Data Management Issues in Electronic Commerce

1
Data Management Issues in Electronic Commerce

• M. Tamer Özsu
• Department of Computing Science
• University of Alberta

2
Disclaimer

• If you are looking for answers to all your
  questions
• Get a cup of coffee and relax.
• I dont have many answers.

3
What are the issues?

• Catalog development
• Intelligent querying
• Workflow management
• EC transaction atomicity
• Management of distribution and heterogeneity

4
Electronic Commerce

• Electronic commerce, in its most general
  definition, refers to selling and buying using
  electronic means.
• Has roots in Electronic Data Interchange (EDI)
• 1960s ? not new
• What is new?
• Internet

5
Forms of Electronic Commerce

• Business-to-business
• EDI-based
• Procedures well-defined
• Parties know and (generally) trust each other
• Low transaction rate high value
• Business-to-consumer
• Internet-based
• One-of-a-kind procedures
• Almost no trust
• High transaction rate lower value

6
What is common?

• Someone has goods/services to sell
• Someone is looking for something to buy
• There has to be a mechanism for the buyer to find
  who sells what he/she wants
• There has to be a mechanism to engage in a
  transaction
• There has to be a way of exchanging money and
  goods/services

7
Why Databases?

• Many repositories
• Catalogs
• Working repositories
• Collaborative work
• Sharing of repositories

8
Smart and Virtual Catalogs

• Smart Catalog
• Searchable, annotated combinations of
  machine-readable (i.e., minimally processable)
  and machine-sensible (i.e., actually understood
  by computer) product data
• Virtual Catalog
• Systems that dynamically retrieve information
  from multiple smart catalogs and present these
  product data in a unified manner with their own
  look and feel, not that of the smart catalog
• A. Keller, Smart Catalogs and Virtual Catalogs,
  In Readings in Electronic Commerce, R. Kakakota
  and A.B. Whinston (eds.), 1997, pp. 259-274.

9
Issues and Challenges

• Fully multimedia catalogs
• Existing catalogs are mostly text-based virtual
  catalogs with embedded images
• Add more support for modeling images for
  similarity searches
• Add audio and video to provide product
  demonstrations, voice-overs and even commercials
• Structured document databases (SGML,XML) image
  databases continuous media systems

10
Issues and Challenges

• More open virtual catalog environment
• Single-vendor or proprietary virtual catalogs
• open to more flexible integration of various
  vendor catalogs
• Distribution
• move to the Internet environment ? scalability
• provide search capabilities
• Browsing-based access limited search

11
Handling Document Types (DTDs)
lt!ELEMENT ..gt ..... lt!ATTLIST...gt
DTD file
DTDs
DTD Manager
Query Interface
DTD Parser
Type System
Type Generator
SGML Documents
Users
Multimedia DBMS
12
Automatic Document Insertion
Authoring Tools
SGML Document Instance
DTD
DTDs
lt!ELEMENT ..gt ..... lt!ATTLIST...gt
SGML Parser
Query Interface
Type System
C Types
Parse Tree
SGML Documents
Instance Generator
Users
C Objects
Multimedia DBMS
13
EC Transactions

• Electronic commerce consists of processes which
  can be modeled as workflows.
• Workflow A collection of tasks organized to
  accomplish some business process.
• D. Georgakopoulos

14
Workflow Types

• Human-oriented workflows
• Involve humans in performing the tasks.
• System support for collaboration and
  coordination but no system-wide consistency
  definition
• System-oriented workflows
• Computation-intensive specialized tasks that
  can be executed by a computer
• System support for concurrency control and
  recovery, automatic task execution, notification,
  etc.
• Transactional workflows
• In between the previous two may involve humans,
  require access to heterogeneous, autonomous
  and/or distributed systems, ACID properties

15
Workflow Example

• T1 Search vendor catalog T4 Send shipping
  instructions
• T2 Place the order T5 Pay invoice
• T3 Receive proforma invoice T6 Receive goods

16
ACID Properties

• ATOMICITY
• all or nothing
• CONSISTENCY
• no violation of integrity constraints
• ISOLATION
• concurrent changes invisible ? serializable
• DURABILITY
• committed updates persist

17
EC Transaction Atomicity

• EC transaction has a slightly different meaning
• refers to the actual transaction of exchanging
  goods for money
• Atomicity of this type of transaction is important

18
Atomicity Types

• Money atomicity
• transfer of funds from one party to another
  without the possibility of the creation or
  destruction of money
• Cash transaction
• Goods atomicity
• Money atomicity exact transfer of goods for
  money
• COD transactions
• Certified delivery
• Goods atomicity both the merchant and the
  customer can prove exactly what goods were
  delivered

19
Atomicity Protocols
P
P
P
P
C
C
C
P
P
P
P
ready?
yes/no
commit/abort?
commited/aborted
Phase 1
Phase 2
20
NetBill
NetBill Server
(2) PO invoice
(3) Approval
Customer
Merchant
(1) Payment order
(4) Delivery of goods
21
Intelligent Querying

• Passive queries
• Find me all Montblanc fountain pens whose price
  is lt 500.
• Active queries
• Whenever a Montblanc fountain pen of 1920s
  vintage becomes available for under 1000, let me
  know.
• Agent technology

22
Distributed System - Users Vision
Distributed Catalog
23
Reality
Catalog Software
Catalog Software
Communication Subsystem (Internet)
Catalog Software
User Query
User Application
User Query
Catalog Software
Catalog Software
User Application
24
Issues

• Distribution
• Heterogeneity
• Autonomy
• Openness to Internet

25
If you have ideas please send me an
email ozsu_at_cs.ualberta.ca and watch out for
the ACM SIGMOD Panel on this issue