The NeTS Program. Networking Research at NSF

Transcription

1 The NeTS Program Networking Research at NSF Dr. Joseph B. Evans Program Director Computer and Network Systems Computer & Information Science & Engineering National Science Foundation Networking Research Center Industry Day 2004 Penn State University 21 April 2004 State College, Pennsylvania

2 Context for NeTS

3 NeTS Goals More Impact from Our Research Better Service to Community CISE Reorganization New Program Improved Practices 17 May 2004 NeTS Program 3

4 CISE Reorganization Goals Realign divisions for coherence Cluster similar programs Increase productivity and efficiency for investigators and program officers Increase grant size and duration Sharpen focus and increase agility Develop themes for cross-cutting initiatives Integrate education and research Broaden participation Build on success of ITR 17 May 2004 NeTS Program 4

5 CISE Organization Office of the Assistant Director Freeman Crawford Division of Computing and Communication Foundations (CCF) Division of Computer and Network Systems (CNS) Division of Information and Intelligent Systems (IIS) Division of Shared Cyber Infrastructure (SCI) Abdali Andrews Pazzani Kim Anger Fisher Bainbridge Munoz Themes 17 May 2004 NeTS Program 5

6 CISE Organization Office of the Assistant Director Freeman Crawford Division of Computing and Communication Foundations (CCF) Division of Computer and Network Systems (CNS) Division of Information and Intelligent Systems (IIS) Division of Shared Cyber Infrastructure (SCI) Network Systems Computer Systems Computing Research Infrastructure Clusters 17 May 2004 NeTS Program 6

7 People Network Systems Cluster Program Directors Joe Evans, Admela Jukan, and Guru Parulkar Others Darleen Fisher, Ty Znati (part-time) Current activities New NeTS program within the cluster Networks Cluster Themes Cyber Trust 17 May 2004 NeTS Program 7

8 NeTS Program

9 NeTS Elements Three elements (this time) Broadly Defined Networking Research Focus Area 1: Programmable Wireless Networks Focus Area 2: Networking of Sensor Systems A well-coordinated realignment of networking to Sharpen focus and increase topical agility Increase impact Increase efficiency, e.g. grant size and duration Improve service to investigator Increase efficiency for program officers 17 May 2004 NeTS Program 9

10 NeTS Program Elements Embedded Systems, Distributed Systems Past Networking Programs Broadly Defined Networking Research Cyber Trust Networking of Sensor Systems Programmable Wireless Networks Communications 17 May 2004 NeTS Program 10

11 NeTS Broadly Defined Networking Research

12 Topics in Broadly Defined Area Network and Protocol Architectures Fundamental Understanding and Design Network Control and Management Innovative Networking Technologies Extensible Networks Strategic Research for the Internet 17 May 2004 NeTS Program 12

13 Broad Area Research Context Next Generation Networks Clean slate design Wide scale deployment in ten years Make appropriate assumptions about technologies, applications, etc. Internet Research Need compelling evidence for the target problems Solutions must Be integrated into existing protocol stack and systems Not break something else Too many projects Focused on Internet but do not meet basic requirements for success Hardly any projects on new paradigms and next generation 17 May 2004 NeTS Program 13

14 Next Generation Networking New Paradigms New Architecture Radical Innovation Strategic Evolution Approaches Testbeds Systems Theory Scalability Trustworthy Evolvability Autonomicity Ubiquity Capabilities Optical Wireless SoC Technologies Controllers Sensors 17 May 2004 NeTS Program 14

15 NeTS Focus Area Networking of Sensor Systems

16 Networking of Sensor Systems New Machines New Environments New Applications New Scale Billion to trillion devices! 17 May 2004 NeTS Program 16

17 The Technology Gap Monitoring & Managing Spaces and Things applications data mgmt service Networking of SES protocol system architecture MEMS sensing Comm. Store Proc urobots actuate technology Miniature, low-power connections to the physical world Power 17 May 2004 NeTS Program 17

