‘Cloud Smart’ Agencies Don’t Outsmart the Mainframe—They Automate It
Overview: Learn why government IT must be Cloud Smart (not Cloud First) and use mainframe automation tools to mainstream the platform and leverage it with the cloud.
A recent article in NextGov featuring several predictions from industry leaders on the direction of government IT makes it clear: 2019 will be a year of modernization. We saw this shift coming in 2017 with the passing of the Modernizing Government Technology (MGT) Act, generating buzz around how to use funds to migrate workloads from on-premises legacy systems to the cloud.
“Cloud First,” policy introduced during the Obama administration, has become a romanticized state of digital nirvana for many CIOs of large, established enterprises and government agencies who wish to migrate as many workloads as possible from on-premises computing platforms to the cloud. But it’s really a “cloud only” strategy that CIOs should be wary of adopting outright.
Instead, they should focus on becoming “Cloud Smart,” as Federal CIO Suzette Kent says and I recently echoed in an article I wrote for GCN. While Kent encourages agencies to continue leveraging the cloud more, they should also “ensure that the technology fits the mission,” as FedScoop reported. In this way, we see Cloud Smart as being “Citizens First” because it shifts the focus from technology projects to creating valuable products for the people agencies serve.
Cloud: Not Bad, but not Everything
Off-prem, light-weight, scalable and modern, the cloud is certainly a superior alternative to the ungovernable, disorderly tangle of x86 infrastructure in most data centers. Large organizations should embrace cloud computing if they expect to compete in the digital age.
However, a Cloud First mentality won’t help government agencies deliver more value to citizens who are worn out from the digital impediments that agencies can’t seem to rise above. Fortunately, more thought leaders in public sector IT, such as Russel Davis, CTO and COO of Vcinity, Inc., are flagging Cloud First as problematic. Here’s Davis’s prediction for 2019 in NextGov:
We predict that the government will drop its ‘Cloud First’ position…
…in recognition of the high costs of data egress, the evolution of hybrid cloud architectures, lower costs of hyper-converged infrastructure solutions, and better utilization of geographically disparate data and compute resources that operate on data where it’s generated or at the edge.”
The Need for the Mainframe
Despite Kent’s, Davis’s and others’ calls for government agencies to become Cloud Smart, there remains a push for Cloud First in the public and private sectors. As for the former, it seems there’s a misunderstanding of the MGT Act for some agencies, which states a key use of funds should be “to improve, retire, or replace existing information technology systems in the covered agency to enhance cybersecurity and to improve efficiency and effectiveness,” the keyword being improve.
The mainframe is unequivocally one of those systems agencies can and should improve to deliver a better experience to citizens alongside the cloud. One prime example is the state of Wisconsin, whose CIO, David Cagigal, told GCN “mainframes are here to stay” because states like Wisconsin need a secure, reliable platform capable of processing billions of transactions a year.
These transactions are increasing as agencies develop new web-, mobile- and cloud-based applications that serve citizens’ needs better. Those applications often require access to mainframe applications and data to be effective.
But while the mainframe is technologically modern and irreplaceable, it often lacks the modern culture, processes and tools required for it to function in an agile cross-platform ecosystem. In this way, its necessity makes it a constraint to the modern technologies of engagement citizens want so bad.
Improving development on the mainframe and, therefore, the delivery of innovation to citizens, requires establishing the right foundation for mainframe developers to increase quality, velocity and efficiency. Mainframe automation tools are critical to this effort.
Mainframe Automation Tools Are Critical
Modern mainframe automation tools are what enable modern culture (collaborative, communicative, transparent) and processes (Agile development and DevOps CI/CD, continuous improvement, etc.) to thrive on the platform. The following are key areas where agencies can and should begin this improvement.
Currently, testing is primarily a manual process Operations supports. In manually writing tests or creating test data, these folks insist on achieving a bug-free implementation. While this is an honorable mission, it slows things down, it’s error-prone and it requires too much specialized knowledge.
Automated testing is critical to mainframe DevOps. By shifting testing left, automation brings testing closer to the origins of code and, therefore, the origins of problems that can be solved faster with more clarity, reducing cycle time and improving quality.
Source Code Management Automation
While older version control tools provide some basic functionality around configuration tracking, environmental dependencies, operating system details and hardware, agencies need a tool that can drive a delivery pipeline with automation.
A modern SCM tool equipped with automation will eliminate manual steps to help mainframe teams iterate quickly throughout the SDLC by integrating into a DevOps toolchain with modern development and testing, ITSM and other tools accessible through REST APIs. Automating much of the underlying process of moving code through a pipeline accelerates the process of analyzing, editing, building, testing and deploying it.
Build and Deployment Automation
Again, applications span platforms and require not only systems like the cloud but also systems like the mainframe to be effective—hence developing a cross-platform DevOps strategy. This makes automated cross-platform code deployments essential.
A mainframe tool that automates, standardizes and monitors deployments throughout every phase of a pipeline across platforms, then gives you a way to visualize that process, makes it possible to see where bottlenecks occur and address impediments. It also eliminates the manual process of what should be a repeatable step in your delivery pipeline for real speed and efficiency.
Automated Batch Processing
As mainframe workloads grow from citizen-facing advancements in web, mobile and cloud, those workloads become more difficult to manage without increasing costs—especially as batch system management experts retire.
A mainframe tool that automates and optimizes batch cycles will make it easy to define business objectives and priorities and rest assured OLTP workload remain intact. Automation will also help you reduce monthly license costs and even defer costly mainframe upgrades.
Citizens First Needs Automation
Moving forward into 2019, there is an ever-increasing focus on the federal government’s effective use of citizens’ tax dollars and the system delivery mechanisms that continue to bring value to those we serve. The mainframe is a critical platform that can’t be given up on or replaced—not even by the cloud. Instead, the mainframe must be improved upon. Doing so will help your agency remain on the path of being Citizens First.
Without improving your toolchain with mainframe automation tools that increase quality, velocity and efficiency, your application development teams will be crippled, and the citizens you serve will be left unsatisfied with what they’re receiving in return for their tax dollars. Focusing on automation across your DevOps architecture will support the modernization goals for your federal agency in 2019.