Hive Computing and Transaction Processing

Application servers are incredibly capable tools. They save organizations significant time when it comes to building things like the security and presentation components of a system.

However application servers also have their limits.

In particular, application servers are not designed for transaction processing. Instead, most transaction processing is done using TP Monitors like Tuxedo or CICS. The application server is responsible for moderating the interaction between the user and the TP Monitor.

That isn't to say that application servers can't be and aren't being used for transaction processing.

They are.

However, all of the transaction processing code — code for handling issues such as load balancing and fail-over — must be developed by the organization. While this can work, organizations that attempt to do transaction processing in an application server end up spending significant amounts of time and money creating that code.

Hive Computing solves that problem.

Hive Computing enables organizations to perform transaction processing in a manner that is more reliable than in an application server and more affordable than a TP Monitor.

Connecting a Hive to an application server is a straightforward process.

The first step is to create the Applications to be deployed in the Hive and do the actual work. These Hive Applications can be written in C/C++ or Java and can interface with external resources like file systems and databases.

Once the Hive Applications have been developed, the second step is to create the code that interacts with the Hive Applications through the HiveCreator API. 

The advantage of using a Hive to perform these transaction processing functions is that there is no need for developers to write code to handle the failure of any Workers that make up the Hive. The Hive handles that, leaving the developer free to focus on their Application.