Course Project Book (Mobile Computing and Wireless Networks) CS 395T - PowerPoint PPT Presentation

1 / 13
About This Presentation
Title:

Course Project Book (Mobile Computing and Wireless Networks) CS 395T

Description:

... Session Migration Intelligent Routing Decisions in AODV Security in Ad-hoc Networks Secure Group Management in ... Protocol for MANET Comparison of Energy ... – PowerPoint PPT presentation

Number of Views:42
Avg rating:3.0/5.0
Slides: 14
Provided by: Yonggua6
Category:

less

Transcript and Presenter's Notes

Title: Course Project Book (Mobile Computing and Wireless Networks) CS 395T


1
Course Project Book(Mobile Computing and
Wireless Networks)CS 395T Spring 2002
  • Course Instructor
  • Dr. Yongguang Zhang (ygz_at_cs.utexas.edu)
  • Course URL http//www.cs.utexas.edu/users/ygz/395
    T-02S
  • May 17, 2002

2
  • Project List
  • Shruti Session Migration
  • Intelligent Routing Decisions in AODV
  • Security in Ad-hoc Networks
  • Secure Group Management in Ad-hoc Network
  • UBQuards A Smartcard Project
  • Smart Cards Career Fair Experience
  • Content Based Routing on Ad-Hoc Networks
  • A New Service Discovery Protocol for MANET
  • Comparison of Energy-Aware Ad-hoc Routing
  • Multi-Mode TCP

3
Shruti Session Migration
  • Challenges Faced, Lessons Learnt
  • Modifications to legacy applications
  • Maintaining portability
  • Implementation Integration
  • 32-bit IP insufficient for unique PK abstraction
    that combines security and authentication
  • Pre-publish or perish
  • http//www.cs.cmu.edu/softagents/migsock/MIGSOCK-
    ThesisPresentation.ppt

Application
App State
Network level
To other end host
  • Design/Approach
  • Application layer
  • Application API to get/set application state
  • Session Manager daemon
  • Provides interaction between Home Agent and all
    user state
  • Home Agent
  • Stores the users session state
  • Socket Migrator
  • Uses IP redirection to offer same abstraction
  • Transfers TCP state
  • Future work
  • Migrate
  • Other applications (mp3, ram)
  • User Mode Linux
  • Bosch virtual machine simulator
  • Application layer
  • Use condor style-checkpointing
  • Create new language abstractions
  • Home agent using smartcard

Project teamAmit, Amol, Ravi, Deepak, Divya,
Young-Ri URL http//www.cs.utexas.edu/users/amitj
i/utOnly/Shruti.html
4
Intelligent Routing Decisions in AODV
  • New Ideas
  • Using signal strength values to predict the link
    life time and hence stability of a route
  • Choose more stable routes
  • Maintaining multiple paths for a destination
  • Decreases route discovery latency
  • Reduce the effect of route failures
  • Key Design/Approach
  • Figure out how to measure life time of a link
  • Use Friis law to predict the relative position of
    a neighbouring node
  • Keep a secondary routing table
  • If primary route fails, use secondary route
  • If both routes fail, initiate route discovery
  • Implementation Issues
  • Simulated these extensions on ns-2
  • Experimented with varying parameters
  • Optimum Beaconing Interval
  • Dynamically adjust wait time at destination
  • Performance comparison with base AODV
  • Used Different mobility patterns of pause time
    and speeds

Results/Lesson Learned
Project team Gokhan Uluderya, Jared G. Boone
Ashish Ahuja URL http//www.cs.utexas.edu/users/u
luderya/MWNET/project.html
5
Security in Ad-hoc Networks
  • Algorithms Implemented
  • Password-authenticated group key establishment
  • Threshold cryptography shared secret division by
    trusted party, then recollection by clients
  • Paper Themes Surveyed
  • Group key establishment
  • Password-authenticated
  • Unauthenticated
  • Threshold cryptography
  • Secret sharing
  • Distributed cryptographic operations
  • Web of trust
  • Introducers
  • Key Design/Approach
  • Provide a broad reading list for research in the
    area
  • Leverage existing work in security for basis of
    solutions
  • Adapt existing algorithms to ad-hoc networks
  • Utilize existing toolkits for cryptographic
    algorithms
  • Results/Lesson Learned
  • Implemented in Java
  • Used Java Cryptographic Extensions (JCE)
  • Suns JCE provider for Triple-DES
  • BouncyCastle.orgs for AES and RSA
  • Lesson learned
  • Many existing security solutions are unsuitable
    due to lack of on-line server or predefined
    hierarchy
  • We must assume a prior context

