Transcript ppt - UCI

CS234 – Advanced Topics in
Networking
Mondays, Wednesdays 2:00-3:20p.m.
DBH 1200
Prof. Nalini Venkatasubramanian
[email protected]
Course logistics and details

Course webpage: http://www.ics.uci.edu/~dsm/cs234



Lectures - MW 2:00-3:20p.m
Must read:



Must check it frequently
Collection of technical papers and reports by topic
Interesting article on how to read a paper
http://www.stanford.edu/class/cs244/2012/papers/HowtoRead
Paper
Reference books


Computer Networking: A Top-Down Approach (5th
Edition), by Jim Kurose and Keith Ross, Addison Wesley
2009. (preferred)
Multimedia Systems, by Ralf Steinmetz and Klara
Nahrstedt, Springer 2004.
Prerequisite knowledge

Requirements



Undergraduate level course in computer
networks.
Understanding of basic networking protocols is
desirable.
Basic programming skills in Java, C++…
Course logistics and details

Homework


Midterm examination



Paper summaries (4 sets)
Tentatively scheduled for end of Week 8
Team presentations on one topic
Course project or research paper

Maybe done in groups (preferred)




Project proposal due end of Week 3
Survey of related works due end of Week 6
Final project presentations/demos/reports – Finals week
Potential projects will be available on webpage
Grading policy

Homework - 20% of final grade



Midterm - 35% of final grade



Tentatively in Week 7
Team presentation - 10% of final grade


4 topic summaries.
Due the week after a specific topic is covered.
Due date: Based on course topic schedule
Class project/paper - 35% of final grade
Final assignment of grades will be based
on a curve.
Lecture Schedule
Part A : Topics in Wired Networks (Weeks 1-5)

Week 1,2 : Internet Technologies


Top-Down overview of networks
Internet: Application Layer Concepts/Protocols



Software Defined Networks
Week 3,4 : Peer-to-Peer Networks

Overview of P2P networks





Search/lookup; content delivery; storage/filesystems
Load Balancing, Latency, Throughput, Heterogeneity
Unstructured P2P – Gnutella, BitTorrent, KaZaa
Structured P2P – Chord, Pastry, CAN
Application Layer Multicasting




Web-HTTP, Email-SMTP, FTP, DNS, Web Caching
Basic Tree-based ALM - Locality, Load-balancing
Forest/Mesh-based ALM - Maximizing bandwidth utilization
Hybrid and Gossip-Based ALM – Reliability/perf w/ failure
Week 5 : Multimedia Networking



Quality of Service and Differentiated Services
Traffic Shaping, Rate control, Error Control
Potential Case Studies – Token Ring, FDDI, ATM
Lecture Schedule
Part B : Topics in Wireless Networking (Weeks 6-10)

Week 6: Infrastructure Based Wireless Networks


Cellular (GSM, CDMA, LTE.... ), Wireless LAN (802.11)
Week 7:Non-Infrastructure Based Wireless Networks

MANETs (mobile adhoc networks), Disruption Tolerant Networking

Week 8: Midterm Review and Midterm

Week 9, 10 : Mobile Pervasive Computing and Sensor Networks



Media Streaming, Power Awareness, Mobile Social Networks
Sensor Networks and Pervasive Networking – Zigbee, BlueTooth,
RFID
CyberPhysical Systems and Networking Technologies
Some tools for projects

Theoretical papers


Simulators


E.g. Quantitative/probabilistic/MATLAB analysis
E.g. NS2, Qualnet, Tossim, PeerSIM, CloudSIM,
Real testbed
E.g.Android apps, Cloud platforms (Amazon
EC2, Microsoft Azure,Google AppEngine, TinyOS,

ns2

Aim: Support networking research and education



For protocol design, traffic studies
Free open source - on Linux, Windows and Mac
Includes

Wired networks






Wireless networks



Routing Distance Vector (DV), Link States (LS)
Transportation: TCP and UDP
Traffic: Ftp, Telnet, Cbr (Constant bit ratess)
Queuing models: Drop-tail, RED
QoS support frameworks: IntServ and Diffserv
Ad hoc routing protocols (AODV, DSR) and mobile IP
Utilities for tracing and visualization
Details: http://www.isi.edu/nsnam/ns/
Qualnet

A good simulator for studying networks


Rapid prototyping of protocols with GUI tools and the
modular, layered stack design
Scalability via support for parallel execution

Includes

Many types of networks: Wired and wireless (WLANs, Cellul
ar, Ad Hoc)
 Various existing protocols
 Easy for tracing and visualization.
 More scalable than NS2 (support thousands of nodes)
www.scalable-networks.com/products/qualnet/

OverSim


http://www.oversim.org/
OverSim



Is a C++ based open-source overlay and peer-topeer network simulation framework for the OMNet++
simulation environment.
Built-in Modules : Structured (Chord, Pastry,
Bamboo, Koorde, Broose, Kademlia), Unstructured
(GIA) P2P and other Overlay Protocols (NICE, NTree,
Quon, Vast, Publish-Subscribe for MMOGs)
You can add a C++ module to run your own protocol
on OverSim
PeerSim


http://peersim.sourceforge.net/
PeerSim



Is a Java-based peer-to-peer system simulator
with two simulation engines (a cycle-based and
an event driven).
Built-in Modules: Pastry, Chord, Kademlia,
Skpnet, Bittorrent, TMan, Cloudcast.
You can build a java package of your own
protocol to run on PeerSim.
Mobile computing systems




Android apps
Android Emulators
Mobile + cloud platforms
Heterogeneous networking
Sensor systems

Sensor network platforms


Sensor network simulators



TinyOS
TOSSIM, extensions to NS2, QualNET to suppor
t Zigbee
RFID platforms and simulators
Robot simulation platforms to study mobile
sensing
Current research directions

Clean Slate Internet Design ( more in next lecture)



Emphasis on security, reliability, mobility, scalability
Note: Changing/upgrading Internet completely is hard/costly.
Better Access to the Internet



Conveniently (anywhere)
Fast (bandwidth)
Efficiently (low power consumption, billing cost)



Employ multiple networks (wireless and wired) and provide sea
mless vertical handover (switch to different networks seamless
ly).
Use clouds or multi-layer systems (e.g., use brokers between
Internet and users) to reduce latency or computation cost at
terminal devices.
Design solutions to optimize network resource allocation…
Growth of the Internet
Number of Hosts on the
Internet:

Aug. 1981
Oct. 1984
Dec. 1987
Oct. 1990
Oct. 1993
Apr. 1995
Jan. 1997
Jan. 1999
Jan. 2001
Jan. 2003
Jul 2004
Jul 2005
213
1,024
28,174
313,000
2,056,000
5,706,000
16,146,000
56,218,000
109,374,000
171,638,297
285,139,107
353,284,187
CSci5221:
Introduc
tion
16
People-to-Computer Ratio Over Time
From David Culler (Berkeley)
Systems Today
Principles of Operating Systems Lecture 1
18
Irvine Sensorium
Distributed Computing Envts.
Globus Grid Computing Toolkit
PlanetLab
Cloud Computing Offerings
Gnutella P2P Network
20
Requirements for Today’s Internet (“ilities”)

Availability and reliability


Quality-of-service (QoS) for applications


protect against malicious attacks, accountability of user actions?
Manageability



untethered access, mobile users, devices, …
Security (and Privacy?)


millions or more of users, devices, …
Mobility


fast response time, adequate quality for VoIP, IPTV, etc.
Scalability


“Always on”, fault-tolerant, fast recovery from failures, …
configure, operate and manage networks
trouble-shooting network problems
Flexibility, Extensibility, Evolvability, ……?


