Weak linkage”
dopamine
Most important “modules”
Blood
Not weakly connected to others,
but highly connected
Modularity 2.0
dopamine
work family community nature
Universal reward/metabolic systems
Robust and adaptive, yet …
food sex toolmaking sports music dance crafts art
Blood Organs
Tissues
Cells
Molecules
Reward Drive Control Memory
sex food toolmaking sports music dance crafts art
work family community nature
cocaine amphetamine
dopamine
Blood Organs
Tissues
Cells
Molecules
Reward Drive Control Memory
work family community nature
sex toolmaking sports music dance crafts art
dopamine
Vicarious
money
salt sugar/fat nicotine alcohol
industrial agriculture
market/ consumer culture
Organs Tissues Cells Molecules
Reward Drive Control Memory
Vicarious
money
salt sugar/fat nicotine alcohol
high sodium
obesity
overwork
smoking
alcoholism
drug abuse
hyper- tension
athero- sclerosis
diabetes
inflammation
immune suppression
coronary, cerebro- vascular, reno- vascular
cancer
cirrhosis
accidents/ homicide/ suicide
From Sterling
VTA
Prefrontal cortex Accumbens
dopamine
Universal reward systems sports music dance crafts art toolmaking sex food
Glucose
Oxygen
Organs
Tissues
Cells
Molecules
Universal metabolic system
Blood
food
VTA dopamine
Glucose
Oxygen
Vicarious
money
salt sugar/fat nicotine alcohol
high sodium
obesity
overwork
smoking
alcoholism
drug abuse
hyper- tension
athero- sclerosis
diabetes
inflammation
immune suppression
coronary, cerebro- vascular, reno- vascular
cancer
cirrhosis
accidents/ homicide/ suicide
Yet Fragile
3. Universal architecture 4. Practical implementation
indirect
hard
relevance
easy
direct
Brain
Tech nets
Cells
? ?
Bewildering
Frivolous!
Frivolous!
Cloth
Thread
Fiber
Garments
Other examples
Words
Lego
Clothing
Cell biology
Internet
Cyberphysical
Money
Software
Hardware
Digital
Analog
“solution sets” (a la Marder, Prinze, etc)
large, thin, nonconvex
All systems
Letters and words
• 9 letters: adeginorz
• 9!= 362,880 sequences of 9 letters
• Only “organized” is a word
1 << (# words) << (# non-words)
large thin
Computer programs
• Almost any computer language
• Large # of working programs
• Much larger # of non-working programs
• “Nonconvex” = simple mashups of working programs don’t work
1 << (# programs) << (# non-programs)
large thin
large thin
1 << # toys << # piles
toys
pile
large thin
1 << # toys << # piles
pile
“order for free?”
edge of chaos
self-organized criticality
scale-free
???
statistical physics
random ensembles
minimally tuned
phase transitions
bifurcations
large thin
1 << # toys << # piles
toys
pile
“order for free?”
nonconvex
Software
Hardware
Digital
Analog
large, thin,
nonconvex
Analog
All “code”
Csete and Doyle
large thin
1 << # outfits << # heaps
outfit
wardrobe
heap
Shirt
Slacks
Jacket sweater
T-Shirt
Socks
Shoes Coat Shorts
Inn
er
Mid
dle
Ou
ter
Skin
Hidden
Robust to variations
and requirements in
• weather
• activity
• appearance
• wear and tear
• washing
System
constraints
Wool Cotton Nylon Silk Polyester Rayon
Cloth
Shirt
Slacks
Jacket Tie
T-Shirt
Socks
Shoes Coat Shorts
Sewing
Wool Cotton Nylon Silk Polyester Rayon
Cloth
Component constraints
Fiber
Geographically diverse sources
Diverse fabric
Functionally diverse garments
General
purpose
machines
Diverse Thread
sew
knit, weave
spin
Cloth
Thread
Fiber
Garments
Xform
Xform
Xform
Architecture and Modularity 2.0
Prevents unraveling of lower layers
sew
knit, weave
spin
Cloth
Thread
Fiber
Garments
Xform
Xform
Xform
Hidden,
large, thin,
nonconvex
Outfits
Assemble
are
necessary
denim
Cloth
Thread
Fiber
500m
.5mm
denim cheesecloth
needle &
thread
paper 500m
cheesecloth
needle &
thread
denim
paper
500m
.5mm x .35 mm cotton paper
500m
paper
Fiber
Paper
Minimal architecture
500m
paper
A neonatal rat
pyramidal neuron filled with Lucifer Yellow
imaged on the BioRad
MRC600 confocal
microscope using a 20X oil
objective, NA=0.8. Image
size is 320 x 425 x 120 μm.
