Google Cloud plans to introduce what it’s calling a simpler, more risk-averse way for enterprises to move their legacy mainframe estates to its cloud with a new service built on technology originally developed by Banco Santander.
That service is Dual Run, and it enables parallel processing, allowing enterprises to make digital copies of their legacy mainframe systems and run them simultaneously on Google Cloud Platform. The service addresses a big challenge with mainframes: the tight coupling of data to the application layer. It allows real-time testing by customers to ensure their cloud workloads are performing as expected, running securely, and meeting regulatory compliance needs — without stopping an application or negatively impacting their end-user experiences — before transitioning to GCP as their primary system.
“This is a simple concept, but hard to implement — hasn't been done so far,” Nirav Mehta, Google Cloud’s senior director of product management for cloud infrastructure solutions and growth, told Protocol. “It will substantially reduce the risk of moving mainframe applications to the cloud.”
Nirav Mehta, Google Cloud’s senior director of product management for cloud infrastructure solutions and growthPhoto: Google Cloud
Dual Run uses virtual machines on GCP to create parallel instances of mainframe workloads. A launcher/splitter includes the required mechanisms to duplicate activity — and return the “primary” system response — at each of the architectural interfaces that drive incoming requests or trigger scheduled workloads, according to Mehta.
A real-time monitoring dashboard shows the differences in transaction responses between the mainframe and GCP deployment. A single output hub also ensures one point of contact during the roll-out period for all batch information sent.
And then, once they’re comfortable, customers can use their mainframes for backup or retire them.
“For some time, you can keep your mainframe as the primary system that responds to customer requests and let the cloud instance actually just be a secondary system which will also be running the same request,” