Orie 480 Study Sheet Midterm

ORIE 480 Study Sheet: MIDTERM James Yu Jane Thipphavong

Lecture 1: Database Systems – Intro − − − − − −

−

−

−

−

Data: Raw facts that are described, observed, or measured Information: Data that has been organized or prepared Knowledge: Data/Information/Rules that are used for actual decision making Database Management System (DBMS): A collection of programs that define/manipulate/maintain databases Database: A collection of interrelated data File Systems are unattractive o Data-program dependence: must write custom programs to query o Redundancy o Inconsistency o Lack of security o Have to write your own o If a query has not been foreseen, an expert programmer is needed Different types of DBMS o Hierarchical o Network o Relational o Object-oriented Database Tables Restrictions o Columns contain same type of data o One value per cell o One (or more) column(s) contains a unique value for each row  Primary Key o Rows have the same size (# of columns)  Order is unimportant, and are dynamic Transactions o Logical unit of access to a DBMS o Concurrency and Recovery from system crashes o Locks, Write-Ahead Logs and Checkpoints 2 Components of RDBMS o Data Definition Language (DDL)  Allows you to define metadata o Data Manipulation Language (DML)  Allows you to insert/edit/delete/ask questions about data  QBE, SQL

Lecture 2: Designing Databases Entity-Relationship Diagrams − − − −

Entities: things about which one wants to store info Attributes: the info one wants to store about the entity Relationships: Associations between two or more entities Cardinality: 1:1, 1:N, N:M

−

Participation o Total Participation: every student must have an advisor o Partial Participation: a professor may not be teaching any classes o Circle: Denotes partial participation for the opposing box (think race relations and throwing the balls) o Referential Integrity:A database entry should refer to something that exists

− Weak Entity: Part of its key is in another entity − Primary Key: An attribute that uniquely determines a record − Foreign key: An attribute in one entity which is the primary key of another entity to which −

the table is related Mapping ERDs to Tables o 1:1  Put the primary key of 1 entity to the table of the other 1 entity o 1:N  Put the primary key of 1 entity to the table of the N entity o N:M  Create new table consisting of primary keys of both tables

Lecture 3: WWW −

− −

Client Server Model o Client opens a connection, makes a request o Server returns file o Connection closed WWW runs on HyperText Transfer Protocol o GET requests an item o HEAD requests information about an item (then check if stored in cache) Many IT systems and definitely the WWW are too large for a person to use effectively, so develop searching tools

Lecture 4: Query by Example and SQL −

Done by Examples

Lecture 5: Internet Markup Languages and Regular Expressions −

− −

HTML o Very simple o Fail-soft principle: If your tags are wrong, something still appears o Apply “formatting tags” to basic elements o Meta tags provide information o Structural versus formatting tags SGML o Standard generalized markup language o What HTML is based on, pre-defined tags XML o eXtensible Markup Language for describing information o user-defined tags o extensibility o precise structure o valid Document Type Definition (DTD)  DTD is a “grammar” for the XML document

o o o

eXtensible Stylesheet Language (XSL) XML is becoming the standard front end of databases on the Web DBMS are incorporating XML output capabilities

Lecture 6 and 7: Networks

o Circuit Switched: end systems are connected to senders by an end-to-end reserved circuit.

o Datagram Networks: Every packet has full addressing information, and can be routed independently of others.

o Layers of Networks: Host -> TCP or UDP -> IP -> Subnetwork -> Physical Media o

o o

o

o o

o

o

o

An identifier for a computer or device on a TCP/IP network. Networks using the TCP/IP protocol route messages based on the IP address of the destination. The format of an IP address is a 32-bit numeric address written as four numbers separated by periods. Each number can be zero to 255. For example, 1.160.10.240 could be an IP address. UDP: User Datagram Protocol, o Demuxing layer o In order to know which process to send the packet to (email, ftp, ,telnet, web, etc) TCP (Transmission Control Protocol) o Packet Sequencing o Flow control Domain Name System (DNS): distributed Internet directory service. DNS is used mostly to translate between domain names and IP addresses, and to control Internet email delivery. Most Internet services rely on DNS to work, and if DNS fails, web sites cannot be located and email delivery stalls. Email: Client sends to Sendmail server, which sends to Mail Transfer Agent (MTA), which queries the DNS and sends to the destination MTA, which sends to the destination Getmail server, and finally to the receiving client. Firewalls o Kind of router o Prevents unwanted packets from entering a site o Hackers could still hack into the mail server and gain access to site o DMZ (Demilitarized Zones) evolved let mail server reside outside firewall to prevent hackers from gaining access to site through mail server (and other such servers) IP Address Distribution Problem o Improve the efficiency of address assignment o Design a new IP with more address space o Dynamic IP address assignment o Network Address Translation; Gives hosts non-global addresses and translates them to global addresses on the fly Virtual Private Network (VPN) o The concept of including a node in a public network (internet) into a site’s private network o Using tunneling and encryption for security Dial-up Roaming o ISPs having roaming agreements with other ISPs via a common mediator o Allows you to access your ISP in many different cities around the country

