Rutgers,%20The%20State%20University%20of%20New%20Jersey

1
New Jersey Digital Government SummitInternet2
and the Future of the Internet - Wireless

• Rutgers, The State University of New Jersey
• www.winlab.rutgers.edu
• Contact Ivan Seskar, Associate Director
• seskar at winlab . rutgers . edu

2
Introduction Wireless as the key driver for the
future Internet

• Historic shift from PCs to mobile computing and
  embedded devices
• gt2.5 B cell phones vs. 600M Internet-connected
  PCs in 2006
• gt500M cell phones worldwide with IP service,
  rising rapidly
• Cellular data devices serve as primary access to
  Internet in India and China
• Sensor deployment just starting, with some
  estimates 5-10B units by 2015

750M servers/PCs, gt1B laptops, PDAs, cell
phones, sensors
500M server/PCs, 100M laptops/PDAs
Wireless Edge Network
INTERNET
INTERNET
Wireless Edge Network
2005
2010
3
Network Architecture Evolution
MSC
Internet (IP-based ? clean slate)
Public Switched Network (PSTN)
Mobile network overlays, etc.
Seamless Internet extension or just a local
network??
Custom Mobile Infrastructure (e.g. GSM, 3G)
BSC
Open BTS
WLAN Access Point
Infostation cache
BTS
Dynamic Spectrum Reuse
WLAN Hot-Spot
VOIP
Ad-hoc network extension
CDMA, GSM or 3G radio access network
VOIP (dual-mode)
Broadband Media cluster (e.g. UWB or MIMO)
Low-tier clusters (e.g. low power 802.11 sensor)
Today
Future?
4
Wireless Testbeds

• Motivated by
• cost time needed to develop experimental
  prototypes
• need for reproducible protocol evaluations
• large-scale system studies (...emergent behavior)
• growing importance of cross-layer protocol
  studies
• creation of communities for wireless network
  research

5
Orbit

• ORBIT open-access multi-user facility for
  experimental wireless networking research
  primarily in unlicensed bands
• 24/7 service facility with remote access
• open interfaces for flexible layer 2,3
  cross-layer protocols
• extensive measurements at PHY, MAC and Net layers
• support for wide range of radio system scenarios

6
ORBIT Radio Grid (Phase I)

• ORBIT testbed currently supports experimentation
  with up to 400 nodes (both end-points and
  routers)
• Heterogeneous radio environment (802.11,
  Bluetooth, ZigBee, GNU Radio, etc.) gt 1000
  radios in a single testing domain
• Integration with wired network testbeds available
  (PlanetLab, VINI)

Current ORBIT sandbox with GNU radio
400-node Radio Grid Facility at WINLAB Tech Center
Planned upgrade (2007-08)
Radio Mapping Concept for ORBIT Emulator
URSP2 CR board
Programmable ORBIT radio node
7
ORBIT Field Trial (Phase II)
3G Coverage Area
802.11 Access Points / Radio Routers
3G Base Station
RU BUS Route (Lines A H)
8
State of ORBIT

• Released for general use in Oct 2005
• Currently consists of 436 nodes, 26 servers, and
  48 Ethernet switches
• Main emulator grid with 400 nodes
• 9 sandboxes with 2 nodes each
• Outdoor network with 10 fixed and 5 mobile nodes
• gt260 experimenters from 64 organizations
• gt12000 total reservations/year

9
ORBIT Resource Utilization
Main grid utilization of 89 with average 10.4
(2 hrs.) slots/day occupancy
10
Future Internet
11
Why Future Internet?

• Future Internet research motivated by three
  distinct factors
• Limitations of current Internet protocols
  (security, QoS, mobility, )
• New core technologies (storage, wireless,
  virtualization, )
• Emerging new applications (content delivery,
  vehicular networks, pervasive systems,..)

• Outputs
• Novel network/protocol concepts
• Future Internet architecture ideas
• Influence on current IP standards
• Approaches to cellular-Internet convergence
• Possible clean-slate deployments
• New classes of applications
• etc.

Limitations of Current Internet
Clean-Slate Internet Research
New Core Technologies for Networking
Emerging Networked Applications
12
Research Programs

• Broad categories of programs on future Internet
• Ongoing corporate RD and academic projects on
  new networking technologies (wireless, optical,
  programmable networks, etc.) and applications
  (sensors, vehicular, etc.)
• Mainstream industry efforts on Internet evolution
  and new services (IRTF, VOIP/SIP, video services,
  content overlays, )
• New clean-slate initiatives, mainly supported
  by NSF FIND and GENI programs, possibly expanding
  to include other agencies

13
FIND (Future Internet Design)

• Started by NSF in 2006
• NSF NeTS research program on clean-slate
  architectures and protocols
• Over 50 funded projects ranging from security to
  virtualized networks to cognitive radio (..mostly
  at academic institutions)
• Duration for first phase of projects (concepts,
  initial evaluation) expected to be 2006-08/09
• Second phase (converged architectures,
  large-scale implementation, real-users) expected
  to start 2009

