ISM 50 - Business Information Systems · PDF fileISM 50 - Business Information Systems Lecture 16 ... Jiawei Han ©2006 Jiawei Han and Micheline Kamber, ... (OLTP vs. OLAP):

ISM 50 - Business Information Systems

Lecture 16

Instructor: Magdalini EirinakiUC Santa CruzMay 24, 2007

Outline

Announcements MySQL case (cont’d) Student Presentation - FedEx Student Presentation - UPS OLTP OLAP

Announcements Reading for 5/29

Messerchmitt 11.2 (pp.333-335) Messerchmitt 18.1 (pp.498 - 505) Akamai Case (reader pp.217-236)

(Read book first)

Forthcoming Student Presentations 5/29

Evan Price (business paper) Matthew Payne (Akamai case)

5/31 Eric Gonzalez (business paper: WalMart) Derek Stern (business paper: Home Depot)

News Folio #3 due next Thursday (5/31)

mySQL Case

mySQLWhat does mySQL make?

How Successful is mySQL? Visibility: Fortune magazine, more mentions on www Reaction from giants Revenue growth 2001 700k, 2002 6.2m, 2003 10m Good performance reviews Recent SAP alliance But Market share tiny:

$10 million out of $10 billion market!

Why Success? Good Technology Large DBMS bloated with features most don’t need Innovative OSS model

mySQLHow does OSS work?

Two Types of License: GPL

Free No Support Any software that uses MySQL as a module must itself be

made GPL Commercial License

Support Could be distributed with non-open source software Not Free:

MySQL: Classic $250, Pro $495 (for ~ 50 users) Compare to:

MSFT $3150 single proc for 50 users IBM $33000 single proc for 50 users Oracle $40000 single proc for 50 users

Aside: DB’s in different software stacks

Operating System

Middleware(DBMS)

Application

MS Windows or other OS

Oracleor MySQL, IBM, etc

SAPOr Oracle, Axtapa, etc.

ERPSoftware

Stack

Linux or other OS

MySQL or other DB

Apache Web Server

WebApplication

SoftwareStack

Proprietary Business Logic

BankingSoftware

Stack

IBM z/OSor other OS

GeneralSoftware

Stack

Oracleor other DB

ProprietaryBanking App.

Which companies are competitors? Which are complimentary to each other? Which are both?

mySQL

Which segments of market is mySQLstrong in? Large Companies or Small Companies? Web applications or Critical Enterprise data?

Why would a major enterprise want to payso much more for an Oracle or IBM DB?

My SQL: market

Small 20% Medium 30% Large 50%

Enterprise wide data 90%

Web Sites 10%

Microsoft Oracle IBM

My SQL

ReliabilityScalabilitySupportLongevity

Cost

Figure Adapted from “Teaching Note for MySQL Open Source Database,” 6/1/04, Stanford GSB.

How should mySQL grow in order to meet it’s stated goal of getting to $100 million In revenue?

My SQL: Growth Strategy

- Lack of Brand identity in this segement - MySQL lacks the organization to offer support - Large enterprises have high switching costs



Web Sites 10%


My SQL


Cost



- Not a big enough market to reach stated $100 million goal.



Web Sites 10%


My SQL


Cost


Stay Put?


+ Many of these customers already using MySQL with websites + Less emphasis on global organization + Leverage SAP alliance - Up against Microsoft.



Web Sites 10%


My SQL


Cost


Maybe?


+ builds on existing brand and strengths - Market not so big



Web Sites 10%


My SQL


Cost


Maybe?

Student Presentations

Audrey McCloskey FedEx

Mark Sheldon UPS

Databases & OLTP

Databases

Application I Application II

Aggregation: accessingmultiple databases

Sharing: two or moreapplications accessing thesame databases

Recall - Two capabilities

Travelocity.com CheapTickets.com

Example - Travel Agency

Hotels Cars Airtickets

What can go wrong?




