(This post is sponsored by VirtualZ Computing)

Some of the largest enterprises are finding new uses for their mainframes. And instead of competing with cloud and distributed computing, the mainframe has become a complementary asset that adds new productivity and a level of cost-effective scale to existing data and applications. 

While the cloud does quite well at elastically scaling up resources as application and data demands increase, the mainframe is purpose-built for the largest-scale digital applications. But more importantly, it has kept pace as these demands have mushroomed over its 60-year reign, and why so many large enterprises continue to use them. Having them as part of a distributed enterprise application portfolio could be a significant and savvy use case, and be a reason for increasing their future role and importance.

Estimates suggest that there are about 10,000 mainframes in use today, which may not seem a lot except that they can be found across the board in more than two-thirds of Fortune 500 companies, In the past, they used proprietary protocols such as Systems Network Architecture, had applications written in now-obsolete coding languages such as COBOL, and ran on custom CPU hardware. Those days are behind us: instead, the latest mainframes run Linux and TCP/IP across hundreds of multi-core  microprocessors. 

But even speaking cloud-friendly Linux and TCP/IP doesn’t remove two main problems for mainframe-based data. First off, many mainframe COBOL apps are their own island, isolated from the end-user Java experience and coding pipelines and programming tools. To break this isolation usually means an expensive effort to convert and audit the code. 

A second issue has to do with data lakes and data warehouses. These applications have become popular ways that businesses can spot trends quickly and adjust IT solutions as their customer’s data needs evolve. But the underlying applications typically require having near real-time access to existing mainframe data, such as financial transactions, sales and inventory levels or airline reservations. At the core of any lake or warehouse is conducting a series of “extract, transform and load” operations that move data back and forth between the mainframe and the cloud. These efforts only transform data at a particular moment in time, and also require custom programming efforts to accomplish.

What was needed was an additional step to make mainframes easier for IT managers to integrate with other cloud and distributed computing resources, and that means a new set of software tools. The first step was thanks to initiatives such as the use of IBM’s z/OS Connect. This enabled distributed applications to access mainframe data. But it continued the mindset of a custom programming effort and didn’t really provide direct access to distributed applications.

To fully realize the vision of mainframe data as equal cloud nodes required a major makeover, thanks to companies such as VirtualZ Computing. They latched on to the OpenAPI effort, which was previously part of the cloud and distributed world. Using this protocol, they created connectors that made it easier for vendors to access real-time data and integrate with a variety of distributed data products, such as MuleSoft, Tableau, TIBCO, Dell Boomi, Microsoft Power BI, Snowflake and Salesforce. Instead of complex, single-use data transformations, VirtualZ enables real-time read and write access to business applications. This means the mainframe can now become a full-fledged and efficient cloud computer. 

VirtualZ CEO Jeanne Glass says, “Because data stays securely and safely on the mainframe, it is a single source of truth for the customer and still leverages existing mainframe security protocols.” There isn’t any need to convert COBOL code, and no need to do any cumbersome data transformations and extractions.

The net effect is an overall cost reduction since an enterprise isn’t paying for expensive high-resource cloud instances. It makes the business operation more agile, since data is still located in one place and is available at the moment it is needed for a particular application. These uses extend the effective life of a mainframe without having to go through any costly data or process conversions, and do so while reducing risk and complexity. These uses also help solve complex data access and report migration challenges efficiently and at scale, which is key for organizations transitioning to hybrid cloud architectures. And the ultimate result is that one of these hybrid architectures includes the mainframe itself.