Bio-Networking Architecture

1
Bio-Networking Architecture

• Michael Wang
• Tatsuya Suda
• Univ. of California, Irvine

2
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

3
Key Requirements

• Network services/applications need to be
• scalable (to millions of users)
• adaptable (to heterogeneous and dynamic network
  conditions)
• survivable (from security attacks) and (always)
  available
• simple (easy to design and maintain)

4
Large Scale Biological Systems

• Example bee colony key features
• scalable
• adapt to environment
• evolve to more optimal forms
• secure and survivable
• relatively simple components (individual bees)

5
Key Biological Concepts

• Emergent behavior/characteristics
• simple behaviors
• Lifecycle
• food/energy, reproduction, death
• Evolution
• diversity, natural selection
• Local interactions
• Scent/pheromones
• Self-protection

6
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

7
Bio-Net Emergent Behavior

• Biological systems
• useful group behavior emerge from autonomous
  local interaction of individuals with simple
  behaviors

8

• Bio-Network
• applications/services are implemented by a
  distributed, collective entity, super-entity
• super-entity consists of multiple autonomous
  entities, cyber-entities
• each cyber-entity behaves similarly to biological
  entities (e.g., migration, replication/reproductio
  n, energy exchange)
• each cyber-entity has basic functionality related
  to its application and service

9
super-entity
individual cyber-entities

• Desirable characteristics emerge in the
  super-entity from the actions/interactions of its
  cyber-entities

10
Bio-Net Food and Energy

• Biological systems
• biological entities strive to gain energy by
  seeking and consuming food

11

• Bio-Network
• cyber-entity stores and expends energy (food,
  money)
• energy exchange
• gain energy from a user in exchange for
  performing a service
• expend energy to use network/computing resources
• energy can be used as a control mechanism
• abundance induces replication or reproduction
• scarcity induces death

12
Bio-Net Evolution and Adaptation

• Biological systems
• the biological system specializes and optimizes
  itself for environmental changes of long-term
  (evolution) and short-term (adaptation)
• key components
• diversity from mutations and crossovers during
  replication and reproduction
• natural selection keeps entities with beneficial
  features alive and increase reproduction
  probability

13

• Bio-Network
• super-entity and cyber-entities evolve, adapt,
  and localize through diversity and natural
  selection
• diversity
• cyber-entities replicate/reproduce with mutation
  and crossover
• human designers can introduce diversity in
  cyber-entities

14

• natural selection
• death from energy starvation
• tendency to replicate/reproduce from energy
  abundance

15
Bio-Net Local Interactions (Social Networking)

• Human society six degrees of separation
  concept
• any two humans being separated by at most six
  relationships
• A asks his friends to find B friends of A ask
  their friends to find B etc.

16

• Bio-Network
• directory service capability (Social Networking)
• cyber-entities establish relationships
• cyber-entities in same super-entity are of the
  same family
• cyber-entities in different super-entities may
  form friends
• when a cyber-entity wants to find a cyber-entity
  belonging to another super-entity, it asks its
  family and friends
• if they dont have a relationship with a
  cyber-entity in the target super-entity, they ask
  their family and friends

17
Bio-Net Pheromones

• Biological systems
• pheromones are emitted to attract partners, find
  partners, deter enemies, and sense danger
• pheromones decay with time and distance

18

• Bio-Network
• cyber-entities emit pheromone packets
• to improve the social networking efficiency
• check the adjacent nodes for pheromones
• follow the strength of pheromones
• to keep cyber-entities of the same type apart

19
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design (design only, not
  implemented)
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

20
Design Node Architecture
Resource Cyber-entity
Other Cyber-entities
Resource negotiation Energy exchange
Resource configuration control
Resource access
Bio-Networking Platform Software
Virtual Machine (e.g. Java Virtual Machine)
Unmodified, commercial software
Heterogeneous OS Hardware
21
Design Cyber-Entity
ID
Super-entity ID
Attributes
Type
Stored Energy
Age
Non-Executable Data
Body
Cyber-Entity
Executable Code
Migration
Replication
Reproduction
Behavior
Protection
Service
Communication
Pheromone Emission
...
22
Basic Cyber-entity Behaviors

• Migration
• Replication, Reproduction
• Social Networking
• Pheromone Emission
• Protection
• Communication
• Energy Exchange

23
Design Behaviors

• A cyber-entity behavior can be implemented by a
  number of algorithms

24
Design Migration Behavior
Example algorithm

calculate the direction a user request came from
- W2
calculate cost of migration
W1
gt M
If
then migrate
Factors which affect migration
W1, W2, and M (aggressiveness) are parameters
25
Design Reproduction