Lecture 8: Dynamic and Active Web Technologies o Static: a file resides on the web server o

Pros: cheap and fast

Cons: inflexible, revisions are hard, may need many documents compared to dynamic Dynamic: document is created upon request (by server) o Pros: flexible; current o Cons: uses server resources, slower, costly to create and debug; can’t display changing information; security issues for the server o State Information  Information that program saves between invocations Active: Server returns a program which runs in user’s browser o Pros: flexible o Cons: Costly and difficult to create and debug; security risk for the client o Java active documents are called “applets” o Javascript: browsers read and interpret javascript embedded in html o

o

o

Lecture 9: Introduction to Decision Analysis − Goal: Make decisions −

o Well grounded probability Probability o P(A|B) = P(AB) / P(B)

Lecture 10: Decision Analysis Biases −

−

−

Human Limitations on Information Processing o Short-term memory o Unreliable recall of information o Slow numerical calculations o Inaccuracy from Biases Two Models of Decision Making o Fully Rational Approach − Requires complete, perfect, and instantaneous information − Requires solution to hard computational problems − Rationality = Expected utility maximization o Satisficing (bounded-rational) approach − Use of limited amount of information, heuristics Heuristics o Availability Heuristics

− What is easily recalled must be most likely −

o

o

Biases • Ease of recall • Retrievability Representativeness Heuristics − Beliefs about what are representative examples of particular groups of events − Biases • Insensitivity to base rates • Misconceptions of chance Anchoring and adjustment heuristics − Starting somewhere and adjusting − Biases • Over reliance on initial guess and overconfidence

−

Decision Analysis o Advantages − Models rich problem structure − Well grounded in theory − Formal methods to compute optimal solutions − Avoids human biases − Sensitivity analysis o Disadvantages − Abstract model: may be to simple − Uncertainty can still exist − Model can include biases − Size and complexity of model can become horrendous

Lecture 11: Decision Trees − − − − −

Maximax: Choose decision that yields maximum profit Maximin: Choose decision that would make you lose the least amount of money Regret Matrix: Difference in the amount of profit if you had chosen Least Regret Decision: Choose decision that generates least amount of regret in any case Decision Tree o Solving a Decision Tree  Compute the posterior conditional probabilities • Ie. P(good market | forecast good)  Compute the values at the outcome nodes  Work back from outcome nodes towards the root node  Read off the optimal decision at teach decision node o Advantages  Structures a decision process  Evaluates all possible outcomes  Decision making process is succinct  Focuses decision on financial figures, probability, and assumptions  Sensitivity analysis

Lecture 12: Information Systems in Enterprises −

Information Architecture Structure o Hierarchy of Levels  Strategic  Management  Knowledge  Operational o Departments  Sales and Marketing  Finance  Accounting  Manufacturing  Human Resources o IT Infrastructure  Hardware  Software

o

−

OLTP o o

 Data and Storage  Networks 6 Major types of Information Systems  Executive Support Systems (ESS) (Strategic Level)  Management Information Systems (Management Level)  Decision Support Systems (Management Level)  Knowledge Work Systems (Knowledge Level)  Work Stations (Knowledge Level)  Transaction Processing Systems (Operational Level) A type of computer processing in which the computer responds immediately to user requests. Each request is considered to be a transaction. Automatic teller machines for banks are an example of transaction processing. ACID test

    −

−

− −

−

−

Atomicity (all or nothing) Consistency (The database is transformed from one valid state to another valid state.) Isolation (The results of a transaction are invisible to other transactions until the transaction is complete) Durability (Once committed (completed), the results of a transaction are permanent and survive future system and media failures)