Project team Kevin Kane URL http//www.cs.utexas
.edu/users/kane/395T-02S/
6
Secure Group Management in Ad-hoc Network
  • Motivation
  • Security in ad hoc networks is a necessity.
  • Not many simulations or implementations for
    security have been done in ad hoc networks.
  • Centralized approach is not suitable.
  • Investigate a distributed solution for key
    management.
  • Architecture
  • Works on top of existing routing protocols.

Simulation Results
  • Conclusion
  • Has clear advantage in terms of connection time
    against the centralized approach.
  • Need further optimization to adapt frequent
    topology changes.

Project team Chun-Chi Chen, Edwin Ng, Chwan-Ming
Wang URL http//www.cs.utexas.edu/users/ccchen/cl
asses/cs395t-mwc/proposal.html
7
UBQuards A Smartcard Project
  • New Ideas
  • Specification for using smartcards as the medium
    for ubiquitous web services
  • User state information on card
  • stateless web service servers
  • Web service workflow description
  • Hardware/Software specification for universal
    communication module

CM
Cardlet
Device
Web Service
  • Key Design/Approach
  • Web service language interpreter
  • Run as a Cardlet
  • 32K Javacard, Stripped down JVM
  • 1 input, 1 output instruction set
  • only integer types
  • control flow using predicates
  • Communication Module
  • Consumer Device
  • Web Services
  • Microsoft.NET Platform, SOAP
  • Results/Lesson Learned
  • Proof of Concept
  • Smartcards Ubiquitous
  • Software Implementation
  • Ubiquitous Communication Module
  • USB, HTTP/SOAP, WFDL
  • Smartcards Computing Power Increasing
  • Example Scenario
  • Plug-n-Play Internet Capability
  • Communication, Storage, device stuff completely
    detached.

Project team Changkyu Kim, Karu Sankaralingam,
Youngin Shin URL http//www.cs.utexas.edu/users/k
aru/mobile/
8
Smart Cards Career Fair Experience
  • New Ideas
  • Current literature focuses on the security aspect
    of smart cards
  • Exploit smartcards for personalization
  • Saves effort and more efficient and easier for
    students and recruiters
  • Recruiters anyway have to scan the resume into
    the database
  • Design/Approach
  • VB front end for sending/receiving resume
  • Resume and BizCard in XML format
  • Standardized format and parsing
  • Applet on smart card has a simple parser
  • Students have an interface to personalize the
    data they want from recruiter
  • Implementation platform
  • SLB 4.1 toolkit and REFLEX 72 card reader
  • Metrowerks CodeWarrior IDE for debugging and
    simulator on Windows 2000 platform
  • Results and Future work
  • Challenging to program with limited data types
    and memory
  • Developed prototype to demonstrate simple
    processing on and off the card.
  • Smart cards still not ready
  • Memory too low, (use compression on host side),
    garbage collection (Java Card)
  • Tool kit unstable and Simulator not very reliable
  • Few data types, array of objects or 2-D arrays
    not available

