As
Organizations grow the level of heterogeneity
across systems increases. This is true whether
you grow internally or grow through acquisition.
The ability to share or unify information decreases.
The effort required to make decisions from comprehensive
information skyrockets. Organizations settle for
doing what they can. Over the years a number of
companies have developed ETL (Extract, Transfer,
Load) tools in an attempt to resolve some of these
"Islands of Information" issues. Through
the acquisition and deployment of these tools
Organizations have discovered that there are a
large family of issues arising from the heterogeneous
set of systems, databases, data models, requirements,
and collection facilities. This family includes
:
- The
same data types on different platforms are often
very different - often in several ways,
- Different
data types do not always have uniform mappings,
- Related
attributes in different systems may often vary
in their actual semantics,
- Attributes
of identical intended semantics on different
systems often vary in their actual semantics,
- Data
quality management varies across systems (in
focus, definition, mechanisms, and results),
- Unifying
Data Quality metrics and remedial measures has
a family tree of its own,
- Moving
very large volumes of data results in very large
issues of time, processing capacity, resource
allocation, parallelism, fault tolerance (check
pointing), and scalability,
- What
ETL operations are performed on what platforms
is more important than it should be,
- These
ETL operations require more time, human resources,
expertise, and maintenance than they should,
- Changes
on any one system are frequent and disruptive,
- Isn't
all of the intelligence embedded in this work/system
organizational meta-data?
- This
'derived' meta-data should be brought to bear
on the problem,
- All
of this is more than humans should be expected
to address effectively.
We
take a "collective problem solving approach"
to data management. We achieve this through Ab
Initio's (ab initio, Latin "from the
beginning") collection of software products.
The Ab Initio Co-Operating system provides a layer
over the operating system (MVS, OS390, UNIX, Win2000/NT)
that renders all platforms homogeneous from a
runtime perspective. The Ab Initio Graphical Development
Environment provides the high level graphical
interface for designing/defining the entire dataflow
(Source(s), Operation(s), Target(s), and their
connections). The GDE enables data engineers to
focus on the conceptual model. The underlying
software addresses the low level mechanics. Physical
data objects serving as Sources and Targets can
be virtually anything from flat files, to ISAM
files, Databases (virtually any relational, network,
hierarchical), SAS data object, Object store,
Excel files, ..., almost anything. The Operations
are supported by a very large and comprehensive
set of 'out-of-the-box' operations. Any operation
one wishes to build can be built and used as an
Operation. It does not matter what language it
is used. A co-operating system wrapper can be
applied to the operation and it can run "anywhere".
Since the GDE is "data aware" (for all
Sources and Targets), any change to any attribute
in any table of any Source or Target is automatically
registered by the GDE in its internal data management
log and the appropriate runtime component(s) can
be generated. The GDE is also heavily parameter
driven. These parameters provide all of the configuration
information.
The
runtime products of the GDE can be run on any
platform running the Co-Operating system. They
can be instructed to run in parallel, on specific
CPU, or specific assignment criteria (load balancing,
round-robin, ...). The executable dataflow objects
are fully scalable. Data movement, transformations,
defect pooling, defect reporting, accepted data
capture, are supported through the Operations.
The
system wide meta-data is automatically generated
and maintained in a current state through access
to a common meta-data repository that is populated
and used by many of the Ab Initio software components.
The interface provides an efficient method for
change impact analysis, both a priori and
a postiori.
We
can resolve the "Islands of Information"
for virtually any organization.
|
|
Capabilities
Summary
Our
Process - A Summary
Our
People - the Team
Whitepapers
Partners
|