Transcript Lecture

E-Commerce Technology
e-Commerce Technology Overview
Commerce (8000 B.C.)
BUYER
LOCATES
SELLER
SELECTION
OF GOODS
NEGOTIATION
SALE
PAYMENT
DELIVERY
INFORMATION
PHYSICAL
POST-SALE
ACTIVITY
Electronic Commerce (2002)
SOME TECHNOLOGIES USED:
SEARCH ENGINE
ON-LINE CATALOG
RECOMMENDER AGENT
SOME INFORMATION GATHERED:
BUYER
LOCATES
SELLER
SEARCH BEHAVIOR
BROWSING BEHAVIOR
CUSTOMER PREFERENCES
CONFIGURATOR
SHOPPING BOT
SELECTION
OF GOODS
EFFECTIVENESS OF PROMOTIONS
BARGAINING STRATEGIES
AGGREGATOR
AUTOMATED AGENTS
TRANSACTION PROCESSOR
NEGOTIATION
PRICE SENSITIVITIES
PERSONAL DATA
SALE
MARKET BASKET
DATA INTERCHANGE
CRYPTOGRAPHY
PAYMENT
E-PAYMENT SYSTEMS
TRACKING AGENT
CREDIT/PAYMENT INFORMATION
DELIVERY REQUIREMENTS
DELIVERY
ON-LINE PROBLEM REPORTS
ON-LINE HELP
INFORMATION
PHYSICAL
BROWSER SHARING
INTERNET TELEPHONY
POST-SALE
ACTIVITY
CUSTOMER SATISFACTION
FOLLOW-ON SALES OPPORTUNITIES
The Electronic Marketplace
BUYER
LOCATES
SELLER
SELECTION
AGENT
CREDIT FILE
SELECTION
OF GOODS
NEGOTIATION
ORDER
TRACKING
INSTALL
BID PREP
SALE
PAYMENT
DATA
ANALYSIS
DIRECT
SELL
SECURE
PAYMENT
DELIVERY
DELIVERY
POST-SALE
ACTIVITY
CRM
The eCommerce Process
• Buyers and sellers find each other
– Communication (via Networking, the Internet, and
Web-Based Information Architectures)
– Human-Computer Interaction, Multimedia
– Intermediaries
• Negotiation
The eCommerce Process
• Transaction
– Transaction processing, Databases
– Electronic Payment Systems,
– Computer Security,
– eCommerce Architecture
• Order fulfillment
– Manufacture (manufacturing systems)
– Delivery (tracking systems)
– Supply Chain Management
The eCommerce Process
• Post-sale events
– Customer Service and Help Facilities
– Reorder, restock
• Accounting
– Transaction processing
– Interoperability between online and legacy
systems
• Data analysis
– Data Mining
eCommerce Technology
•
•
•
•
•
Infrastructure
Wireless technologies
Search engines
Access security
Data interchange
• Cryptographic security
•
•
•
•
•
Electronic payments
Content delivery
Intelligent agents
Data mining
Mass personalization
E-Commerce Infrastructure
• What worldwide structure is required to support eCommerce?
• Network
• Machines
• Protocols
• Security
• Payment
How Does an Optical Fiber
Transmit Light?
•
•
•
Suppose you want to shine a flashlight beam down a long, straight hallway.
Just point the beam straight down the hallway -- light travels in straight lines, so it is no
problem.
What if the hallway has a bend in it? You could place a mirror at the bend to reflect the light
beam around the corner. What if the hallway was very winding with multiple bends? You
might line the walls with mirrors and angle the beam so that it bounces from side-to-side all
along the hallway. This is exactly what happens in an optical fiber.
Client/Server Architecture
•
•
•
•
Fundamental Internet structure
Client requests service; server provides it
Data exchanged only through real-time messages
Server may become a client to a different server
1
Server 2
responds
to client 1
Client 1 requests
service from server 2
The Internet
2
Client 2 requests
service from server 3
3
Server 3 responds
to client 2
Routers
NORTEL
3COM
CISCO
Router Tables
Internet Server
• The server is the heart of the technical
architecture, receiving requests from Internet
users, retrieving the information locally or
from networked devices and replying.