18 NoSS Approach Deployed Infrastructure Applications Commercial Solutions Focus of This Program Close Coupling Required Without this progress will slow down significantly Reusable Systems & Science Foundations Research Experimental Systems Programmable Manageable Durable Secured Networks of SES to Enable Seamless Physical Webs 17 May 2004 NeTS Program 18

19 Expected Results Foundations research Theoretical, algorithms and systems Systems Several platforms; OS and protocols stacks; networked programming environments Many local testbeds with applications Education and training Graduate/undergraduate courses with experimental projects Many graduate students with hands-on experience Identifiable distinct community Members coming from sensors, embedded hardware, OS, networking, middleware, applications 17 May 2004 NeTS Program 19

20 NeTS Focus Area Programmable Wireless Networks

21 Recent Wireless Themes Applications End-user interest in mobile access to information Requirements of national defense & homeland security Recognition that spectrum is woefully under-utilized Proposals to dramatically re-architect the modes & mechanisms for using radio frequencies Policy Opportunity - Wireless Networks That Exploit Flexibility Advances in radio system engineering flexible radios are just becoming available Technology 17 May 2004 NeTS Program 21

22 State of Wireless Networking Wireless systems today Inflexible, wasteful static spectrum allocations Fixed radio functions Limited network and systems coordination Implications Proliferation of standards, such as Wi-Fi/802.11, Bluetooth, 3G, 4G, CDMA, GSM Encourages stovepipe architectures and services Discourages innovation and growth Source: FCC Source: Cisco Source: Linksys 17 May 2004 NeTS Program 22

23 Critical Problems Interference in unlicensed bands Underutilization in many other bands Source: B. Becker Source: M. McHenry 17 May 2004 NeTS Program 23

24 Evolution of Wireless Systems Receive Transmit Capture radio waves Generate radio waves Extract information from radio waves Implant information into radio waves Present information to user Accept information from user Circa 1900 Vacuum Tube Analog Processing Circuits Today s Systems Analog Processing Circuits Digitizer Digital Signal Processing Software Tomorrow Analog Processing Circuits Digitizer Digital Signal Processing Software Derived from version by R. Sternowski 17 May 2004 NeTS Program 24

25 Advances in Radios Software radio attributes and capabilities Wide operational frequency supports novel use of multiple bands Multiple waveforms in a single hardware unit provides interoperability Impact Dynamic spectrum management helps prevent interference Adaptable to local & current situation; flexible frequency use provides opportunities for quality of service Rapid deployment and service creation Enables new network architectures through flexible & dynamic connectivity Microphone Video Fax Data Narrowband A/D-D/A (Optional Integral Source Coding) N-/RT Software Programmable Processor(s)? Wideband A/D-D/A? Antenna Tightly Integrated Host Hardware Systems and networking issues remain unexplored and unexploited! 17 May 2004 NeTS Program 25

26 Programmable Wireless Networks Route messages through network and interoperate with larger Internet Dynamically and cooperatively manage spectrum resources Self-organize with rapid initial configuration Provide for mobility Support variety of network services Use adaptation to assure quality of service Support multiple users & domains 17 May 2004 NeTS Program 26

27 Research Area Evolution Programmable Wireless Networking Research Focus Deployed Infrastructure Applications Commercial Solutions Architectures, algorithms, & protocols for auto-configuration, topology, routing, management Reusable Systems & Science Required for accelerated progress Programmable wireless systems that provide improved and more efficient connectivity & services Foundations Research Experimental Systems Programmable wireless systems with radio & software capabilities for research in networking, management, and control Close coupling required 17 May 2004 NeTS Program 27

28 Programmable Wireless Focus Research Area Dynamic spectrum management architectures and techniques Topology discovery, optimization and network self-configuration Spectrum Coordinated Networks or SCN Changes Needs Architectures that are secure & robust, with quality of service and policy enforcement How to choose among possible topologies, and evaluate novel network architectures Interaction between routing, topology, and administration Source: H. Rajan Which approaches for diverse applications, communication modes, security & policy domains 17 May 2004 NeTS Program 28

