Computer Science & Computer Engineering Department Everything - - PowerPoint PPT Presentation

Computer Science & Computer Engineering Department

Everything is Alive Agent System & RFID Natural Language Interfaces Database on a Grid

• Dr. Craig Thompson

http://csce.uark.edu/~cwt

Context: Knowledge Cycle

… NSF, DARPA, … University Research

Funded projects Class projects Special projects

Teaching Wal-Mart Acxiom

problems solutions new hires, interns, coops, supported students Students learn the basics Students learn the latest ideas and how to solve problems

Everything is Alive Agent System & RFID Enterprise Integration Architecture

Everything is Alive

I see a tank! Need fuel!

• bservations &

recommendations

• rders &

subscriptions Any threats?

An RFID Tag

RFID Reader

R R R R R R R R R R R R

Conveyor belt Storage Loading dock Pallets

R R R R = RFID Reader

Camera

Distribution Center

Pallet/Container/Item

Reader2 Conveyor Belt Tag Printer Reader1 Camera GUI Dashboard DBMS

…

XML/WSDL messages sent between “agents” agent wrappers

• Uses a Service Oriented Architecture
• Compatible with WSDL

…

XML Commands XML Data Device Wrappers

Natural Language Interfaces

Basic MBNLI

User interface Including menus and answers Predictive Parser Including translations

Grammar & Lexicon

Target System e.g. a DBMS or resource

Other info e.g. DBMS schema

S NP VP Light – turn on

DBMS on a Grid

HASHi TREEj … … … …

Indexing Layer Queryi

Query Tree

consists of relational algebra

• perators like join, project, select

but also stored procedure calls

Query Execution

results in tables

• ptimized

translated to

…

Disk-based Storage Layer Query Layer

Relational DBMS Architecture Data Grid Architecture … HASHj TREEk … … … …

Indexing Layer Grid-node-based Storage Layer Workflowi

scheduled

Workflow Execution

results in data sources

Workflow Layer

automated

BACKUP

What is a Grid?

• A grid or cluster is a collection of computers that work together to solve a

problem

• Types of grids

– Computational grids – reducing execution time – Data grids – large scale data management problems

Why Use a Grid?

• Commodity processors - Custom supercomputer are expensive
• Scalable - Cheap storage and large datasets
• Take advantage of parallelism

DBMS on a Grid

• In certain high-end data-centric applications, practitioners are discovering

that traditional relational database technology is not meeting their requirements for huge data sets, high transaction throughput, cheap storage, and flexible workflow support – so they are switching to grid

• technology. Can we have our cake and eat it too? Do we have to give up

relational capabilities to get grid benefits?

Web Page

User interface Grammar & Lexicon

Web Page

User interface MBNLI I/F descriptor

Web-enabled MBNLI

Parser Farm(s) Target System(s) e.g. a DBMS or resource