• Selection and sizing of this machine is critical
task, typically presenting a tradeoff between
performance and cost.
Web Server
Web server - A Web server is a piece of computer software
that can respond to a browser's request for a page, and deliver
the page to the Web browser through the Internet.
You can think of a Web server as an apartment complex, with
each apartment housing someone's Web page.
In order to store your page in the complex, you need to pay
rent on the space.
Pages that live in this complex can be displayed to and
viewed by anyone all over the world.
Your landlord is called your host, and your rent is usually
called your hosting charge.
Every day, there are millions of Web servers delivering pages
to the browsers of tens of millions of people through the
network we call the Internet.
UNIX v.s. NT
•
•
•
•
•
The two basic options are
– UNIX based platforms(IBM, Sun, HP)
– Microsoft NT based, Intel platforms
MS products generally cost less than UNIX platforms.
UNIX is a more mature OS than NT. As a result it delivers a better
performance for the same hardware configuration.
UNIX administration, requires more complex skills.
If you don’t have in-house UNIX expertise, investing in an UNIX based
server may require a large maintenance cost.
Server Workflow
Client / Server
In general, all of the machines on the Internet can be categorized
as two types:
• Server
• Clients
Those machines that provide services (like Web servers or FTP servers) to other machines
are servers.
And the machines that are used to connect to those services are clients.
When you connect to Yahoo at www.yahoo.com to read a page, Yahoo is providing a machine
(probably a cluster of very large machines), for use on the Internet, to service your request.
Yahoo is providing a server.
Your machine, on the other hand, is probably providing no services to anyone else on the Internet.
Therefore it is a user machine, also known as a client.
It is possible and common for a machine to be both a server and a client, but for our
purposes here you can think of most machines as one or the other
Client / Server
• A server machine may provide one or more services on the
Internet.
– For example, a server machine might have software running
on it that allows it to act as a Web server, an e-mail server
and an FTP server.
– Clients that come to a server machine do so with a specific
intent, so clients direct their requests to a specific software
server running on the overall server machine.
• For example, if you are running a Web browser on your
machine, it will most likely want to talk to the Web server
on the server machine.
• Your e-mail application will talk to the e-mail server, and
so on...
Understanding a simple Email
Server
•
•
The simplest possible e-mail server might look like this:
It would have a list of e-mail accounts,
–
–
•
It would have a text file for each account in the list.
–
•
•
•
•
•
•
with one account for each person who can receive e-mail on the server.
My account name bozdogan, John Smith's might be jsmith, and so on.
So the server would have a text file in its directory named bozdogan.TXT, another named
JSMITH.TXT, and so on.
When someone wants to send me a message, the person composes a text message
(“Barbaros, Can we have lunch Monday? John") in an e-mail client, and indicates that the
message should go to bozdogan.
When the person presses the Send button, the e-mail client would attach to the e-mail
server and pass to the server the name of the recipient (bozdogan), the name of the sender
(jsmith) and the body of the message.
The server would format those pieces of information and append them to the bottom of the
MBRAIN.TXT file. The entry in the file might look like this:
From: jsmith
To: mbrain Marshall,
Can we have lunch Monday? John
Email
•
There are several other pieces of information that the server might save into the file,
–
–
–
•
•
•
•
•
•
The time and date of receipt and a subject line, but overall you can see that this is an extremely
simple process!
As other people send mail to bozdogan, the server would simply append those messages to the
bottom of the file in the order that they arrive. The text file would accumulate a series of five or 10
messages, and eventually I would log in to read them.
When I want to look at my e-mail, my e-mail client would connect to the server machine. In the
simplest possible system it would:
Ask the server to send a copy of the bozdogan.TXT file.
Ask the server to erase and reset the bozdogan.TXT file.
Save the bozdogan.TXT file on my local machine.
Parse the file into the separate messages (using the word "From:" as the separator).
Show me all of the message headers in a list.
When I double-click on a message header, it would find that message in the text file and
show me its body.
