1. Marie Curie Chairs (EXC) IP FLows over Optical and Wireless (“IP-FLOW”) “Kick-off” Meeting – Monday 28 June, 2004 Yoram Ofek Department of Information and Communications Technology University of Trento, Italy

2. Prof. Yoram Ofek - DIT - University of Trento, Italy 2 Agenda  10:30 Opening:  Prof. Fausto Giunchiglia – Vice Rector – University of Trento  Prof. Roberto Battiti – University of Trento  Mr. Philippe Froissard – European Union  10:45 Prof. Yoram Ofek (University of Trento) – IP-FLOW Overview and Planning  Discussion/planning Research Area 1:  11:30 Prof. Mario Baldi (Politecnico di Torino) and Prof. Achille Pattavina (Politecnico di Milano) – “Minimum Complexity Non-blocking Switching”  11:50 Prof. Mario Baldi and Ing. Guido Marchetto (Politecnico di Torino) – "Time-based Router”  11:40 Prof. Yoram Ofek “40 Tb/s Time-based Non-blocking Switch Design”  Lunch 12:00 – 13:00  Discussion/planning Research Area 2:  13:00 Prof. Moti Yung (Columbia University - USA) – “Trusted Flow: Why, How, and Where?”  13:15 Prof. Fabio Massacci (University of Trento) – “Malware Protection”  13:30 Prof. Roberto Battiti (University of Trento) – “Trust and Location”  13:45 Dr. Paolo Falcarin and Dr. Maurizio Morisio (Politecnico di Torino) – “A proposed TrustedFlow System Architecture with Aspect-oriented Implementation”  14:15 Dott. Davide D'Aprile Dr. Daniele Mazzocchi (Istituto Superiore Mario Boella - Torino) – “A TrustedFlow Enabled Run-time Environment and Trusted Flow Formal Specification”  14:30 Dr. Luca Tagliaferri and Prof. Paolo Prinetto (Politecnico di Torino) – “Code Self-checking Techniques for TrustedFlow Implementation” Discussion/planning Research Area 3:  15:00 Prof. Bezalel Gavish (SMU - USA) – Placement of multidirectional antennas  15:15 Prof. Andrea Massa / Prof. Renato Lo Cigno (University of Trento) – “the inter-relationship: APs placement, antenna design, Radio wave propagation properties”  15:30 Prof. Yoram Ofek (University of Trento) – Concluding Discussion  16:00 – 16:30 Refreshment  16:30 – 18:00 Separate discussions and detailed planning  20:00 – Dinner – Ristorante Chiesa

3. Prof. Yoram Ofek - DIT - University of Trento, Italy 3 IP-FLOW Overview  Focus on:  1. Flow (switching/forwarding) of data over the Internet: optical and wireless  [Rather than on routing]  2. Internet/telecom business and security

4. Prof. Yoram Ofek - DIT - University of Trento, Italy 4 IP-FLOW Overview  Research Area 1 – UTC-based pipeline forwarding for solving switch and link bottlenecks delivery of very high capacity streaming data  Research Area 2 – TrustedFlow for run-time software authentication trusted delivery and processing of data  Research Area 3 – B-GLOW - Broadband services via GLobal Optical and Wireless integration - delivering streaming data to Wi-Fi clients

5. Prof. Yoram Ofek - DIT - University of Trento, Italy 5 Research Area 1 - Motivation The Telecom (Unfinished) Story  Why the telecom industry: vendors and providers are more or less in “bankruptcy” (  two trillions $ debts)  They ignored cash flow:  1. spending money wisely …  2. earn more money from new services  Optimal UTC-based pipeline forwarding ensures:  1. Efficient expenditure on equipment/infrastructure:  Single fully utilized network – convergence to 1 wire 1 network  2. Revenue from services with deterministic QoS:  Streaming/interactive media: video, audio, multimedia

6. Prof. Yoram Ofek - DIT - University of Trento, Italy 6 Pipelines are deployed to increase efficiency in  Factory (automotive), Computers (CPU), … NOW (deterministic) pipeline over the Internet! Thanks to GPS//... that provides UTC – Coordinated Universal Time everywhere NOW (deterministic) pipeline over the Internet! Thanks to GPS/ GALILEO /... that provides UTC – Coordinated Universal Time everywhere Research Area 1 UTC-based Pipeline Forwarding/Switching Time-driven Switching (TDS) 12 1000 Time Cycle0 12 1000 Time Cycle1 12 1000 Time Cycle 79 UTC second Time-of-Day or UTC 0 beginning of a UTC second Time-driven Priority (TDP) Optimal for Video-Conferencing – Why?

7. Prof. Yoram Ofek - DIT - University of Trento, Italy 7 Research Area 1 UTC-based for Streaming Media  Deterministic performance:  No loss,  Minimum delay and  Constant jitter  end-to-end delay < 100 ms - interactive voice/video  Minimum buffers for all-optical realization  Enables huge capacity increase for streaming  1000 capacity increase may create major switching bottlenecks

8. Prof. Yoram Ofek - DIT - University of Trento, Italy 8 Research Area 1 UTC-based Switching Complexity  Minimum complexity non-blocking switching – thus very high scalability  Analytically proven  UTC-based switch = Banyan  Minimum complexity  Crossbar:  Clos:

9. Prof. Yoram Ofek - DIT - University of Trento, Italy 9 Crosspoint Switches: Vitesse –VSC3140VC, or MindSpeed – M21150 32-by-32 320 Gbps 1 32 32-by-32 320 Gbps 1 32 1 32-by-32 320 Gbps 1 32 32-by-32 320 Gbps 1 32 1 32x320: 10 Tbps Switching Module Electrical Interconnection Research Area 1 UTC-based 40 Tera-bit/s Switch

10. Prof. Yoram Ofek - DIT - University of Trento, Italy 10 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 32-by-32 1 32 32-by-32 1 32 1 Only off-the-shelf components Thus, simple to prototype Pipeline Forwarding Switch Controller 32-by-32 1 1 32 10 Tbps Switching Module Interconnection: Electrical or Optical UTC Research Area 1 UTC-based 40 Tera-bit/s Switch 50 times less complexity than HFR (Cisco newest switch)

11. Prof. Yoram Ofek - DIT - University of Trento, Italy 11 Research Area 1 Optical to Wireless Challenge  Optical/wireless bandwidth mismatch – factor of 1000 or more  UTC-based forwarding ensures:  Minimum delay buffering and delay  Deterministic no loss Bandwidth Mismatch Animation Bandwidth Mismatch Animation

12. Prof. Yoram Ofek - DIT - University of Trento, Italy 12 Ph.D. Thesis Opportunities in Research Area 1  The research activities in this area will be open for students to propose and define their activities since there are still many open issues with regard to performance and realization.  Blocking probability analysis of global IP network with time-driven switching and time-driven priority  Optics vs. electronics: qualitative and quantitative comparison of UTC-based pipeline forwarding  Bandwidth mismatch: streaming from optical to wireless  Network architecture design and performance evaluation of multi terabit/second IP packet switches  Performance study of time-driven “best-effort” traffic over IP networks

13. Prof. Yoram Ofek - DIT - University of Trento, Italy 13 Research Area 2 Computing and Networking  Fundamental inventions in the past 50 years:  Transistor  Laser & optical fiber  Lead to exponential growth in  Computing  Networking  Doubling performance every 18-24 mo. - a.k.a. Moore's "Law“  Led to the convergence of:  Computing and Networking

14. Prof. Yoram Ofek - DIT - University of Trento, Italy 14 Research Area 2 Computing and Networking  Today,  Internet: computing/networking are ONE  Thus, trusted operation is a major challenge  Critical to business:  Temporal (and spatial) manipulation of protocols  On-line bidding / stocks trading  DoS, DDoS – loss of revenues  Stealing: content, SW and information

15. Is it possible to use the network to alleviate networking trust problem?

16. Prof. Yoram Ofek - DIT - University of Trento, Italy 16 Research Area 2 The Context of TrustedFlow  A method that combines:  Computing and networking  For distribution of trust or entrusting  Stand-alone computer can be trusted, by definition !

17. Prof. Yoram Ofek - DIT - University of Trento, Italy 17 Research Area 2 What Do We Want to Achieve? Remote authentication of code during execution Trusted 1 st computer ensure that SW executed on Untrusted 2 nd computer was not modified Thereby, entrusting the 2nd computer

18. Prof. Yoram Ofek - DIT - University of Trento, Italy 18 Research Area 2 The Trust Problem Definition  How to ensure run-time SW authenticity  Focusing on two generic protocols:  1. Sending packets:  To avoid unfair usage/attacks on networks/servers  TCP, SLA, 802.11, …  2. Receiving packets (e.g., content):  To ensure digital right management (DRM)  Audio, video …

19. Prof. Yoram Ofek - DIT - University of Trento, Italy 19 Research Area 2 How: Entrusting  (Stream of) Tags are EMANATED from a program=code=software at run-time  ENTRUSTING by verifying the Tags Tags Trusted Computer UNtrusted Computer Entrusting NetworkingComputing

20. Prof. Yoram Ofek - DIT - University of Trento, Italy 20 Research Area 2 How: Entrusting Untrusted Computing Environment Trusted Computing Environment Tag Checker (TC) “HARDENED” with Special Hardware/Software (e.g., TCPA) (Stream of) Tags Associated with Periodic Replacement of SW parts & Parameters Trusted Computer Entrusted SW on UNtrusted Computer Entrusting SW by Verifying Tags SW with Tag Generator (TG) Sending Packets Receiving Packets

21. Prof. Yoram Ofek - DIT - University of Trento, Italy 21 Research Area 2 Quality of Remote SW Authentication with TrustedFlow TIME: periodic replacement of SW components and parameters SPACE: interlocking and Hiding Of Tag Generation BetterRemoteAuthenticationOperations

22. Prof. Yoram Ofek - DIT - University of Trento, Italy 22 Ph.D. Thesis Opportunities in Research Area 2  TrustedFlow for streaming media content protection and management  TrustedFlow for 802.11 MAC protocol  Reverse TrustedFlow for securing mobile users  TrustedFlow for Grid computing  TrustedFlow for Internet voting software  TrustedFlow for TCP (transport control protocol)  Design of the “hiding” function of the tag generator  Design of the “interlocking” function of the tag generator  TrustedFlow for SLA (service level agreement)

23. Prof. Yoram Ofek - DIT - University of Trento, Italy 23 Wireless Wi-Fi 802.11 Optical Backbone Security Revenue Generating Services Mismatch: Bandwidth / Availability / Reliability Secure Data and Content Flows B-GLOW High Fidelity Data Flows Over Wireless & Optical Backbone Research Area 3 Research Area 3 B-GLOW Overview Ubiquitous Computing, Storage, Position & Time

24. Prof. Yoram Ofek - DIT - University of Trento, Italy 24 Predictable High Quality Content Delivery via IP-based Wireless: IEEE 802.11 – Wi-Fi Optical Backbone Interfaces: GE/10GE/IP-MPLS/SDH/ … Scalable and adaptable Audio/Video Content Desired by Mobile Users Mobile Users with Handheld Devices Who Desired to Receive (and Pay for) Quality Content Research Area 3 B-GLOW Overview

25. Prof. Yoram Ofek - DIT - University of Trento, Italy 25 Predictable Payments for quality content protected that can be protected by DRM With scalable flat rate/mini/micro payments With multilingual user interfaces Predictable delivery of scalable content To provide scalable audio and video content with coding and representation with run-time SLA protocol and interfaces with UMTS To provide capacity to user to satisfy their need for content while minimizing infrastructure cost Optimized access point location with/without directional Antenna Efficient and timely packet flow and monitoring End-to-end Measurements with Feedback to Apps and Users Packet Transmission/Scheduling Protocol for Overcoming Optical/Wireless Bandwidth Mismatch Objective n. 4 Objective n. 3 Objective n. 2 Objective n. 1 IP-FLOW FOCUS Research Area 3 B-GLOW Overview

26. Prof. Yoram Ofek - DIT - University of Trento, Italy 26 USB 802.11 with External Antenna Omni/Directional 802.11 Access Point with/out SDMA USB 802.11 with External Antenna Omni/Directional Multimedia Server Streaming Media to Mobile Users Mobile Device PC/PDA/Phone Mobile Device PC/PDA/Phone Experimental Study: - “Hot-spot” and open-space - Utilizing both: - Network/protocol analyzer - RF spectrum analyzer Research Area 3 Focus on Wi-Fi Physical/Network Layer

27. Prof. Yoram Ofek - DIT - University of Trento, Italy 27 Antenna Gain: G(db) = 10 log 10 (g max ) = 10 log 10 [4* π *A / Lambda 2 ] g max  4*(3.14)*{(L1*L2)/(Lambda 2 )} [Lambda = speed-of-light/Frequency] [A=L1*L2 is the rectangular area of antenna aperture in cm 2 ] L1 L2 5GHz 6 cm  60  12 cm  30  G(db)  14 5GHz 6 cm  60  24 cm  15  G(db)  17 5GHz 6cm  60  48 cm  7.5  G(db)  20 5GHz 6 cm  60  G(db)  11 L2 Research Area 3 Focus on Directional Antennas 2 cm  60  4 cm  30  8 cm  15  16cm  7.5  2 cm  60  15GHz L1 L2 802.16 Smaller Directional Antennas But higher loss 802.11

