Data Caching Strategies for Checking Integrity Constraints of Mobile Database Zarina Dzolkhifli,...

Preview:

Citation preview

Data Caching Strategies for Checking Integrity Constraints of

Mobile Database

Zarina Dzolkhifli, Hamidah Ibrahi, Lilly Suriani AffendeyPresented by: Phillip Cunningham

Title

Data Caching Strategies Store only relevant and necessary

information.

Checking Integrity Constraints How will we determine if the rules of our

database have not been violated by an update?

Title (cont.)

Mobile Database Any mobile application of a database which

stores the data locally. In particular for this article the concern is with

mobile databases without a permanent connection to a centralized database.

What is the point?(In English)

• To verify that our database will be happy with the changes made to it on a mobile device when it is not connected to a data network, and how we can do that most effectively.

• Put simply, we want to make sure that we won't break our database.

Is there anything that already does this?

Yes, to a degree. There is a system which verifies with a base

system over a network that the database is still okay.

The mobile host is required to connect to at least one 'node' of the distributed database system that is wired and kept up-to-date with all data.

Is there anything that already does this?

The goal of obtaining a mobile system that does not need to connect to a node to verify the consistency of the database after the update is unique.

Checking constraints efficiently has been done in centralized systems, but not for a mobile device that can do it on its own.

Method

1.) Given an update operation, U, determine which integrity tests might be violated

2.) Group the tests based on criteria such as:

Relation specified in tests

Grouped by type of test (complete/sufficient)

Where they must be evaluated (local or not)

Each group has at least one test from each integrity constraint.

Method (cont.)

3.) Derive internal representation of integrity tests.

Method (cont.)

4.) Analyze relationships between the tests.

Redundant Subsumptive

Rule 1: If two tests require the same set of data they are redundant

Rule 2: If one test requires a subset of the data of another test it is subsumptive.

Experiments

This compares different caching methods for a specific database and two transactions.

Strict Strategy and Relaxed Strategy are those put forth by the authors.

More Experiments

When adding more tuples to the tables involved RS and SS still cached the same amount of tuples.

Considerations

These are good strategies to improve speed and keep the mobile device from needing to download the entire database.

This won't actually allow for verifying transactions if you don't have the right data already on your phone and are disconnected.

Questions

?

Recommended