–
You have to admit that this is a VERY simple system. Surprisingly, the real e-mail system that you use
every day is not much more complicated than this!
Understanding the Real Email
System
•
For the vast majority of people right now, the real e-mail system consists of two
different servers running on a server machine.
– One is called the SMTP Server, where SMTP stands for Simple Mail
Transfer Protocol. The SMTP server handles outgoing mail.
– The other is a POP3 Server, where POP stands for Post Office Protocol.
• The POP3 server handles incoming mail.
• The SMTP server listens on well-known port number 25, while POP3
listens on port 110
A typical Email Server looks like this:
Understanding SMTP
•
•
Whenever you send a piece of e-mail, your e-mail client interacts with the SMTP server to
do the sending.
The SMTP server on your host may have conversations with other SMTP servers to actually
deliver the e-mail.
Understanding SMTP
•
I sent an email using outlook express to Johnny
•
Outlook Express connects to the SMTP server at mail.mercynet.edu using port
25.
Outlook Express has a conversation with the SMTP server. Outlook express tells
the SMTP server the address of the sender and the address of the recipient, as
well as the body of the message.
The SMTP server takes the "TO" address (for example, [email protected])
and breaks it into two parts: 1) the recipient name (jsmith) and 2) the domain
name (mindspring.com). Since the recipient is at another domain, SMTP needs to
communicate with that domain.
The SMTP server has a conversation with a Domain Name Server and says,
"Can you give me the IP address of the SMTP server for mindspring.com?" The
DNS replies with the one or more IP addresses for the SMTP server(s) that
Mindspring operates.
The SMTP server at mercynet.edu connects with the SMTP server at Mindspring
using port 25. It has the same simple text conversation that my e-mail client had
with the SMTP server for Mercynet.edu, and gives the message to the Mindspring
server. The Mindspring server recognizes that the domain name for jsmith is at
Mindspring, so it hands the message to Mindspring's POP3 server, which puts
the message in jsmith's mailbox.
•
•
•
•
SMTP
•
The actual conversation that an e-mail client has with an SMTP server is incredibly simple and human
readable. It is specified in public documents called Requests For Comments (RFC) (see the links section)
and a typical conversation might look something like this:
•
•
•
•
•
•
•
•
•
•
•
•
•
helo test
250 mx1.mindspring.com Hello abc.sample.com
[220.57.69.37], pleased to meet you
mail from: [email protected]
250 2.1.0 [email protected]... Sender ok
rcpt to: [email protected]
250 2.1.5 jsmith... Recipient ok data
354 Enter mail, end with "." on a line by itself
from: [email protected] to:[email protected] subject: testing John, I am testing... .
250 2.0.0 e1NMajH24604 Message accepted for delivery
quit
221 2.0.0 mx1.mindspring.com closing connection
Connection closed by foreign host.
–
What the e-mail client says is in red, and what the SMTP server replies with is in green. The e-mail client introduces itself,
indicates the from and to addresses, delivers the body of the message and then quits.
Domain Name Servers
• If you spend any time on the Internet sending email
or browsing the web, then you use Domain Name
Servers without even realizing it.
• Domain Name Servers, or DNS, are an incredibly
important but completely hidden part of the Internet,
and they are fascinating!
• The DNS system forms one of the largest and most
active distributed databases on the planet, and
without DNS the Internet would shut down very
quickly.
DNS Resolution
Root server
Request for
uchicago.edu
DNS IP number
Internet
128.135.4.2
Request for
gsbkip IP number
Local DNS
server
Internet
GSB DNS server
128.135.4.2
128.135.130.201
Request for
gsbkip.uchicago.edu
IP number
128.135.130.201
File request
Internet
Desktop
File returned
Enterprise Web
Server
http://www.stamey.nu/DNS/DNSHowItWorks.asp
The Basic Idea
• For example, the machine that humans refer to as
www.mercynet.edu has an IP address of 216.27.61.137. Every
time you use a domain name, you use the Internet's domain
name servers (DNS) to translate the human-readable domain
name into the machine-readable IP address.
• During a day of browsing and emailing, you might access the
domain name servers hundreds of times!