29 Spectrum Resources The spectrum resource space consists of Frequency the radio frequencies used to carry a signal Time the time duration a signal is transmitted Space the volume over which the signal transmission is effectively communicated or causes interference Signal format the manner in which information is encoded on the radio frequency signal Frequency TDMA Space Time CDMA TV FM Source: G. Minden RF resources illustrating a few signals in time, frequency, and space 17 May 2004 NeTS Program 29

30 Dynamic Spectrum Management Architectures Broker-based and/or sensing-based Issues Implementing policies Secure and robust Quality of service Contract enforcement Evaluation and innovation needed Obtaining Frequencies for Programmable Wireless via Broker Obtaining Frequencies for Programmable Radio via Sensing 17 May 2004 NeTS Program 30

31 Towards Flexible Topologies Flexible topologies enabled by multiple simultaneous frequencies to multiple adjacent nodes Novel network architectures possible Significant service and performance improvements possible, including enhanced reliability and immunity Need to determine how to choose among possible topologies, and evaluate network architectures 17 May 2004 NeTS Program 31

32 Routing and Administration Flexibility creates substantial administration and security challenges Already have very hard problems to solve in traditional networks, and flexibility makes these worse Which associations are appropriate and trusted? Policy & its implementation will be critical BGP currently the only multi-domain option, with static policies tuned to particular peerings Needs to be more automatic & dynamic Example - separate public safety, cellular, hot-spot networks How do they talk? Under what circumstances? With what permissions? At what layer IP, upper and lower layers? Public Safety Hot-Spot Cellular 17 May 2004 NeTS Program 32

33 Some Desirable Radio Attributes Flexible in RF carrier frequency (~0-6 GHz) Flexible in bandwidth (several 10 s MHz) Flexible in waveform A/D and D/A driven, generated/processed by programmable DSP and/or FPGAs Dials to observe Traffic characteristics measured at network layer Error rate & characteristics (BER and distribution) MAC layer per packet error information Network and transport layer per flow correlations Receive characteristics Physical layer signal strength, interfering signals, background noise MAC layer transmit power, antenna in use Knobs to influence Physical layer Frequency & bandwidth Transmit power Beam width & direction Data rate, code, & chipping rate MAC protocol FEC strength Retransmit scheme MTU size Encryption & parameters Network layer Routing protocol Addressing plan(s) ACLs Interface framework needed to allow interoperation, with a flexible, usable set of scalable parameters 17 May 2004 NeTS Program 33

34 Impact of Programmable Wireless Networks Vastly improved connectivity Remote areas use more power and better frequencies where utilization relatively low Urban areas provide more capacity where utilization is relatively high More opportunities for networking devices such as sensors and controllers by providing capacity & adaptability More efficient use of a shared national resource 17 May 2004 NeTS Program 34

35 NeTS Logistics

36 Classes of Proposals Individual investigator and small group awards up to three years and average about $150,000 per investigator per year Large group awards up to four years and request up to $1,000,000 per year Facilities awards up to three years and average about $500,000 per year Workshops, planning grants, exploratory grants, etc. contact Program Director 17 May 2004 NeTS Program 36

37 Help from the Community Send your best ideas to NSF Consistent with focus & goals of the program We want high risk / high reward proposals Suggest and encourage good panelists who can do justice to the proposals and our focus Volunteer to be a reviewer and panelist 17 May 2004 NeTS Program 37

38 Contacts & Further Information Broadly Defined Networking Research Admela Jukan Networking of Sensor Systems Guru Parulkar Programmable Wireless Networks Joseph Evans 17 May 2004 NeTS Program 38

39 NeTS Status Update

40 Final Remarks Now is a time of excitement for networks and networking systems CISE reorganization has resulted in more sustained support of networking research Networking cluster is committed to Making a difference Serving research community and supporting visionary and far-reaching research 17 May 2004 NeTS Program 40

41 The NeTS Program Networking Research at NSF 17 May 2004 NeTS Program 41