• Factors that affect reproduction behavior
• stored energy level
• availability of desirable mate (age, energy
  level, type)
• reproductive aggressiveness
• Crossover and mutation

26
Mutation

• Concatenate behavior factor weights into a bit
  string
• Flip a random bit to mutate

Aggressiveness
Energy source proximity
Mutual repulsion
Migration Cost
gt

-
0011
1001
1011
1001
0011
1001
1001
1001
27
Mutation

• Arrange algorithm in a tree

gt
gt
-
-
mutation


Aggressiveness
Aggressiveness
Migration Cost
Migration Cost
Fairness
Mutual repulsion
Energy source proximity
Energy source proximity
28
Crossover
Parent A behaviors
Parent B behaviors
Child behaviors
29
Increasing Diversity

• Human designers can introduce behavioral
  diversity into the cyber-entity population
• Diversity will
• ensure large domain for evolution
• make the population more immune to specific
  attacks

30
Design Social Networking

• Cyber-entities establish relationships
• family relationship (cyber-entities belonging to
  the same super-entity)
• friend relationship (cyber-entities not in the
  same super-entity)
• Cyber-entities with relationships inform each
  other of their super-entity membership and
  current network location
• existing multicast algorithms will be used

31

• For cyber-entity s to find super-entity T,
• s multicasts inquiry to a subset of its family
  members.
• if family members share a node or have
  relationship with any members of T or can detect
  a member of T using pheromones, T is found.
• otherwise, family members contacted by s ask
  their friends to find T
• this process repeats until T is found.

32
Finding a Super-Entity
Red Group Found at R

• Green cyber-entity wants to find/join Red
  super-entity
• Green finds Blue Group
• Green multicasts inquiry to Blue Group
• if Blue Group does not find Red Group, Blue Group
  asks its friends

G
33
Finding Nearest Cyber-Entity

• Starting from R0
• navigate toward the adjacent node with smallest
  distance to G
• stop navigation when all other adjacent nodes are
  farther in distance

R0
G
R
34
Social Networking Issues

• Recursive querying large amounts of traffic
• sequential search vs. parallel search
• a maximum propagation count set for queries
• initial queries with low propagation count
  increase the count if previous queries did not
  succeed
• Social network may be partitioned so that one set
  of cyber-entities cannot find another set

35
Design Pheromone Emission

• Cyber-entity can emit pheromone packets
  containing cyber-entity ID, super-entity
  membership
• Factors of the pheromone emission
• strength of pheromone
• decays with distance from the source
• decays with time
• rate of pheromone emission

36

• Pheromone
• allows a cyber-entity to know what other
  cyber-entities on neighboring nodes
• to improve the performance of the social net
• to keep cyber-entities of same type apart
• to find a mate for reproduction

37
Design Protection

• Cyber-entities protection behaviors include
• information fragmentation and dispersion

38
Fragmentation and Dispersion
A file or document is fragmented and given to
multiple cyber-entities Cyber-entities disperse
by migrating to different places in the network
b
a
a
c
b
c
original file
abc
39
Variable Rate of State Change
Not all cyber-entities change immediately
Authenticated (signed) Update Request
40
Consistency Checking
All cyber-entities periodically check consistency
with their replicas
Successful attack
41
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design (partially implemented)
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

42
Design Node Architecture
Resource Cyber-entity
Other Cyber-entities
Resource negotiation Energy exchange
Resource configuration control
Resource access
Bio-Networking Platform Software
Virtual Machine (e.g. Java Virtual Machine)
Unmodified, commercial software
Heterogeneous OS Hardware
43
Design Platform Software

• Platform software provides
• cyber-entity execution environment
• strict resource control
• various system services for cyber-entities
• Platform software can implement some of
  cyber-entity behaviors

44

• Implementing functionality in platform software
  vs. implementing functionality in cyber-entity
  behaviors
• platform software implementation
• more secure
• to reduce size of cyber-entities
• implementation can be optimized
• cyber-entity behavior implementation
• easily changed and can evolve

45
Design Platform Software
Cyber-Entity Execution Environment
Cyber-Entity Context
Cyber-Entity Registry
Cyber-Entity LIfecycle
Registry Lookup
Group Membership
Social Networking
Migration
Cyber-Entity Commun.
Energy Management
Pheromone Emission
Security
46
Design Platform Software

• Some components have been implemented

Cyber-Entity Execution Environment
Cyber-Entity Context
Cyber-Entity Registry
Cyber-Entity LIfecycle
Registry Lookup
Group Membership
Social Networking
Migration
Cyber-Entity Commun.
Energy Management
Pheromone Emission
Security
implemented
47
Demo, migration
48
Same age, do nothing.
Update agent file.
Update local file.
Demo, updating files
49
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