ease of new service creation and deployment?
evolvable to meet future needs?
CSci5221:
Internet Design
21
Always Best Connected(ABC)Access
Wired/Wireless
Always Best Connected(ABC)Access
Multiple Wireless
Improve throughput by
WiFi offload
Use Multi-hop networks
Increase coverage
WiZi-Cloud
Use Zigbee to turn on/off WiFi
interface at smartphone to
save energy – APs also have
two interfaces
MINA: A Multinetwork Information Architecture
Observe-Analyze-Adapt
1. Tier based overlay architecture
(Using Network centrality, clustering )
2. Heterogeneous Networks
and devices
3. Diverse services and
applications
Mobile cloud
Two tier mobile cloud system
Research Problems






Switch networks seamlessly with context
awareness
Optimize network resource usage
Rich data (multimedia) streaming
Network activity and power consumption
Provide information, notification (events)
on time
…
Sensor Networks


Sensor network platforms
Sensor network simulators

Extensions to NS2, QualNET, TOSSIM



http://www.cscjournals.org/csc/manuscript/
Journals/IJCN/volume2/Issue6/IJCN-72.pdf
Participatory Sensing
Crowdsensing
SAFIRENET – Next Generation MultiNetworks
Information need

Multitude of technologies


WiFi (infrastructure, ad-hoc), WS
N, UWB, mesh networks, DTN, zi
gbee
SAFIRE Data needs

Timeliness
Multiple network
s
NEEDS
DATA


Reliability


Resource Constraints




Dead Reckoning
(don’t send
Irrelevant data)
Video, imagery
Transmission Power, Coverage,
Failures and Unpredictability
Goal

31
accuracy levels needed for CO
monitoring
Limitations

Sensors
immediate medical triage to a FF
with significant CO exposure
Reliable delivery of data over un
predictable infrastructure
UCI I-Sensorium Infrastructure
Campus-wide infrastructure to instrument, experiments, monit
or, disaster drills & to validate technologies
sensing, communicating, storage & computing infrastructure
Software for real-time collection, analysis, and processing of sen
sor information
used to create real time information awareness & post-drill analy
sis
32
32
Mote Sensor Deployment
Heart Rate
Proprietary EMF t
ransmission
Polar T31 Heart rate st
rap transmitter
Inertial positioning
Polar Heart
Rate Module
Crossbow MIB510
Serial Gateway
Crossbow MDA 300CA
Data Acquisition board
on MICAz 2.4Ghz Mote
IMU (5 degrees o
f freedom)
IEEE 802.15.4 (zigbee)
To SAFIR
E Server
Carbon monoxide
Temperature, humidity
Carboxyhaemoglobin, light
33
UC Irvine Sensorium Boxes
(building on Caltech CSN project)
●
Humidity
●
●
Camera
●
Microphone / accelerometer
●
Microphone / light sensor
●
●
●
●
●
SheevaPlug computer
Accelerometer
Ethernet
Battery backup
Additional Sensors
●
Wi-Fi dongle, Smoke, Toxic gas
es (e.g. CO), Radiation, Humid
ity, Microphone, Camera
boiling pot, monitor pet's food and water, face rec
ognition
●
●
●
control (de)humidifer, particularly for individuals wi
th respiratory ailments
detect gunshot in an apartment building / complex
monitor thunderstorm activity
“Green” networks through energy harvesting
Multiple
Energy
Harvesting
sources
Energy harvesting
Prediction model
Network models Changing
Hierarchical
Application
topology
& Query model
Energy
Harvesting
WSN prototypes
SQuARES
Sensor
Irvine Sensorium
Decision
Maker
Server
Offline phase
Online phase
Advantages: Perpetual system lifetime, minimum
maintenance cost
Challenges: Temporal and Spatial variations in
energy harvesting sources
Solution: Smart and adaptive energy harvesting
management schemes, exploiting applications’
tolerance to quality degradation to manage
system energy consumption and satisfy energy
harvesting constraints
Sustainability and CPS

Smart campus

The Software Defined Building?
Integrating sensing with smart grid

Smart water systems (opportunity)?



Integrating Evs
Simulation, Emulation F/ws