14
Emerging Wireless Scenarios
15
4G Cellular and WiMax

• Next-generation of cellular technology aimed at
  2015
• Higher radio access speeds 100 Mbps using MIMO
  and OFDMA technologies
• Decentralized control of radio resources
• Support for inter BTS mesh networks, etc.
• WiMax or cellular BSR has potential for lower
  cost, commodity equipment model
• Simplifications to cellular architecture ? flat
  network of IP base stations
• Ideally, plugs into future IP network with
  integrated mobility support

16
Ad-Hoc Mesh Networks

• Multi-hop radio (ad hoc, mesh, vehicular, sensor)
  technologies now entering the mainstream
• Leverages Moores law cost/performance gains of
  commodity radios such as IEEE 802.x
• Distributed solution with short-range radios will
  eventually outperform centralized (cellular) ?
  analogous to PC/mainframe evolution
• Involves new routing discovery protocols
• Interactions between lower layers (PHY, MAC) and
  routing in dense deployments
• Problems with TCP end-to-end model due to
  changing BW and channel quality

Wired Internet Infrastructure
Mesh GW or AP
Mesh Router
Hierarchical Mesh Network
The 49 Mesh Router from Meraki Networks
17
Dynamic Spectrum Cognitive Radio

• New techniques for spectrum coexistence needed as
  radio density increases
• Smart radios with fast scan, agility, etiquette
  under consideration by FCC
• Dynamic adaptation of radio (PHY/MAC)
  implications for networking

Data Signal
Spectrum Policy Server
Spectrum Coordination
Dynamic Spectrum Protocols For Coordination
Wired Internet
Next-gen wireless devices with dynamic spectrum
capability (fast RF scan, agile, adaptive PHY/MAC)
18
P2P and DTN

• P2P and DTN network protocols expected to migrate
  from niche scenarios to wider usage
• Router mobility
• Network may be disconnected at times delay
  tolerant protocols
• Caching and opportunistic data delivery .
  In-network storage
• Content- and location- aware protocols

Internet
Mobile DTN Router
Opportunistic High-Speed Link (MB/s)
Ad-Hoc Network
Mobile DTN Router
Roadway Sensors
Static DTN Router
Mobile P2P User
19
Sensor Nets
Pervasive Application Agents
Compute Storage Servers
User interfaces for information control
Mobile Internet (IP-based)
Sensor net/IP gateway
Overlay Sensor Network Infrastructure
3G/4G BTS
GW
ZigBee, UWB, etc.
Relay Node
Sensor/ Actuator
Ad-Hoc Sensor Net A
Ad-Hoc Sensor Net B

• Sensor net scenarios involve
• Large scale
• Limited CPU speed and transmit power
• Intermittent connectivity, low-speeds, ad-hoc
  modes
• Location and content-awareness
• May involve closed loop control in real-time

Virtualized Physical World Object or Event
20
Vehicular Networks

• Vehicle safety and information/convenience
• Potentially high density
• Networking involves location awareness
• Ad hoc network formation and disconnections
• Network (group) mobility
• V2V and V2I modes

Irrelevant vehicles in radio range for few
seconds
Following vehicle, in radio range for minutes
Passing vehicle, in radio range for tens of
seconds
Desired message delivery zone
(Idealized) Broadcast range
Projects at UCLA, Rutgers WINLAB, UC Berkeley, ..
21
Integrating the Physical World with the Internet
Ambient interfaces
Application Management Control Software
Human in the Loop
Global Pervasive Network (Future Internet)
Computation Storage
To Actuators
Protocol module
Content Location Aware Routers
Network Connectivity Computation
Hospital with Embedded Monitoring
Smart Public Space
Vehicles with Sensors Wireless
Virtualized physical world object
From Sensors
Multiple radio standards, ? Cognitive radios
Autonomous Wireless Clusters (ecosystems)
Robotics Application
22
GENI Implementation Wireless Subnets Overall
Wireless Deployment Plan

• Five types of experimental wireless networks
  planned necessary to support full range of
  protocol research and to enable new applications
• 1. Wireless emulation and simulation (repeatable
  protocol validations)
• 2. Urban 802.11-based mesh/ad-hoc network
  (real-world networking experience with emerging
  short-range radios)
• 3. Wide-area suburban network with both 3G/WiMax
  (wide area) and 802.11 radios
• 4. Sensor networks (application specific,
  specific system TBD via proposal process may
  include environmental, vehicular, smart spaces,
  etc.)
• 5. Cognitive radio network advanced technology
  demonstrator (adaptive, spectrum efficient
  networks using emerging CR platforms)
• also some common network facilities such as
  location dynamic binding services
• Each network at a different geographic location
  new spectrum allocation may be needed at some
  sites

23
Web Sites for More Information

• GENI www.geni.net
• FIND www.nets-find.net
• ORBIT www.orbit-lab.org