0 50 100 150 200 250 300 0
50
100
150
200
250
300
350
400
100m
0 100 200 300
0
100
200
300
100m
.5mm x .35 mm cotton paper
.5mm x .35 mm cotton paper
100m
100m
47x35 micron
cotton fibers
60x50 microns
100m
100m
100m
10m
cotton fibers
47x35 micron
5m
5m
5m
1 m
Intel Tukwila
quad core chip
with more than
2 billion
transistors
10 mm
1 m
5m
5m
100m
.5mm x .35 mm cotton paper
.5mm x .35 mm cotton paper
100m
cheesecloth
500m
.5mm
cheesecloth
denim
cheesecloth
denim
cheesecloth
denim
500m
.5mm
500×500m
cheesecloth
denim
.5×.5mm
5×5 mm
500×500m
.5×.5mm
5×5 mm
5×5 mm
5×5 mm
21.5×32.5 mm
0 5 10 15 20 25 30 0
5
10
15
20
Intel Tukwila quad core chip with
more than 2 billion transistors
21.5×32.5 mm
5×5 mm
5×5 mm
5×5 mm
5×5 mm
5×5 mm
5×5 mm
500×500m
.5×.5mm
500×500m
cheesecloth
denim
.5×.5mm
cheesecloth
denim
cheesecloth
denim
cheesecloth
denim
500m
.5mm
500m
cheesecloth
500m
1.15mm x .85mm polyester plain weave
.5mm x .35 mm
cotton paper
500m
cheesecloth
needle &
thread
denim
polyester
plain
weave
paper
500m
cheesecloth
needle &
thread
denim
polyester
plain
weave
paper
500m
cheesecloth
needle
&
thread
denim
500m
Cloth
Thread
Fiber
Xform
Xform
Hidden,
large, thin,
nonconvex
are
necessary
Cloth
Thread
Fiber
Garments
Xform
Xform
Xform
Hidden,
large, thin,
nonconvex
Outfits
Assemble
are
necessary
Cloth
Thread
Fiber
Garments
Xform
Xform
Xform
Universal strategies?
Garments have
limited access to
threads and fibers
constraints on
cross-layer
interactions
quantization
for robustness
Even though
garments seem
analog/continuous
Prevents unraveling of lower layers
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Geographically diverse sources
Diverse outfits
Constraints
that
deconstrain
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Diverse outfits
Constraints
that
deconstrain
System
constraints
Component
constraints
Protocol
constraints
Diverse sources
Diverse outfits
Protocol Cloth
Thread
Fiber
Garments
Xform
Xform
Xform
Constraints
that
deconstrain
?
Fiber
Diverse outfits
?
Fiber
Diverse outfits
SOC/EOC?
Self-Organized Clothing
Edge of Couture
Scale-Free Fashion Random
Fiber
Diverse outfits
Small gap
No architecture
Self-Organized Clothing
Standard error
Edge of Couture
Scale-Free Fashion
Fiber
Diverse outfits
Huge gap
No architecture
Supernatural
Standard error
Fiber
Diverse outfits
Huge gap
Supernatural
Fiber
Small gap
Emergence
Self-organized
Phase transition
Edge-of
Scale-free
Diverse outfits
Mysteries in the gaps
No architecture
Huge gap
Supernatural Mysteries in the gaps
No architecture
Mainstream Small gap
Emergent
Different worlds
Huge gap
Supernatural Mysteries in the gaps
No architecture
Mainstream Small gap
Emergent
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Diverse outfits
Constraints
that
deconstrain Protocols
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Diverse outfits
Constraints
that
deconstrain Protocols
Cloth
Thread
Fiber
Cloth
Thread
Fiber
Garments
Cloth
Thread
Garments
Cloth
Thread
Fiber
Cloth
Thread
Garments
Layered, large, thin
Components
Organized
Components
Organized
large thin
1 << # outfits << # heaps
outfit
heap
Components
Organized
Components
Components
large thin
1 << # outfits << # heaps outfit
Organized
Software
Hardware
Digital
Analog
Cloth
Thread
Fiber
Cloth
Thread
Garments
Components
Organized
Components
Organized
What is the “true functional unit?”
