Скачать презентацию The Princeton EDGE Lab Opening Remarks Mung Chiang Скачать презентацию The Princeton EDGE Lab Opening Remarks Mung Chiang

1fd90f47bd6713ae959dd7d3038af318.ppt

  • Количество слайдов: 24

The Princeton EDGE Lab Opening Remarks Mung Chiang November 8, 2009 1 The Princeton EDGE Lab Opening Remarks Mung Chiang November 8, 2009 1

Welcome ¾ ¾ Informal, local event today Formal, global event is annual open house Welcome ¾ ¾ Informal, local event today Formal, global event is annual open house in August every year starting 2010 Current Status: ¾ Research projects ¾ Staff hiring ¾ Equipment purchase ¾ Publicity ¾ Outreach 2

Theory and Practice ¾ “There is nothing more practical than a good theory” “Theory Theory and Practice ¾ “There is nothing more practical than a good theory” “Theory as the brain and systems as the heart” “Rigor or relevance: either but not both? ” “Relevant question or elegant answer? ” … ¾ Why theory is inalienable: ¾ ¾ q q ¾ Explanatory (rather than descriptive) models Top-down design with predictive power Why theory is incomplete: q q Sensitivity to math crystallization Need to make a difference in live networks 3

Theory-Practice Divide Industry Tear Down This Wall ! Modeling Reality Model Transfer Theory Mathematics Theory-Practice Divide Industry Tear Down This Wall ! Modeling Reality Model Transfer Theory Mathematics 4

Theory-Practice Synergy Theory Bigger Overlap Practice New Questions Do them both 5 Theory-Practice Synergy Theory Bigger Overlap Practice New Questions Do them both 5

An Edge Between Theory and Practice ¾ From Dichotomy to Union: Practice Theory Practice An Edge Between Theory and Practice ¾ From Dichotomy to Union: Practice Theory Practice The Princeton EDGE Lab 6

Not Just Yet Another Testbed 1. Focus on edge networking (rather than backbone and Not Just Yet Another Testbed 1. Focus on edge networking (rather than backbone and sensor networks) 2. Theory-inspired implementation: ¾ ¾ Validate predictions Falsify assumptions Tighten characterizations Inspire new questions Technology transfers Innovation incubation Student education 7

History 2003 -2007: ¾ Many developments in theory: www. princeton. edu/~chiangm/publications. html ¾ Several History 2003 -2007: ¾ Many developments in theory: www. princeton. edu/~chiangm/publications. html ¾ Several tech transfers to industry ¾ EE-CS synergies ¾ ¾ Year 0: 2008 planning Year 1: 2009 fund raising and recruiting Research arm scenic. princeton. edu Lab arm scenic. princeton. edu/edgelab 8

Research -- Motivations Networked end users with access that is ¾ Universal ¾ (Politically) Research -- Motivations Networked end users with access that is ¾ Universal ¾ (Politically) Free ¾ Shared ¾ Scalable ¾ (Economically) Sustainable 9

Research -- Topics ¾ Social q q q ¾ Clean q q ¾ Content Research -- Topics ¾ Social q q q ¾ Clean q q ¾ Content distribution and sharing Content pipe divide Online social networks Internet for freedom Economic policies Price of greening Healthcare wireless delivery Edge q q End user oriented Access and home networks 10

Research -- Projects ¾ ¾ ¾ ¾ ¾ Network optimization Dynamics of information in Research -- Projects ¾ ¾ ¾ ¾ ¾ Network optimization Dynamics of information in networks via geometry Content aware networking P 2 P global streaming Utility Optimal CSMA Socio-Tech networking Sharing. Mart Freedom. Network economics Green IT 11

The Language is Math Methodologies: ¾ Distributed optimization ¾ Stochastic control ¾ Games and The Language is Math Methodologies: ¾ Distributed optimization ¾ Stochastic control ¾ Games and economics ¾ Graphs and random processes… Functionalities: ¾ Power control and scheduling ¾ Congestion control and routing ¾ Topology control and distribution ¾ Pricing… 12

Research -- Grand Challenges ¾ ¾ ¾ How to deal with nonconvexity? How to Research -- Grand Challenges ¾ ¾ ¾ How to deal with nonconvexity? How to be robust and opportunistic to dynamics? How to reduce high dimensionality? 13