28. Prof. Yoram Ofek - DIT - University of Trento, Italy 28 Research Area 3 Challenge  Small angle-high gain continuously changing best receiving direction, due to:  1. environment factors  2. motion - mobility  3. interference (unlicensed band)  Possible solution:  multiple sectors Antenna Sector 4 Antenna Sector 5 Antenna Sector 6 Antenna Sector 3 Antenna Sector 2 Antenna Sector 1Each Antenna Sector is Flat Panel with Multiple “Patches”

29. Prof. Yoram Ofek - DIT - University of Trento, Italy 29 Antenna Sector 1 Antenna Control Software USB 802.11 Access Control Adapter USB 802.11 Access Control Adapter Antenna Sector 6 USB HUB USB Connection Mobile Device Research Area 3 1st Experimental Prototype Only software

30. Prof. Yoram Ofek - DIT - University of Trento, Italy 30 Antenna Sector 1 Antenna Control Software USB 802.11 Access Control Adapter Antenna Sector 6 Mobile Device RF Switch USB Connection Research Area 3 2 nd Experimental Prototype with RF Switch

31. Prof. Yoram Ofek - DIT - University of Trento, Italy 31 Ph.D. Thesis Opportunities in Research Area 3  In-depth experimentation and measurement of content delivery capacity in “hot-spot” as function of the number of active mobile user – using high-end spectrum and network analyzers  Studying the inter-relationship between directional antenna and access point placement both analytically and experimentally  Design and analysis of the potential benefits of directional antennas for mobile users  Studying and analysing how to provide scalable and adaptive audio and video content with necessary coding and representation, including a run-time and adaptive service level agreement protocol for matching capabilities for (paid) quality content delivery to mobile devices  Studying and analysing how to provide scalable content with protection using DRM (digital right management) with capability for differentiated payments and micro-payments, that are dependent on the quality of content delivered and on prior classification of users

32. Prof. Yoram Ofek - DIT - University of Trento, Italy 32 Wireless Wi-Fi 802.11 Optical Backbone Security Services Mismatch: Bandwidth / Availability / Reliability Secure Data and Content Flows B-GLOW High Fidelity Data Flows Over Wireless & Optical Backbone Research Area 3 TrustedFlow Authentication of Data Flows Research Area 2 Research Area 1 Time-based Pipeline Forwarding of Data IP-FLOW Summary Ubiquitous Computing, Storage, Position & Time

33. Prof. Yoram Ofek - DIT - University of Trento, Italy 33 IP-FLOW Overview  Milestones 1 ST YEAR2 ND YEAR3 RD YEAR Research Area 1 PCB electronic/heat dissipation design Success  Implementation and testing Success  Integration and demonstration switch and interfaces Research Area 2 Implementation of TrustedFlow protocol Success  Integration for trusted content delivery Success  Integration and demonstration with trusted computing Research Area 3 Understanding mobile w/directional antenna Success  Evaluating content delivery quality Success  Integration and demonstration with micro-payment / DRM

34. Prof. Yoram Ofek - DIT - University of Trento, Italy 34 IP-FLOW Overview  PROJECT GANTT

35. Prof. Yoram Ofek - DIT - University of Trento, Italy 35 IP-FLOW Overview  Lecture series: Three graduate courses  Course 1: Satisfying QoS Requirements over a Single Networks  Course 2: Optical Networking  Course 3: The Inter-relationship of QoS, Mobility and Security  Seminar Series to Broad Audience  Seminar 1: User’s satisfaction and flow of packet over the Internet  Seminar 2: The use of time in future packet networks  Seminar 3: Trusted computing and run-time authentication of protocols  Seminar 4: Moore’s Law and the evolution of optical networking  Seminar 5: The inter-relationship of QoS, mobility and security

36. Prof. Yoram Ofek - DIT - University of Trento, Italy 36 IP-FLOW Overview  teaching, training, dissemination and Workshops 1 ST YEAR2 ND YEAR3 RD YEAR Graduate coursesCourse 1 / Course 2Course 3 / Course 2Course 1 / Course 3 Dissémination:Seminars: 1, 2, 3Seminars: 3, 4, 5Seminars: 1, 2, 3, 4, 5 Dissemination: Short tutorials Based on: Course 1 / Course 2 Based on: Course 3 / Course 2 Based on: Course 1 / Course 3 Workshops organized by the Chair holder on: Streaming media over optical and wireless - 5/1/06 Trusted computing and content protection - 11/1/06 Possible technology transfers: -Directional antennas for mobile users -TrustedFlow applications -Pipeline forwarding Switch