 |
 |
5 Utilizations for 1 Development
|
|
The TierBroker message is very easy to understand - with one interface you get a complete solution for the five most important data integration requirements.
- Enterprise Application Integration (EAI)
Connections to MQ Series, TIBCO, JMS and other message oriented middleware.
- High Speed Batch Processing
The ability to process flat-file formats that are character and position delimited, with mixed record structures at speeds of +10,000 records per second.
- Application Connectors
Connecting real-time information and transactions to desktop applications. Just paste the Web Services URL into MS Excel to web-enable a spreadsheet. Create customer gateways for SAP, PeopleSoft and Siebel using the same plug-and-play approach.
- XML, SOAP, WSDL
Access Web Services as a client. Publish your own Web Services API as a host. TierBroker makes Web Services as easy to use as the Browser.
- HTML and the Browser
Web Forms without programming! No CGI, Perl, JSP, ASP, ADO, .NET or J2EE. TierBroker allows you to create complex Web Forms with just plain-text HTML.
TierBroker means leverage, because the code that powers your web site is the same code that runs the nightly batch. There is no duplication of effort to support the five key requirements of data integration.
Self-Contained Solution

The TierBroker Server is a self-contained solution. All of the examples in this white paper use the TierBroker Server in combination with scripting. There is no compiling. There is no app server, web server or platform required.
The TierBroker Server works the same way the web works. The Browser is powered by HTML, which provides structure and JavaScript, which provides program logic. The TierBroker Server is powered by XML, which provides application structure and TBScript, which has the same syntax as JavaScript.
To use the TierBroker Server all you need is the server installation package and a TierBroker project file.
return
to top
What
is TierBroker?

TierBroker is
a toolkit that enables the creation and implementation of middleware.
Middleware is the glue that connects different applications by translating
and transforming data. Applications also use middleware to read
from and write to database managers such as Oracle and Sybase and
messaging applications such as MQ Series.
TierBroker has
two basic components: the TierBroker IDE (Integrated Development
Environment) and the TierBroker Server. The IDE is a powerful programmer's
editor that you use to edit an XML file specifying the details of
converting your input data into TierBroker objects and how and where
you want those objects converted for your output queues. These queues
may be disk files or real-time running processes. The TierBroker
Server, which runs in a tiny footprint on a variety of platforms,
uses the metadata that you specified with the IDE to perform the
actual conversions.
return
to top
Typical
TierBroker Applications

- Extract Transform Load
Data conversion between systems. Reading customer or transaction
files and loading that information into packages such as Siebel,
PeopleSoft,
BroadVision,
and PowerPlay.
- Messaging Applications
Communication within the enterprise. Taking events such as "update
customer address" and propagating that information to the customer
master database, call center or contact management, billing, ERP
or other packages. Use of business rules and application logic
to design the communication between the computer systems within
the enterprise.
- Web-based Transaction Processing
Data exchange using protocols such as TCP/IP, SMTP, POP3, HTTP,
and FTP. Automated e-mail response and notification. TierBroker
can take the output of submit forms, convert it to XML, and rout
it to applications ranging from catalog database queries to shopping
cart transactions. Business rules and application logic can be
applied without Java or C++ programming.
- High Volume Batch Processing
TierBroker can process large volumes of data because it does not
depend on a messaging infrastructure. TierBroker has been used
to process transaction files with hundreds of thousands of records
at speeds greater than 1,000 records per second.
- Personalization and Customer Relationship Management (CRM)
TierBroker business rules can be used in combination with content
management systems to create customer and opportunity driven web
pages. TierBroker has been deployed with BroadVision and is currently
being integrated into CRM applications for broadcast video.
- Electronic Billing Presentment and Payment (EBPP)
TierBroker can manage B2B transaction processing including invoices,
vouchers, and statements with on-line transaction review and error
handling capability. TierBroker can be used to transform information
between markets and portals such as CommerceOne
and Ariba.
return
to top
Business
Benefits

- Greatly reduces time-to-market and cost of implementation.
Usage of TierBroker has shown that solutions can be achieved at
40% or less of the effort normally required.
- Applications can be created without software engineers
because the TierBroker does not use compiled source code.
- Business logic is separated from workflow so that applications
can be assembled and tested in a software factory using spreadsheets
to simulate real time message flows.
- XML native tool. Requires no C++ or Java coding. TierBroker
can be implemented without senior programming resources in a fraction
of the time it takes using competing products.
- Requires no expensive infrastructure. TierBroker can
use messaging environments, such as MQ Series, CORBA and COM.
However, it needs none of these technologies to operate. The TierBroker
Server operates in Windows 98/NT, Solaris, HP/UX and AIX with
ports to other platforms planned for the future.
- High Volume, High Throughput. TierBroker has been used
to implement some of the largest and most complex data integration
projects in the capital markets.
- Allows opportunistic development. TierBroker can be used
for specific data integration requirements. Most competing technologies
are sold as an end-to-end solution only; their cost structure,
both in terms of licensing fees and implementation requirements,
makes them practical only if they are implemented across the entire
enterprise. Instead of being an expensive "complete" solution
that forces you to replace you entire system, TierBroker is a
reasonably priced tool for integrating existing system components
from the latest cutting-edge technology to legacy mainframe databases.
return
to top
Product
Architecture

