Data centers continue to see an increase in the mix of mainframes and servers that support operational and line-of-business (LoB) applications. Not only that, but enterprises across the world, particularly Fortune 500 companies, still have a huge investment in legacy code running on mainframes (think CISC and IMS). Because this code may be running on mainframes for a long time to come, systems integrators and application programmers must be able to access those resources, and that requires integration. Simply put, integration with mainframes isn’t optional—it’s a must-have.
Knowing this, let’s break down what integration really means. Conceptually, a mainframe, like any server, is a resource pool offering services such as transaction processing and databases, not unlike a cluster of Microsoft servers running Microsoft Transaction Server (MTS) and SQL Server. From an integration angle, we can then see our needs as:
• Database Integration:
• Transaction Integration:
• User Interface (UI) Integration:
• Security Integration:
In this article, we’ll delve deeper into database integration, saving transaction, UI, and security integration for later.
Because mainframes have such a long history in the enterprise they house a huge amount of mission-critical data. Often, this data is stored within DB2, although it may be in other RDMS products or files. Unlike SQL Server or Oracle, however, DBMS’ on mainframes aren’t really designed with a Windows network in mind. Instead, they are generally inward-facing and used by applications running on the same mainframe.
In other words, a lot of valuable data—all of which could potentially play a much larger role within the organization--is often locked away on the mainframe.
But that doesn’t have to be the case. Integration solutions such as Microsoft’s Host Integration Server (HIS) provide widely supported interfaces to mainframe data. For example, HIS provides both Open Database Connectivity (ODBC) and Object Linking and Embedding DB (OLE DB) interfaces to programmers.
Now, let’s take two examples to illustrate how database integration can be hugely beneficial to organizations that want to get more out of both their mainframe investment and their data.
The Big Report
A large national retail company, Big Mart, has been storing sales and inventory data on its IBM mainframe for years. Furthermore, the company has invested heavily in highly customized code, and it doesn’t want to lose that investment any time soon. However, Big Mart also wants to find a faster way to develop new trending reports on inventory management and customer buying decisions to increase profitability.
In the past, Big Mart developed new reports by contracting programmers that were familiar with COBOL. However, this was both expensive and inflexible. Now, Big Mart wants to leverage the large pool of programming expertise in the .NET world, while still being able to utilize its mainframe.
So Big Mart deploys HIS, which gives them an ODBC interface to their mainframe data. And by utilizing Microsoft’s Distributed Query Processor (DQP), a feature of Microsoft SQL Server, Big Mart provides an enterprise-wide view of their data to their .NET programmers. This allows the .NET programmers to have an anywhere-anytime view of the data that drives the company.
The Little Report
As great as it sounds to develop an enterprise-wide view of their data using HIS and DQP, there are still substantial wins in more targeted use cases. Take, for example, a user in Big Mart’s HR that needs to see a daily listing of new and terminated employees. Like many of the other LoB applications at the company, the HR system runs on the mainframe. Yet, most end-users are more familiar with Windows and Office. This means that users were trained on both Windows and the mainframe: Twice the training, twice the cost.
By leveraging the ODBC interface from HIS instead, Big Mart’s developers are able to develop an Excel spreadsheet application that automatically pulls the latest employee data, updates pivot tables and charts, and allows the HR end-user to manipulate the data in any way she sees fit—all within Windows.
As you can see, mainframes can be tightly integrated into a Windows environment. So tightly in fact that end-users don’t even need to know that the mainframe exists. In this article, we dug into database integration, but in later articles we’ll discuss many other ways of integrating mainframe systems with both servers and desktops.
For more information about interoperability at Microsoft, go to http://www.microsoft.com/interop.