Introduction to Data Management
Managing Data on the Tandem Systems
15873 Tandem Computers Incorporated 2-7
Location. In conjunction with the EXPAND software, TMF enforces consistency
across all nodes of a database spread over a network. TMF guarantees this
consistency no matter where a transaction originates or how it is geographically
distributed. As an example, a transaction might generate updates to files stored at
three different nodes. If the updating required at one node cannot be completed
for any reason, the database potentially becomes inconsistent because only two of
the three nodes reflect the transaction. TMF, however, removes (backs out) every
part of the transaction at nodes that started updating activity before the transaction
failed. This action restores the database to its original consistent state.
Application
Expandability to Match
Your Business Growth
Just as your business grows in size, your application software must be readily
expandable to accommodate new kinds of tasks. Several Tandem system features
combine to offer this kind of expandability. On the Tandem system, an application
can grow to include large databases accessed by thousands of users.
Local and Locally
Distributed Databases
A local database is composed of a set of files located on a single disk volume. A
locally distributed database, however, consists of files that reside on more than one
disk volume, either in a single Tandem system or in multiple systems in a local
network.
Because you can add new programs to your system without recoding any of the
current software, you can easily expand a local database or extend such a database
into a locally distributed one. You code the program that performs the function you
want to add. Then you implement that program under the GUARDIAN 90 operating
system as a new process that communicates with existing processes. Because all
processes run under control of the GUARDIAN 90 operating system, you never have
to convert them to run under another operating system, no matter how large your
application grows.
As Figure 2-3 shows, this same capacity for expansion makes it convenient to add
entire applications that manage completely different tasks and consist of many new
requester and server processes.