Virtual machines
Software
Hardware
Digital
Analog
Cloth
Thread
Fiber
Cloth
Thread
Garments
Components
Organized
Components
Organized
Organized Software Garments
Implementations
Virtual machines
Digital
Analog
Thread
Fiber
Cloth
Thread Components
Organized
Implementations
Organized Software Garments
Digital Cloth Organized
Software
Hardware
Digital
Analog
Cloth
Thread
Fiber
Cloth
Thread
Garments
Components
Organized
Components
Organized
Layered, large, thin
Hidden
Virtualized
Garm
ent
Garm
ent
Garm
ent
Outfit
Outfit
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment
Garm
ent
Garm
ent
Garm
ent
Outfit
Cloth
Thread
Fiber
Garment
Ga
rme
nt
Cloth
Thread
Fiber
Garment G
arm
en
t
Cloth
Thread
Fiber
Garment
Layering within garments (textiles)
Meta-layers
Fast,
Limited
scope
Slow,
Broad
scope
Unfortunately, we’re not
sure how this all works.
Ba
se
Insu
latio
n
Sh
ell
Outfit
Body Environment
• Complexity Robustness
• Layers must be hidden to be robust
• Choice (management and control) is
more complex than assembly
Ba
se
Insu
latio
n
Sh
ell
Outfit
Body Environment
Choice & Control
Assembly
Shirt
Slacks
Jacket sweater
T-Shirt
Socks
Shoes Coat Shorts
Inn
er
Mid
dle
Ou
ter
Skin
Selection
assembly
Outfit
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment
Outfit
Cloth
Thread
Fiber
Cloth
Thread
Garment
Asse
mb
ly
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Garments
Weave
Sew
Spin
Farm
Do
ma
in s
pe
cific
Asse
mb
ly
Cloth
Thread
Fiber
Garments
Xform Ctrl
Networked,
universal,
layered Xform Ctrl
Xform Ctrl
Xform Ctrl
Do
ma
in s
pe
cific
Cloth
Thread
Fiber
Garments
Xform Ctrl
Networked,
universal,
layered Xform Ctrl
Xform Ctrl
Xform Ctrl
Do
ma
in s
pe
cific
D
em
an
d S
up
ply
Cloth
Thread
Fiber
Garments
Xform Ctrl
Networked,
universal,
layered Xform Ctrl
Xform Ctrl
Xform Ctrl
Do
ma
in s
pe
cific
S
up
ply
Co
ntro
l
Desig
n
Complexity?
Co
ntro
l
De
sig
n
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Diverse outfits
Constraints
that
deconstrain
System
constraints
Component
constraints
Protocol
constraints
Assem
bly
Co
ntro
l
De
sig
n
Cloth
Thread
Fiber
Diverse Garments
Xform
Xform
Xform
Diverse outfits
Constraints
that
deconstrain
System
constraints
Component
constraints
Protocol
constraints
Assem
bly
Fiber
Geographically diverse sources
Diverse fabric
Functionally diverse garments
General
purpose
machines Diverse Thread
sew
knit, weave
spin
Functionally diverse garments
Cloth
Thread
Fiber
Garments
Fragilities?
Cloth
Thread
Fiber
Garments
Scalable Sustainable?
TCP IP
Physical
MAC
Switch
MAC MAC
Pt to Pt Pt to Pt
Diverse applications
Layered architectures
TCP IP
Physical
Diverse applications
Diverse
Too clever?
TCP IP
Deconstrained
(Hardware)
Deconstrained
(Applications)
Layered architectures
Constrained Networks
“constraints that deconstrain” (Gerhart and Kirschner)
TCP/
IP
Deconstrained
(Hardware)
Deconstrained
(Applications)
Constrained
Facilitated wild evolution
Created
• whole new ecosystem
• completely opposite
Why?
• OS better starting point than
phone/comms systems
• Extreme robustness confers
surprising evolvability
• Creative engineers
• Rode hardware evolution
Networked OS
Architecture
OS
Deconstrained
(Hardware)
Deconstrained
(Applications)
Layered architectures
Constrained Control, share, virtualize,
and manage resources
Processing
Memory
I/O
Few global variables
Don’t cross layers
Essentials
Cat
abo
lism
AA
Ribosome
RNA
RNAp
transl.
Proteins
xRNA transc.
Pre
curs
ors
DNA
DNAp
Repl. Gene
ATP
ATP
Enzymes
Building Blocks
Shared
protocols
Deconstrained
(diverse)
Environments
Deconstrained (diverse)
Genomes
Bacterial
biosphere
Architecture
=
Constraints
that
Deconstrain
Layered architectures
Cat
abo
lism
AA
Ribosome
RNA RNAp
transl. Proteins
xRNA transc.