- Object Oriented Projects
The unit of work in TierBroker is the project. A project may inherit
other projects, and multiple inheritance is supported for using
several libraries of TierBroker metadata. TierBroker has an object
oriented philosophy toward inheritance, reuse and distribution
of applications. Projects have version control and effective dating,
which allows an application to be examined or run for a specific
version and asOfDate.
- Objects, Maps and Workflows
There are three moving parts in TierBroker. Objects, which are
data structures, Maps, which represent transformations between
classes of objects, and Workflow, which combines objects and Maps
with real world sources and targets. Each of these items has attributes
that may be configured within the TierBroker IDE.
A good example of an attribute is the txt:TextType attribute of
the TextFile Queue type. The txt:TextType indicates what type
of text (CSV|XML|Plain) the queue should expect to process.
- Workflow in TierBroker
Within a project, Workflow specifies the operation of the server
at runtime. Within a single workflow, input and output queues
indicate the source and destination of objects at runtime. Even
a relatively complicated workflow, such as the one shown below,
is still just a computer program that reads a file and generates
output that is then directed to an XML file, a database, and another
process that transforms that information and writes it to a NULL
Queue for testing purposes. When you use your mouse to assign
objects, maps, and queues to a workflow, you can assemble complex
multi-queue transformations much faster than you could by coding
in a traditional programming language.
- Data Definition
TierBroker uses XML as its primary data definition language. You
can use XML DTD declarations to describe the structure of your
input, regardless of its format, to TierBroker. However, object
structure definitions may also be imported from a variety of sources
including MS query files, SQL query statements, SQL tables, existing
XML DTDs and TierBroker Object Definition Language. Object structure
definitions can also be defined within the TierBroker IDE by creating
a new object and adding fields.
- Object Relational Model
TierBroker has an object relational architecture. Single inheritance,
is-a and has-a relationships are all supported. References between
objects are defined by keys that may combine data attributes and
constant values. For example, the relationship between a swap
trade header and the floating rate leg might be written as: (tradeLeg.tradeId
= 10023 AND tradeLeg.legType = "F"). This looks like an SQL "where"
clause, but it is not a complete SQL statement, nor does it assume
that the object resides in an SQL database. Instead, it's a loose
where clause, because it combines the syntax of the SQL where
clause with the properties of many persistent datastore types.
- Mapping and Transformation
Mapping, transformation, validation and enrichment are the basic
properties of middleware. TierBroker Import Maps specify how to
transform one object type into another. Transformation maps do
not specify many-to-one relationships by themselves; instead,
a map specifies the relationship between one class of objects
and another. Applications may combine maps in workflows to support
one-to-many and many-to-one relationships. TierBroker applications
can also combine maps with script-based business logic to support
content-based many-to-one transformations. The transformation
of transactions into journal entries is one example of this capability.
Creating complex mappings is as easy as filling out a form with
the TierBroker IDE. The following screenshot shows how a TierBroker
user mapped four EVENT_D input fields to an ODS_EVENT_OCCURRENCES
TierBroker object along with five new fields of calculated information:
- Business Rules
TierBroker applications use business rules to specify data enrichment
using lookups and decodes, routing decisions involving queues,
and conditional relationships between objects and transformation
maps.
- Application Logic
TierBroker can extend the power of Maps, Workflow and business
rule scripting using a variety of techniques. XML messaging, stored
procedures and user exit functions are all supported.
- Real Time vs. Batch Processing
In TierBroker, data transformation and mapping requirements are
fully abstracted from the implementation of real time versus batch
processing. The same objects, business logic and transformations
that are tested and developed for real time transaction processing
over the web can also implement high volume batch processing of
files obtained via FTP. TierBroker allows objects and transformations
to be applied to both real time and batch processing.
TierBroker
allows objects and transformations to be applied to both
real time and batch processing |
Some products claim to have the same capability. What they really
mean is that they can do one of two things:
- Apply
message-based architectures (generally CORBA, JMS or MQ Series)
to batch files by loading each file onto the message transport
layer one record at a time.
- Shoe-horn
real time transactions into a batch architecture by loading
messages into a temporary file or database table.
TierBroker takes a different approach by implementing high speed
native connections to both batch and real time data sources.
- The Information Bus--Not the Magic Bus
Real time message-based architectures are great for distributing
on-line transactions. Even high volume real time systems rarely
need to process more than 50 transactions per second. Very few
e-commerce web sites support more than 180,000 transactions per
day, which is the implication of 50 transactions per second for
just one hour! However, many corporate reporting applications
have consolidations that involve millions of records. For example,
calculating the global currency exposure of a financial institution
on a currency/counterparty cross product basis can involve the
aggregation of hundreds of thousands of complex positions. In
other words, trying to process a 500,000 record file using a message-based
architecture is a really bad idea. However, this is the approach
used by many message-based tools, which boast of adapters to legacy
database and flat file formats.
- Native RDBMS Support
TierBroker can communicate via ODBC easily. If you're using Oracle
or Sybase, however, and you need real speed, you don't have to
use ODBC--you can use the built-in drivers for these database
managers. Drivers for other popular RDBMS such as Informix and
MS SQL Server are also available.
- Transaction Support
Batch processing and real time systems have different transaction
boundary requirements. TierBroker imposes no arbitrary transaction
commit points. Instead, the developer selects application level
transaction boundaries, which are based on the packet structure
of the data stream.
- TierView
In addition to letting you monitor the transaction status of the TierBroker Server as it performs production workflows, serves as a powerful debugger for applications in development. You can set breakpoints and then, when execution is interrupted, examine data and perform commands at the debugging prompt before resuming execution.
return
to top
E-mail
or call us at (212) 607-7580 for more information on what
TierBroker can do for your system.
|
 |