The Bubba prototype was implemented using a 40 node FLEX/32 multiprocessor with 40 disk drives.
Although this is a shared-memory multiprocessor, Bubba was designed as a non-shared memory
 system and the shared-memory is only used for message passing. Nodes are divided into three
groups: Interface Processors for communicating with external host processors and coordinating
query execution, Intelligent Repositories for data storage and query execution, and Checkpoint/Logging
Repositories. While Bubba also uses declustering as a storage mechanism (both range and hash
declustering mechanisms are provided) and dataflow processing mechanisms, Bubba is unique in a
number of ways. First, Bubba uses FAD rather than SQL as its interface language. FAD is an extended
relational persistent programming language. FAD provides support for complex objects via a variety of
type constructors including shared sub-objects, set-oriented data manipulation primitives, as well as
more traditional language constructs. The FAD compiler is responsible for detecting operations that can
be executed in parallel according to how the data objects being accessed are declustered. Program
execution is performed using a dataflow execution paradigm. The task of compiling and parallelizing a
FAD program is significantly more difficult than parallelizing a relational query. The other radical
departure taken by Bubba is its use of a "single-level store" mechanism in which the persistent
database at each node is mapped into the virtual memory address space of each process executing at
the node instead of the traditional approach of files and pages. This mechanism, similar to IBM's
AS400 mapping of SQL databases into virtual memory, HP's mapping of the Image Database into
the operating system virtual address space, and Mach's mapped file mechanism, greatly
simplified the implementation of the upper levels of the Bubba software.