Web Architecture
How are web sites constructed?
TIER 4
Database
TIER 3
Applications
TIER 2
Server
TIER 1
SOURCE: INTERSHOP
Firewall
• The firewall is typically a hardware/software
combination that controls the traffic between
your internal network and the public internet.
• Although a firewall can be directly
incorporated into an Internet server, it is most
commonly a specialized computer.
• The configuration is a challenging task and
should be performed by experts.
Firewall
As you can see all inbound and outbound Internet traffic must pass through the firewall
eCommerce Data Exchange Needs
RFQs
Ship Notices
Catalogs
Letters of Credit
Quotations
Purchase Orders
Electronic Payments
Bills of Lading
Invoices
Data Interchange
• How can sites exchange information without prior
agreement?
– What do the data fields mean? price, extended price, unit
price, prix, цена, τιμή, 값, X’AC12’
– XML: Extensible Markup Language
• How can the content be separated from form (visual
appearance)?
• How can data formats and structures be
communicated?
– What does the hex string “65436F6D6D65726365” mean?
– ASN.1, Basic Encoding Rules (BER)
Invoice Example
<UnitPrice>6.05</UnitPrice>
SOURCE: PROF. JEROME YEN
How to Make Data Portable
• Tell what the data means
• Tell how the data is structured
• Tell how it should look
SO COMPUTERS CAN
UNDERSTAND IT
• BUT DO THESE SEPARATELY. MIXING IS BAD
•
•
•
•
The meaning -- XML
The structure -- DTD (document type definition)
The formatting -- XSL (Extensible style sheet)
Example: XML catalog structure
XML at a glance
Well Formed Document:
<Book>
<Author>George Soros</Author>
<Title>The Crisis of Global Capitalism</Title>
<Year>1998</Year>
<Publ>Public Affairs</Publ>
<Price>26.00</Price>
<ISBN>1-891620-27-4</ISBN>
</Book>
DTD: Document Type Definition
<?xml version="1.0">
<!DOCTYPE Book [
<!ELEMENT Book
(Author, Title, Year, Publ, Price, ISBN)> ]>
SOURCE: PROF. JEROME YEN
XML Recipe Example
<?xml version="1.0"?>
<Recipe>
<Name>Apple Pie</Name>
<Ingredients>
<Ingredient>
<Qty unit=pint>1</Qty>
<Item>milk</Item>
</Ingredient>
<Ingredient>
<Qty unit=each>10</Qty>
<Item>apples</Item>
</Ingredient>
</Ingredients>
<Instructions>
<Step>Peel the apples</Step>
<Step>Pour the milk into a 10-inch saucepan</Step>
<!-- And so on... -->
</Instructions>
</Recipe>
Electronic Payment Systems
Electronic Payments
• Forms of money
– token (cash), notational (bank account), hybrid (check)
• Money does not move on the Internet
• Credit-card transactions
– Secure protocols: SSL, SET
•
•
•
•
•
•
Automated clearing and settlement systems
Smart cards
Electronic cash, digital wallets
Micropayments
Electronic delivery of goods
Electronic bill presentment and payment
– BlueGill
Intelligent Agents
•
•
•
•
•
Programs to perform tasks on your behalf
Metasearchers, shopping bots, news agents, stock
agents, auction bots, bank bots
How to make agents “intelligent”
– Rule-based systems
– Knowledge representation
Agents that learn
– Inductive inference
Negotiation agents
Avatars (characters in human form)
SYLVIE from VPERSON
Shopping Agents
Data Mining
• Extracting previously unknown relationships from
large datasets
• Discovery of patterns
• Predicting the future
– past behavior best predictor of future purchasing
• Market basket analysis
– diapers/beer
Data Mining Tools
• Visualization (“seeing” the data) Table Lens
• Predictive Modeling
• Database Segmentation
– Classify the users
• Link Analysis
– Associations discovery
• Neural networks
– Systems that learn from data
• Deviation Detection
– Are any of the data unusual? Fraud detection
Data Mining
• Extracting previously unknown relationships from
large datasets
– discover trends, relationships, dependencies
– make predictions
– target customers
• In eCommerce, data comes from
–
–
–
–
–
–
customers themselves
cookies
external databases
data matching
DoubleClick, etc.
