This talk is also available as a 4-page paper in http://bit.ly/SCVTpaper
(Proceedings of IEEE SCVT 2012)
Abstract—As communication networks become increasingly complex and dynamic, the three functions known as monitoring, control and management prove to be ineffective. It is increasingly difficult to operate large networks, perform diagnostics, prevent cascading failures, or deliver dependable services. I argue that this is because, although the Internet serves more terminals than there are neurons in the brain, we still handle our networks via deterministic protocols. We still try to capture the complex entangle of interconnections by creating deterministic models of the network, its traffic and control system. By the year 2020 the number of interconnected ‘things’ will grow by a factor of a thousand, and networks will be programmed to ‘learn’ how to detect new communication patterns and self-regulate, rather than acting deterministically. I introduce the anatomy of a smart network, discussing what more could be achieved with it.
Farewell to deterministic networks: smarter networks for the Internet of Things
1. Farewell to deterministic networks:
Smarter networks for the Internet of Things
Prof. Antonio Liotta
Eindhoven University of Technology
http://bit.ly/autonomic_networks
http://nl.linkedin.com/in/liotta
https://twitter.com/#!/a_liotta
www.slideshare.net/ucaclio
http://bit.ly/press_articles
2. Three questions about networks
• How have we come to build ‘deterministic’ networks?
• Can we really connect 1 trillion ‘things’ by 2020?
• What’s the anatomy of a smart network?
Prof. A. Liotta 2
3. Advances in communication networks:
What was the first ‘deterministic’ network?
The Telegraph (William Cooke and Charles Wheatstone, 1839)
Prof. A. Liotta 3
4. Advances in communication networks:
What was the first ‘decent’ switched network ?
Prof. A. Liotta
The ‘human’ switch, circa 1940 4
5. Advances in communication networks:
The perfectly engineered Telecommunications Management Network
Architecture:
• Layering
• Abstraction
• Insulation
• Deterministic
Strategy
• Sophisticated
• Dimensioned
• Specialized
• Standardized
Legacy
Prof. A. Liotta 5
6. Advances in communication networks:
A perfectly simple network
Diverse global services
TMN Architecture:
• Layering
General-purpose • Abstraction
transport • Insulation
• Deterministic
Strategy
• Simple
• Over-dimensioned
• General-purpose
• Open
Diverse physical transmission media
Legacy
Prof. A. Liotta 6
7. Advances in communication networks:
The (unsuccessfully) re-engineered IP
Diverse global services
TCP/IP
TMN Architecture:
QoS, DiffServ, MPLS, … • Layering
• Abstraction
• Insulation
• Deterministic
Strategy
• Complex
Diverse physical transmission media • Dimensioned
• General-purpose
QoS-enabled • Standardized
IP has failed!
Legacy
Prof. A. Liotta 7
8. Advances in communication networks:
New direction: keep the network simple, push complexity to the edges
Re-engineered IP
TCP/IP
TMN Architecture:
• Layering
• Abstraction
• Insulation
• Deterministic
Strategy
• Ultra-Complex
• Over-dimensioned
• General-purpose
• Open
Legacy
Prof. A. Liotta 8
9. Next ??? IoT: smart networks to handle
diverse communication requirements
Re-engineered
IP IoT
TCP/IP
TMN Architecture:
• Layering
• Abstraction
• Insulation
• Deterministic
Strategy
• Ultra-Complex
• Over-dimensioned
• General-purpose
We can no longer tackle • Open
complexity via complexity
Legacy
Prof. A. Liotta 9
10. Next ??? IoT: smart networks to handle
diverse communication requirements
Re-engineered
IP IoT
TCP/IP
TMN Architecture:
• Layering
• Abstraction
• Cross-layer data
fusion
• Non-deterministic
Strategy
• Smart
• Evolutionary
• General-purpose
• Open
Legacy
Prof. A. Liotta 10
13. Today: communication protocols assume that infinite energy is available
Tomorrow: communication networks will integrate with energy networks
1trillion, each ~100 mWatt transmission power 100, each 1 Gigawatt supply
Prof. A. Liotta 13
14. Today: networks stay neutral from apps requirements and user perception
Tomorrow: networks no longer ignore QoE!
Sending rate: 2048 Kbps
HIGHER
39.48 dB
- 23.85 dB
Received quality
LOWER
15.65 dB
Prof. A. Liotta 14
15. Today: networks stay neutral from apps requirements and user perception
Tomorrow: networks no longer ignore QoE!
Sending rate: 768 Kbps
LOWER
35.15 dB
- 16.03 dB
Received quality
HIGHER
19.12 dB
Prof. A. Liotta 15
16. Todays network meets the definition of “complexity”
set by “complexity theory”
• Properties of whole can’t be
inferred from properties of
individual parts
• Individual components
interact nonlinearly, leading to
emergent behavior
• Constantly evolves and
unfolds over time
Prof. A. Liotta 16
17. Today: networks can’t catch up with complexity /diversity / dynamics
Smart networks: autonomic, learning, cognitive networks
Nature-like features
Hidden patterns
Emergent behavior
Self-regulation
Learning
Biological
models
Machine learning
Prof. A. Liotta 17
19. Today: we know how to build learning robots
Tomorrow: networks will learn too
Original
Clips
50 50
100 100
150 150
200
Difference in 250
200
250
Temporal 300 300
350
350
Motion 400 400
450
450
500
500
550
550
100 200 300 400 500 600 700
100 200 300 400 500 600 700
Initially we train the Reinforcement Learning
network to handle teaches how to handle
“News over Laptop” “Sport over Phone”
Prof. A. Liotta 19
21. Thank you !
Want to author or
edit a book?
New Springer Series:
Internet of Things –
Technology,
Communications
and Computing
Get in touch!!
http://bit.ly/pervasive-networks liotta.antonio@gmail.com
Prof. A. Liotta 21