A resource might be unavailable




Two applications might try to access & update the same resourceconcurrently




An application or a host might crash before the completion of thetransaction




A customer’s transaction should be completed in its entirety,or aborted

Transaction Processing

“The coordination of multiple resources and theshared access to common resources in a systematicand consistent way”

Examples? Financial applications (stock market, ATMs) Reservations (travel, theatre) Manufacturing (inventory, purchasing, billing) Etc…

Online Transaction Processing(OLTP) Transaction Processing for networked applications

4 Important Properties of transactions: ACID Atomicity Consistency Isolation Durability

The ACID properties

Atomicity All transaction components should either complete together

(commit) or abort E.g. All reservations (airline, hotel, car) should be grouped as a single

transaction that either commits, or aborts Consistency

A transaction must leave the system in a consistent state at the endof the transaction, or else abort

E.g. Either a consistent set of reservations has been made, or none Isolation

Concurrent transactions are allowed only if they don’t interfere witheach other

Two travel agents can concurrently access the same database if thereservations are for different dates/places

Durability A transaction leaves the resources in a permanent state after it

commits

Structure of a Transaction

DurableStarting State

Actions to beperformed

Durable,ConsistentEnd State

Successful completion

Abort

Rollback

OLTP

Simplifies application development

Ensables protection and integrity of mission-criticaldata in a transparent way for the end user for the application developer

Data warehouses &OLAP

“Data Mining: Concepts andTechniques”Jiawei Han©2006 Jiawei Han and Micheline Kamber,All rights reserved

Recall - Decision Support

Knowledge management systems: Turn dataand information into knowledge A Data warehouse stores organization’s historical

data

Provide functionalities for summarizing, aggregating,reporting on these data

Data mining is the process of discovering patternsin large amounts of data

What is a Data Warehouse?

A decision support database that is maintainedseparately from the organization’s operationaldatabase

Focusing on the modeling and analysis of data fordecision makers, not on daily operations or transactionprocessing

Slide adapted from slides for Data Mining: Concepts and TechniquesBy Jiawei Han and Micheline Kamber. Copyright 2006. See copyright notice

Data Warehouse—Nonvolatile

Operational update of data does not occur in thedata warehouse environment Does not require transaction processing, recovery, and

concurrency control mechanisms

Requires only two operations in data accessing:

initial loading of data and access of data


Data Warehouse—Time Variant

The time horizon for the data warehouse is significantlylonger than that of operational systems Operational database: current value data Data warehouse data: provide information from a historical

perspective (e.g., past 5-10 years)

Every key structure in the data warehouse Contains an element of time, explicitly or implicitly


Data Warehouse vs. OperationalDBMSOLTP (on-line transaction processing)

Major task of traditional relational DBMS Day-to-day operations: purchasing, inventory, banking, manufacturing,

payroll, registration, accounting, etc.OLAP (on-line analytical processing)

Major task of data warehouse system Data analysis and decision making

Distinct features (OLTP vs. OLAP): User and system orientation: customer vs. market Data contents: current, detailed vs. historical, consolidated Database design: Relational (ER) vs. Star View: current, local vs. evolutionary, integrated Access patterns: update vs. read-only but complex queries


OLAP Database Data modeling

“Star-schema” “Snowflake-schema”

OLAP Cube: Multidimensional Database

OLAP Database

OLAP Cube Enables analysis in various aggregation levels

OLTP vs. OLAP

OLTP OLAP

users clerk, IT professional knowledge worker

function day to day operations decision support

DB design application-oriented subject-oriented

data current, up-to-date

detailed, flat relational

isolated

historical,

summarized, multidimensional

integrated, consolidated

usage repetitive ad-hoc

access read/write

index/hash on prim. key

lots of scans

unit of work short, simple transaction complex query

# records accessed tens millions

#users thousands hundreds

DB size 100MB-GB 100GB-TB

metric transaction throughput query throughput, response


Why Separate Data Warehouse?High performance for both systems

DBMS— tuned for OLTP: access methods, indexing,concurrency control, recovery

Warehouse—tuned for OLAP: complex OLAP queries,multidimensional view, consolidation

Note: There are more and more systems which performOLAP analysis directly on relational databases


ISM 50 - Business Information Systems · PDF fileISM 50 - Business Information Systems Lecture 16 ... Jiawei Han ©2006 Jiawei Han and Micheline Kamber, ... (OLTP vs. OLAP):

Documents