EJBs%20 XML%20 %20Integrity%20Constraints

1
EJBs XML Integrity Constraints

• Data-Object Modeling and Optimization (DOMO)
• June 2003
• Rajesh Bordawekar, Michael Burke, Mukund
  Raghavachari, Vivek Sarkar, Oded Shmueli

2
A Changing Application Environment

• Frameworks such as Container-Managed Persistence
  simplify integration of application logic with
  databases.
• Programmer specifies mapping between Java classes
  and schema declaratively.
• Relational databases have integrity constraints.
• XML schemas have integrity constraints.
• Application logic may have its own integrity
  constraints.
• There is no mechanism for reasoning about
  constraints on data given a mapping between
  language types and relational schema.
• Applications may fail when seemingly correct
  updates are rejected by the database.
• Domain constraints and referential integrity
  constraints.
• Need to reason about such constraints.

3
Setting
Mapping associates db.X. t and ejb.X.t
Mapping
db.X. t
ejb.X.t
XML data
T
? What can be deduced about db originated values
based on the schema?
Relational data
4
EJBs XML

• Current J2EE implementations provide little
  support for XML data sources.
• Goals
• Understand integration of XML and EJBs.
• Efficient algorithms for instantiating EJBs from
  XML data sources.
• E.g., Streaming algorithms.
• Focus Mechanisms for ensuring that XML (and
  relational) integrity constraints are maintained
  in application logic.
• Past Work
• Streaming algorithms for XPath (ICDE 2003).

5
XML Data Bindings

• XML data is imported into the application
  environment.
• Often, the data conforms to an XML schema or a
  DTD.
• XML schemas support constraints domain,
  uniqueness, key.
• These should be related to relational database
  and application constraints.
• Data travels in both directions (export).

6
Plans

• Examine data binding formalisms.
• Examine XML schema constraints.
• Analyze their interplay at the application level
• Support static analysis of programs to determine
  constraint violations.

7
Example Updating a Record"

• EJB Field SAL is updated in an employee record.
• SAL is associated with min10K, max200K.
• Two salary ranges 10,45, 65,200.
• May check application code to determine possible
  violations.

8
Record XML Schema

• Complex type
• ltcomplexType nameemployee/gt
• ltattribute nameempName typeeid/gt
• ltattribute namedeptNum typednum/gt
• ltattribute namesalary typeempSal/gt
• lt/ complexType gt
• Union type
• ltsimpleType nameempSalgt
• ltunion memberTypelow high/gt
• lt\simpleTypegt
• Restriction
• ltsimpleType namelowgt
• ltrestriction baseintegergt
• ltminInclusive value10/gt
• ltmaxInclusive value45/gt
• lt/restrictiongt
• lt/simpleTypegt

9
Use Cases

• Reading a database record
• Writing/Updating a database record
• Reading an XML element/attribute
• Writing/Updating an XML element/attribute
• Assisting CMP/BMP
• Assisting CMR

10
Using Constraint/Type Information

• Constraints for the same application may
  originate in a relational DB, an XML schema, a
  DTD or other sources.
• We therefore need to reason about and manipulate
  constraints originating in various formalisms.
• Analyze program statically and dynamically to
  verify that constraints are met.
• Failure of a check implies update will fail in
  DB. Success usually provides no definite
  information.
• Much related Dist. DB literature on the subject.

11
XAOS XPath Analysis and Optimization for
Streaming

• Motivation
• XPath lookup is next major bottleneck in XML
  processing (after parsing)
• Streaming access is most efficient mode for
  memory hierarchies and large documents

XML Document
XML Parser
XPath Expression
Parsing Events
Specialized XPath Processor
Xpath-specific Automaton
XML Fragments

• Xpath-specific Automaton supports streaming
  XPath processing for forward and backward axes

12
XAOS Example
XPATH expression //yu//wancestorz/v
XDAG
Root
Root
Root(0)
Root(0)
descendant
descendant
Y
Z
Y
Z
X (1)
child
descendant
descendant
child
Y(10)
Y(2)
Y(2)
U
W
V
U
V
W
Z(4)
U(9)
Z(11)
Z(4)
U(9)
W(3)
U(13)
XML Document
V(5)
V(5)
V(6)
V(6)
W(7)
W(7)
W(7)
W(7)
W(12)
V(5)
V(6)
W(7)
Parser Specialized XPath Processor
Xpath-specific Automaton
W(8)
W(8)
W(8)
W(8)
XPath Matchings in XML Document
13
XAOS Scalability Parsing XPath Execution times

• Experiments run on
• 550 MHz Pentium III,
• w/ 256MB RAM Linux 2.2