50
Adaptation Simulation

• Ability of cyber-entities to respond to changes
  in their local environment
• Cyber-entity behaviors implemented
• migration, replication, parameter mutation, death
• Environment factors considered
• ResourceCostFactor cost of living at current
  node - cost at a neighboring node
• EnergySeekingFactor direction of energy sources
  (which adjacent node the request came from)
• RepulsionFactor congregation of similar
  cyber-entities

51
Migration Behaviors

• W1 ResourceCostFactor
• W3 RepulsionFactor gt Aggressiveness
• Resource Cost Minimizing Entity
• migrate towards regions with lowest resource
  costs
• avoid coexisting on a node with similar entity

52
Migration Behaviors

• W1 ResourceCostFactor W2 EnergySeekingFactor
  
• W3 RepulsionFactor gt Aggressiveness
• Resource Cost Minimizing Entity
• migrate towards regions with lowest resource
  costs
• avoid coexisting on a node with similar entity
• Energy Seeking Entity
• migrate towards source of energy (user requesting
  service)
• avoid coexisting on a node with similar entity

53
Replication Behavior

• If current energy level gt aggressiveness
• create a new entity of same type
• mutate weight factors and aggressiveness value

54
Energy Seeking Entity (Simulation 1)
Entity 3 w1 .575, w2 .45, aggress 4.5
Entity 1 w1 .5, w2 .5, aggress 4
Entity 2 w1 .425, w2 .575, aggress 2.25
55
Resource Cost Minimizing Entity (Simulation 1)
Entity 1 w1 .5, w2 .5, aggress .25
Entity 3 w1 .6, w2 .55, aggress .75
Entity 2 w1 .55, w2 .45, aggress 0
Entity 4 w1 .5, w2 .5, aggress .1
56
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

57
Bio-Net Based Web Caching/Replication System

• Cyber-entities hold one or small set of web pages
  
• User request for a web page is associated with a
  value
• value of a web page request energy unit value
  in a service request in Bio-Net

58

• Cache replacement algorithm in Bio-Net
• market based (users can specify how important
  each web page is)
• The amount of energy units received by a
  cyber-entity affects its behaviors (e.g.
  migration and reproduction)
• Tracking of hit rates in Bio-Net
• active thread of execution is associated with a
  web page (cyber-entity)

59
Users Indicate Value of Web request
BlueHat/prodA
BlueHat/prodA



BlueHat/prodA
FanClub/Zed

FanClub/Zed

user downloads software for business evaluation
user download newsletter for entertainment
- only slight value to user

- valuable to user
60
Effect of value on cache replacement algorithm
As disk and memory resources become scarce on the
Bio-Networking Platform, resource costs on the
node increase. This may cause the FanClub
cyber-entity to migrate away and make room for
other, possibly more valuable cyber-entities.
BlueHat/prodB
(distant) Bio-Networking Platform
BlueHat/prodA
FanClub/Zed
Bio-Networking Platform
user downloads software for business evaluation
user download newsletter for entertainment
- only slight value to user
- valuable to user

61
Effect of value on cache replacement algorithm
As disk and memory resources become scarce on the
Bio-Networking Platform, resource costs on the
node increase. This may cause the FanClub
cyber-entity to migrate away and make room for
other, possibly more valuable cyber-entities.
FanClub/Zed
(distant) Bio-Networking Platform
BlueHat/prodA
BlueHat/prodB
Bio-Networking Platform
user downloads software for business evaluation
user download newsletter for entertainment

- valuable to user

62
Thread of Execution Associated with Web Page

• A web page (i.e., cyber-entity) has a thread of
  execution associated with it.
• tracking page hits or access patterns
• implementing dynamic content, scripts, etc
• intellectual property protection/checking

63
Outline

• Motivation (Observation of Biological Systems)
• Overview of the Bio-Networking Architecture
• how bio concepts are used in the bio net
• Bio-Networking Architecture Design
• cyber-entity design
• platform software design
• Preliminary Adaptation Simulation
• Bio-Net Based Web Application
• Summary

64
Summary and Conclusion

• Bio-Networking Architecture represents a new
  paradigm in the construction of network services
  and applications
• Services and applications consists of multiple,
  autonomous cyber-entities that exhibit emergent
  behavior
• Bio-Networking is adaptable, evolvable, secure,
  survivable, scalable, and simple

65
Research Questions

• What are the beneficial concepts and mechanisms
  from the biological world?
• What is the relationship between individual
  behaviors and emergent behaviors?
• What is the stability/adaptability tradeoff?
• Can Bio-Networking Architecture evolve fast
  enough?
• How secure and survivable is the Bio-Networking
  Architecture?
• What is the performance and overheads of Social
  Networking?