Project team Kaushik Lakshmanan and Ramanathan
Pallassana
9
Content Based Routing on Ad-Hoc Networks
  • Motivation
  • Peer-to-peer applications content driven
  • No centralized DNS for name-lookup in MANETs
  • Current protocols have to decouple steps of name
    to address mapping and route discovery
  • Combine! -- Content based routing
  • Key Ideas/Design
  • Nodes maintain information if they are on a path
    towards content
  • Communication local broadcast
  • Associate a path-id with every path discovered
  • Eliminate IP-Address/node-id
  • For optimization use pseudo random local-id
  • Distance metric hop count
  • Results/Lesson Learned
  • Implemented in NS2
  • Comparison
  • Plain flooding vs Our Content-Routing
  • Metric
  • Traffic over 80 reduction
  • Route cache hit rate about 27
  • Control traffic negligible
  • Lesson learned
  • Think out of the box
  • challenge to implement a whole new protocol in ns

Project team Vivek Subramanian Amish
Gandhi URL http//www.cs.utexas.edu/users/viveksu
b/mobile_proj.html or http//www.schmoogle.com
10
A New Service Discovery Protocol for MANET
  • New Ideas
  • Distributed directory
  • Cache optimizations (ala DSR, AODV)
  • Periodic Broadcast based Service Advertisement.
  • Transparent to underlying routing protocols
  • Each node serves as a Service Agent, Directory
    agent and also User agent

Master
Application
MANET
Agent API
Service Agent
Slave
AdHoc Tool Kit Environment
  • Motivation
  • Service Discovery will become a crucial need for
    upcoming networks/devices for seamless
    interoperability
  • Existing Protocols
  • JINI, UPNP, SLP, Salutation, etc..
  • Existing protocols are unsuitable for Mobile
    AdHoc Networks
  • We Propose and implement a new SDP for MANETs.
  • Results/Lessons Learned
  • File Sharing Application implemented in java
  • Tested using AdHoc toolkit
  • Used Socket Library extension to emulate the
    mobile environment
  • Lessons learned
  • Using AdHoc toolkit
  • Various SDPs

Project team Krupakar V. Pasupuleti, Jisun Park
and Subramanyam Mallela URL http//www.cs.utexas.
edu/users/manyam/cosd.htm
11
Comparison of Energy-Aware Ad-hoc Routing
  • Motivation
  • Important to ensure battery-operated mobile nodes
    operate as long as possible.
  • Several solutions proposed and compared with
    non-energy-aware protocols.
  • Our goal is to compare different solutions under
    DSR using same parameters.
  • Algorithm
  • Span Only coordinator nodes stay awake. The
    non-coordinator nodes go to sleep.
  • Afeca Sleep depending on the number of
    transmitting neighbors.
  • FA Choose the minimum cost routing path.
  • DSR
  • Simulation Setup
  • Use the ns-2 simulator.
  • Parameters traffic load (number of source nodes
    and traffic rate), mobility (pause time, speed).
  • Matrices Packet delivery rate vs. traffic load,
    packet loss rate vs. mobility, network lifetime
    vs. traffic load, network lifetime vs. mobility,
    nodes survival percentage vs. time.

Results/Lesson Learned
Project team Pisai Setthawong, Hari Shankar,
Yihong Zhou URL http//www.ece.utexas.edu/pisai/
cs395t
12
Multi-Mode TCP
  • MM-TCP
  • When the bandwidth of bottlenet changes
    frequently, TCP can save a subset of current TCP
    states (mode).
  • When TCP return to a history network enviorment,
    it can use the saved mode to set some initial
    values, such as cwnd.
  • TCP can adapt the enviorment quickly and then
    improve the performance.

...
...
Experiment Result
  • Conclusion/Lesson Learned
  • Implemented in ns2
  • Simulation results in ns2
  • There are some improvements when TCP switches
    between different network enviorment. But the
    improvement is not large enough.
  • Lesson learned
  • Implementation of TCP stack
  • Simulation an evaluation

Project team Yi Li URL http//www.cs.utexas.edu/
users/ylee/project.html
13
  • Contact
  • Prof. Yongguang Zhang
  • Dept. of Computer Sciences
  • University of Texas at Austin
  • Austin, Texas 78712, U.S.A.
  • E-mail ygz_at_cs.utexas.edu
  • Phone 512-232-7889
Write a Comment
User Comments (0)
About PowerShow.com