Batch o The opposite of transaction processing is batch processing, in which a batch of requests is stored and then executed all at one time. Transaction processing requires interaction with a user, whereas batch processing can take place without a user being present. MIS issues o Multiple entries of same data o Multiple versions of the same element o Conflicting data definitions o Nonstandard systems interfaces o Data glut o Inflexibility o Report writing backlog o Data ownership and hoarding o Lack of timeliness Business Processes o Before: Stovepipe Systems o After: Enterprise System On Line Analytical Processing o a key reporting and analysis tool for operations and strategic planning o Query-driven, usually over multidimensional DB o most OLTP systems unsuitable for OLAP: OLTP will slow down OLAP Data Warehouses o Source of data for OLAP  Includes analytical tools o Key enabler is Metadata  Identifies the technical aspects of the data  Identifies the business meaning of the data  Provides utilization tools Data Mining

−

o Information as the answer to unasked questions o Known as data discovery o Employs scalable math algorithms Infrastructure Design Issues o Distributing functionality across system components  What should be centralized? Distributed?  Basis for decisions?  Capacity planning  Total Cost of Ownership  Asset management

Lecture 13: Understanding the Business Value of Systems o Challenge: Determining intangible benefits of a system (i.e. improved resource control, o

o

o o

more timely information, increased organizational learning, enhanced employee goodwill, satisfaction, client satisfaction, corporate image) Capital Budgeting: The process of analyzing and selecting various alternatives o Payback Method: Time required to pay back the initial investment  No. years to pay back = Original Investment / Annual Net Cash In  Advantage: Simple, easy to communicate  Disadvantage: How big is the project, what happens after? o Accounting Rate of Return on Investment (ROI)  Approximation of the accounting income  Total benefits - -Total costs – Depreciation / Useful life = Net Benefit  Net Benefit / Initial Investment = ROI  Advantage: Simple  Disadvantage: Arbitrary judgments, Bad for long term projects o Cost Benefit Ratio  Total Benefits / Total Costs = Cost Benefit Ration  Advantage: Simple  Disadvantage: No information about timing. o Net Present Value: Time value of money  Future Value / (1 + Interest Rate) ^ N = Present Value o Profitability Index: Compare the profitability of different alternatives  Present value of cash in / Initial Investment = Profitability Index  Advantage: Unlike NPV, we have an idea of how profitable a project is  Disadvantage: complex o Internal Rate of Return (IRR): The rate of return of the investment  The interest rate that sets NPV = 0  Advantage: Time value of money, investment point of view  Disadvantage: Complex, No info about the magnitude of the return Real Options o Hold an option on a future IT investment o Resove uncertainty over time o Start wit a small pilot rather than the big risky investment Scoring Models o Assign project scores based on different objectives. Portfolio Analysis o Often takes long term “strategic” view of IT o Project portfolio should be balanced o Considers multiple projects

Lecture 14: MIS Design and Implementation -

-

-

-

-

-

-

-

-

Key Components of Systems Analysis o Decisions o Transactions and Processing o Information and its flow o Individuals and functions involved o Communications and coordination Systems life cycle o Inception o Systems analysis o Feasibility Study o Design o Specifications o Building the system o Testing o Training o Conversion & Installation o Maintenance & Enhancements Critical Success Factors o A small number of objectives on which IS must focus o Attention on how things should be Data Collection for Analysis and Design o Observation o Interview o Questionnaires Design Methodologies o Structured (e.g. Data flow diagrams) o Objected Oriented Model (UML: universal modeling language) Testing Conversion o Parallel Strategy: run both old and new system o Direct cutover: kill old system and use new one o Pilot Study: implement new system in one part of company o Phased Approach: intro system in stages instead of all at once Training and Documentation User-oriented Design o Use of prototypes o Prototypes are quickly made and have flexible design Outsourcing o Advantages  May pay only what you use  Access to top talent  Frees up company resources o Disadvantages  Underestimate cost of finding and evaluating vendors  Monitoring  Loss of control Causes of successes and failures o Funding o Insufficient user involvement o User-Designer communications gap o Insufficient management support

o

o o

Level of complexity  Project size  Project structure  Experience with technology Politics Non-systematic approach to changes