"Where there's a will;
I will restrict my self from migration issues. The mobile clients will (must)
contact the server, not the other way around, i.e., I will ignore the fact that
mobile computers may wish to migrate between different networks. Readers
interested in this are referred to . I will not take any
considerations with regards to security (e.g., authentication and data
there's a won't..."
- D:A:D (Unowned)
This thesis is a continuation of the work presented in . To
avoid unnecessary repetition of some of the material, I will skip in-depth
treatment of the following issues:
- Serializability-based synchronization (e.g., Two-phase locking,
Optimistic concurrency control, and Timestamps).
- Atomicity and recoverability (e.g., Stable storage, Logging, Shadowing,
and Two-phase commit).
- General description of distributed file systems (e.g., Network File
System, Andrew File System, Little Work, and Coda).
- The role of a mobile host [in transaction processing] (e.g., Remote
terminal, Full-fledged server, and Preprocessing client).
For my own sake, and because my views differ slightly from those given
in , I will be repeating the parts about mobile-computing in
general and basic properties of transactions. Furthermore, I find it natural to
include a full description of the implemented model for consistency.