Research -- Grand Challenges ¾ ¾ ¾ ¾ 1 Mbps global P 2 P Research -- Grand Challenges ¾ ¾ ¾ ¾ 1 Mbps global P 2 P streaming Near optimal scheduling without message passing Tunable control of content-pipe mutual awareness Universal broadband coverage Protect end user from central authorities Price user-generated content online Distribute and recover the price of greening 14

Research – Testbed/Data Partners ¾ VINI (Princeton) WAN in Lab (Caltech) TFA (Rice) Wi. Research – Testbed/Data Partners ¾ VINI (Princeton) WAN in Lab (Caltech) TFA (Rice) Wi. Mesh (KAIST) Fast. Mesh-SIP (HKUST) ¾ NECA ¾ ¾ 15

People – Group and Active Alumni ¾ ¾ ¾ ¾ ¾ Hazer Inaltekin Hongseok People – Group and Active Alumni ¾ ¾ ¾ ¾ ¾ Hazer Inaltekin Hongseok Kim Sangtae Ha Tian Lan Joe Jiang Yiannis Kamitsos Prashanth Hande Michael Yang Arel Lidow 16

People – Princeton Collaborators ¾ ¾ ¾ ¾ Rob Calderbank (EE, Math) Michael Freedman People – Princeton Collaborators ¾ ¾ ¾ ¾ Rob Calderbank (EE, Math) Michael Freedman (CS) Margaret Martonosi (EE) Vince Poor (EE) Jennifer Rexford (CS) Matthew Salganik (Sociology) Jacob Shapiro (WWS) 17

People – US Collaborators ¾ ¾ ¾ Nick Bambos Ed Knightly Xiaojun Lin Steven People – US Collaborators ¾ ¾ ¾ Nick Bambos Ed Knightly Xiaojun Lin Steven Low Urbashi Mitra Ashu Sabharwal Sanjay Shakkottai Ness Shroff R. Srikant Kevin Tang Junshan Zhang 18

People – International Collaborators ¾ ¾ ¾ ¾ Lachlan Andrew Gary Chan Song Chong People – International Collaborators ¾ ¾ ¾ ¾ Lachlan Andrew Gary Chan Song Chong Suhas Diggavi Jianwei Huang Mikael Johansson Zhu Li Soung Liew Marc Moonen Antonis Papachristodoulou Danny Tsang Chee Wei Tan Albert Wong Victor Wong Yung Yi Anglea Zhang 19

People – Industry Collaborators ¾ ¾ ¾ ¾ Phil Chou Bob Fry Victor Glass People – Industry Collaborators ¾ ¾ ¾ ¾ Phil Chou Bob Fry Victor Glass Jason Li Jin Li Linda Ness Alexandre Proutiere Mikael Prytz Sundeep Rangan Raj Savoor Sudipta Sengupta Siamek Sorroshyari Steve Sposato Xinzhou Wu 20

Sponsors ¾ ¾ NSF (Net. SE, CRI, FIND, Ne. TS) ONR (PECASE, YIP, DURIP) Sponsors ¾ ¾ NSF (Net. SE, CRI, FIND, Ne. TS) ONR (PECASE, YIP, DURIP) AFOSR (MURI, Complex Systems, STTR) Qualcomm ¾ Nokia Siemens Microsoft Telcordia AT&T ¾ Princeton Grand Challenge ¾ ¾ ¾ 21

Events ¾ ¾ ¾ Annual open house (approx. ) Weekly group meeting (approx. ) Events ¾ ¾ ¾ Annual open house (approx. ) Weekly group meeting (approx. ) Monthly EDGE Lab seminar talks 22

Future ¾ It’ll keep evolving in an open ended way 23 Future ¾ It’ll keep evolving in an open ended way 23

Today’s Agenda Part I: Samples from the Past ¾ Load Spillage Power Control (Prashanth) Today’s Agenda Part I: Samples from the Past ¾ Load Spillage Power Control (Prashanth) ¾ END Tool (Mike and Arel) Break Part II: Ongoing Ones ¾ P 2 P Global Streaming (Joe) ¾ Sharing Mart (Hazer) ¾ Utility Optimal CSMA (Mung) Part III: Future Plans (Hongseok and Sangtae) Discussion Dinner 24