The never-ending and often contentious debate about the mainframe remains quite vocal in the marketplace. Whether it’s the technical mandarins of the mainframe shouting down anyone who challenges their core religious principles or the cloud intelligentsia looking down their noses at this 60-year-old platform, the arguments seem to show little sign of any reasonable compromise.
To clear the air: one can support the cloud without trashing the technology of the mainframe, and one can support the importance of mission-critical applications without disrespecting the cloud models.
Some applications on the mainframe are there because of their historical implementations but don’t necessarily need the hard-core strengths of this platform. There are several ways to simply move these applications to the cloud without any consideration for them fitting into modern, cloud-native architecture. For many applications, this is an entirely acceptable result. Whether you choose to rehost the applications or transform the COBOL to Java, this can yield real value for many mainframe applications. By removing the MIPS consumption of these applications from the mainframe platform, room for the remaining applications to grow, while postponing an upgrade is of real benefit.
Every mainframe organization has applications that can successfully run somewhere else without violating the sacrosanct beliefs of the mainframe mandarins. The problem is that many mainframe organizations see this as, to use an old metaphor, “letting the camel’s nose under the tent”! Letting ANY workload move opens the door for even MORE to move! For these believers, the correct answer is to implement Linux on the mainframe and move applications to this environment that can benefit from the lower costs of the mainframe specialty engines, specifically the ZIIP engines. These engines can run DB2 and Java workload at a much lower cost than traditional mainframe general purpose (GP) engines.
But moving the workload to the cloud is about more than just reducing cost. It’s about agility and elastic scalability.
So, the simplest way to move applications to the cloud, particularly applications written in the COBOL language, is to rehost them, using mainframe emulators from several vendors or transform the code to Java and use a Linux/JVM runtime environment in the cloud.
The first approach preserves the COBOL implementation and, consequently, requires this skill set’s continued availability. The latter eliminates COBOL and uses the Java language, a more readily available skill. The key to the second approach is that the application continues to meet its current service level agreement (SLA) expectations. An organization may be less concerned over the proper structure of the application as they often require very little maintenance.
The code can’t be rubbish, of course. But rubbish code usually won’t perform anyways. This direct approach provides some benefits that the rehosting approach doesn’t. Since this is a 100% Java application, it can take advantage of all the runtime and development technologies that are available today. These technologies are available from a massive software ecosystem, different from the ecosystem associated with rehosting solutions. These solutions tend to be limited since there are fewer of these runtime environments in the market than Java ones.
Software vendors need to invest in the rehosting emulation environments. In contrast, when applications are migrated to Java, you have brought your application to a wide-open market of innovation associated with this environment for many years.
So why would you take the simplest approach to move the mainframe workload to the cloud? In 2020, IDG surveyed 200 IT executives responded to their modernization initiatives (The State of IT Modernization 2020, IDG). In this survey, IT leaders recognized that it is not ALL about the apps or platforms. It’s about creating IT environments that support the business and its goals! However, you support the business and its goals by using technology and applications that enable IT organizations to act effectively to enable business objectives.
Not surprisingly, the same report also noted that 41% of respondents stated they abandoned or delayed their IT initiatives due to competing priorities, lack of a clear roadmap, and privacy/security concerns. If an over-dependence on the mainframe and the applications that run there drives your IT strategy, there is little chance you will overcome the reasons mentioned above for delay. Straightforward advice: use the simplest approach to move workload when it helps move your IT modernization strategy forward. Stop letting the application workload on this platform continually limit your ability to modernize in a way that “supports the business and its goals”!
Taking the simplest approach to moving mainframe workload shouldn’t drive your modernization strategy, but it can certainly reduce the friction associated with the IT modernization project. Moving workload in this way helps solidify the strategy of modernizing to the cloud. The future of IT may be some kind of hybrid cloud model, but organizations need to show progress and not just lip-service to this idea. When the simplest approach to migrating a mainframe application helps leverage the hybrid cloud model, then don’t hesitate to do it. Moving application workload, demonstrating the benefits of hybrid cloud, and showing how modernization can support the business and its goals should be the tactics for any organization’s IT modernization strategy.
– Guest content from Dale Vecchio, Mainframe Modernization Thought Leader
Find Out More
Watch A Modernization Dilemma: Cloud Application –
Or An Application In The Cloud