Pre
curs
ors
DNA DNAp
Repl. Gene
ATP
ATP
Enzymes
Building
Blocks
Crosslayer
autocatalysis
Macro-layers
Inside every cell
almost
Cat
abo
lism
AA
Ribosome
RNA
RNAp
transl.
Proteins
xRNA transc.
Pre
curs
ors
DNA
DNAp
Repl. Gene
ATP
ATP
Enzymes
Building Blocks
Core conserved
constraints facilitate
tradeoffs
Deconstrained
phenotype
Deconstrained genome
What makes the bacterial
biosphere so adaptable?
Active control of
the genome
(facilitated
variation)
Environment
Action
OS
Deconstrained
(Hardware)
Deconstrained
(Applications)
Layered architectures
Constrained Control, share, virtualize,
and manage resources
Processing
Memory
I/O
Few global variables
Don’t cross layers Direct
access to
physical
memory?
Cata
bolis
m
AA
Ribosome
RNA
RNAp
transl.
Proteins
xRNA transc.
Pre
curs
ors
DNA
DNAp
Repl. Gene
ATP
ATP
Enzymes
Building Blocks
Shared
protocols
Deconstrained
(diverse)
Environments
Deconstrained (diverse)
Genomes
Bacterial
biosphere
Architecture
=
Constraints
that
Deconstrain
Few global variables
Don’t cross layers
Problems with leaky layering
Modularity benefits are lost
• Global variables? @$%*&!^%@&
• Poor portability of applications
• Insecurity of physical address space
• Fragile to application crashes
• No scalability of virtual/real addressing
• Limits optimization/control by duality?
Fragilities of layering/virtualization
“Universal” fragilities that must be avoided
• Hijacking, parasitism, predation
– Universals are vulnerable
– Universals are valuable
• Cryptic, hidden
– breakdowns/failures
– unintended consequences
• Hyper-evolvable but with frozen core
TCP/
IP
Deconstrained
(Hardware)
Deconstrained
(Applications)
Layered architectures
Constrained Control, share, virtualize,
and manage resources
I/O
Comms
Latency?
Storage?
Processing?
Few global variables?
Don’t cross layers?
App App
IPC
Global
and direct
access to
physical
address!
DNS
IP addresses
interfaces
(not nodes)
caltech.edu?
131.215.9.49
App App
IPC
Global
and direct
access to
physical
address!
Robust?
• Secure
• Scalable
• Verifiable
• Evolvable
• Maintainable
• Designable
• …
DNS
IP addresses
interfaces
(not nodes)
Physical
IP
TCP
Application
Naming and addressing need to be
• resolved within layer
• translated between layers
• not exposed outside of layer
Related “issues”
• VPNs
• NATS
• Firewalls
• Multihoming
• Mobility
• Routing table size
• Overlays
• …
TCP
IP
Physical
Diverse
Until late 1980s, no
congestion control, which
led to “congestion collapse”
Original design challenge?
TCP/
IP
Deconstrained
(Hardware)
Deconstrained
(Applications)
Constrained • Expensive mainframes
• Trusted end systems
• Homogeneous
• Sender centric
• Unreliable comms
Facilitated wild evolution
Created
• whole new ecosystem
• completely opposite
Networked OS
Reactions
Assembly
DNA/RNA
control
control
DNA
Ou
tsid
e
Insid
e
Tra
ns
mit
ter
Ligands &
Receptors
Re
ce
ive
r
Responses
control
Receiver
Re
sp
on
se
s
Protein
Cross-layer control
• Highly organized
• Naming and addressing
Coming later:
contrast with cells
?
Deconstrained
(Hardware)
Deconstrained
(Applications)
Next layered architectures
Constrained Control, share, virtualize,
and manage resources
Comms
Memory, storage
Latency
Processing
Cyber-physical
Few global variables
Don’t cross layers
Every layer
has
different
diverse
graphs.
Architecture is least
graph topology.
Architecture
facilitates
arbitrary
graphs.
Persistent errors
and confusion
(“network science”)
Physical
IP
TCP
Application
Notices of the AMS, 2009
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment
Cloth
Thread
Fiber
Cloth
Thread
Garment C
on
trol
Des
ign
Choice &
Control
Assembly
Assem
bly
Fast,
Limited
scope
Slow,
Broad
scope
Software
Hardware
Digital
Analog