Using Behavior Templates To Design Remotely Executing Agents For - - PowerPoint PPT Presentation

▶

Mar 26, 2024 221 likes •420 views

Using Behavior Templates To Design Remotely Executing Agents For Wireless Clients Eugene Hung and Joseph Pasquale Dept. of Computer Science and Engineering University of California, San Diego Motivating Scenario Problem Wireless clients

SLIDE 1

Using Behavior Templates To Design Remotely Executing Agents For Wireless Clients

Eugene Hung and Joseph Pasquale

Dept. of Computer Science and Engineering

University of California, San Diego

SLIDE 2

Motivating Scenario

SLIDE 3

Problem

Wireless clients are diverse, varying in:

Network Bandwidth Network Reliability Graphical Display User requirements

Web services are not flexibly customizable

Scenario: Wireless E-Commerce

SLIDE 4

Outline

Previous Approaches Design Goals Solution – ReAgents Architecture Behavior Library Experiments Conclusions

SLIDE 5

Previous Approaches

Network-based

Active Networks

Server-based

WAP (Wireless Application Protocol)

Intermediary-based

Web Proxies Mobile Agents

SLIDE 6

Design Goals

The ideal customization solution will be:

Flexible enough to handle user needs Transparent to servers (deployable) Easy to program and understand Efficient when used

SLIDE 7

Solution: ReAgents

ReAgents – Remotely Executing Agents

Contain Customizing Logic (CL) “One-shot” mobility to ReAgent host Behavior-based development

SLIDE 8

Behavior: Filter

Reduces server data to client specifications Customizing Logic: Data-reducing algorithm Sample Application: Low-bandwidth filtering

SLIDE 9

Behavior: Encoder

Transforms data for reverse-transform at client Customizing Logic: Reversible data transformation Sample Application: Encrypted transfer for privacy

SLIDE 10

Behavior: Monitor

Polls object on server until desired state is reached,

then reacts to state change

Customizing Logic: Object state test and reaction Sample Application: Custom stock trader

SLIDE 11

Behavior: Cacher

Bypasses server communication by storing

frequently accessed server data close to client

Customizing Logic: Cache management policy Sample Application: Resource-poor client caching

SLIDE 12

Behavior: Collator

Sends same request to many servers and merges results Customizing Logic: Results-collation algorithm Sample Application: Shopping comparison agent

SLIDE 13

Experiment

File transfer time reduced 30-75%

SLIDE 14

Experimental Results

ReAgent overhead is low Overhead scales well as file size increases

50000 100000 150000 200000 220 600 1560 3430 File size (kilobytes) Response time (ms) Total UnCmp Total Cmp R->C R S->R Cmp

SLIDE 15

Conclusion

ReAgents customize for wireless clients

Flexibly

Customizing logic

Transparently

Server is bypassed

Easily

One-shot mobility simplifies security and semantics Behaviors provide structured, patterned development

Efficiently

Results show good performance and scalable overhead

SLIDE 16

Questions?

SLIDE 17

SLIDE 18

ReAgent Architecture

SLIDE 19

Usage

ReAgent created by chaining Behaviors Behaviors created by instantiating with CL Example: Custom Stock Trader

ReAgent reagent = new ReAgent(); Behavior m = new Behavior (“Monitor”, “MyPriceWatch.class”); Behavior t = new Behavior (“Filter”, null); reagent.addBehavior (m, null); (no converter for monitor) reagent.addBehavior (t, “GenerateStockBuyRequest.class”); reagent.launch(“middleman.org”); reagent.process(“GET http://stock.org/viewprice.cgi/?p=GOGL”);