Next: CPISync
Up: thesis_2
Previous: Putting it together
  Contents
Personal digital assistants or
PDAs are periodically linked up to computers in order to
synchronize data being held on the PDA and the data held by
complementary applications on the computer. A secondary benefit of
periodic synchronization is that it serves as a backup mechanism.
Typical application data that needs to be synchronized includes
address books, memo pads, to-do lists, email clients, schedule
planners and off-line internet and web information (like avantgo
[27]). Information in these applications is modified on
the PDA and/or the PC independently. The more often
synchronization is done, more current is the data being held on
the PDA and the PC. A natural result of frequent synchronization
is that the amount of `new' information between two successive
synchronizations is very small compared to the total amount of
information in the databases of the PDA and the PC.
In this chapter we first introduce CPISync (characteristic
polynomial interpolation based synchronization) in
Section 3.1 that is based on a set reconciliation
algorithm first proposed in [14]. We then
describe two variants of the original algorithm, namely
Deterministic CPISync and Probabilistic CPISync in
Sections 3.2 and 3.3. The algorithms'
theory as well as implementation and results are described in
these sections.
The CPISync algorithm is suitable for application in PDA
synchronization because it makes the communication and time
complexity of synchronization a function of the differences
between two databases being compared and not the size of the
databases themselves. In this chapter we first discuss the CPISync
algorithm and then go on to discuss its application in the PC-PDA
synchronization system. Towards the end of the chapter we discuss
the our results of the CPISync PC-PDA synchronization as compared
to Slowsync, discussed in Section 2.2.1.
Subsections
Next: CPISync
Up: thesis_2
Previous: Putting it together
  Contents
Sachin Kumar Agarwal
2002-07-12