Digital rights management tools (what we read and how
much)
– library records
Taxonomy of Data Mining Methods
Data Mining Methods
Predictive
Modeling
• Decision Trees
• Neural Networks
• Naive Bayesian
• Branching criteria
Database
Segmentation
Link
Analysis
Text
Mining
Deviation
Detection
Semantic Maps
• Clustering
• K-Means
Rule Associa tion
Visualization
SOURCE: WELGE & REINCKE, NCSA
Predictive Modeling
• Objective: use data about the past to predict future
behavior
• Sample problems:
– Will this (new) customer pay his bill on time? (classification)
– What will the Dow-Jones Industrial Average be on October
15? (prediction)
• Technique: supervised learning
– decision trees
– neural networks
– naive Bayesian
Mass Personalization
Mass Personalization
• Treating each user as an individual
– key is INFORMATION
• How to acquire and store information about
customers
–
–
–
–
Cookies
Question and response
Clickstream analysis
External databases.
• How to use information effectively and instantly
• Personalization technology
Outline
• What is personalization?
• Personalization is based on data
• How can data about people be acquired?
– From people themselves
– From their clickstream
– From outside data sources
• How can data be used
– To improve the customer’s experience?
– To help the company?
What is Personalization?
• Addressing customers by name and remembering their
preferences
• Showing customers specific content based on who they are and
their past behavior
• Empowering the customer. Examples: Land’s End, llbean
• Product tailoring. Example: dell.com
• Connecting to a human being when necessary.
CallMe
– Adeptra TeleBanner
• Allowing visitors to customize a site for their specific purposes
• Users are 20%-25% more likely to return to a site that they tailored
(Jupiter Communications, Inc.)
Need For Personalization
• In the real-world
– Customer relationship is mediated by people
– Personalization is critical: PEOPLE are PEOPLE
• On the Web
–
–
–
–
Too many customers; too few employees
Orders are entered by machine; follow-up is by machine
Customer relationship is mediated by machines
Personalization is critical
• Uniqueness (everyone is different)
• Efficiency (everyone has limited time)
Store Visitors in the Real World
• Casual store visitor:
– no intention of buying
• Prospecting store visitor:
DATA COLLECTED
ONLY IF VISITOR
BUYS SOMETHING
– wants to buy, maybe not here
• Add, marketing target:
– in store because of ad or promotion
• Customer:
–
–
–
–
buys something
pays cash
uses a credit card
uses a store charge card
IDENTITY UNKNOWN
PRODUCT/TIME KNOWN
IDENTITY KNOWN
IDENTITY, JOB, INCOME KNOWN
Store Visitors in Cyberspace
• Casual site visitor:
– no intention of buying
• Prospecting site visitor:
CAN EASILY DETECT
THE DIFFERENCE
– wants to buy, maybe not here
• Add, marketing target:
– in store because of ad or promotion
WE KNOW HOW HE
GOT HERE AND WHAT
HE WANTS TO BUY
• Customer:
–
–
–
–
buys something
pays cash
uses a credit card
uses a store charge card
WE HAVE HIS WHOLE FILE
WE KNOW WHAT OTHER PEOPLE
LIKE HIM ARE BUYING
Click Behavior
CASUAL VISITOR
STORE
HOME PAGE
OFFICE
PRODUCTS
HOUSEWARES
PRESENTATION
ITEMS
LASER
POINTERS
LASER
1
LASER
2
KITCHEN
TOASTERS
LASER
3
SPORTING
GOODS
HUNTING
RIFLES
GOLF
CLUBS
CALLAWAY
Click Behavior
PROSPECTING VISITOR
STORE
HOME PAGE
OFFICE
PRODUCTS
HOUSEWARES
PRESENTATION
ITEMS
LASER
POINTERS
LASER
1
LASER
2
KITCHEN
TOASTERS
LASER
3
SPORTING
GOODS
HUNTING
RIFLES
GOLF
CLUBS
CALLAWAY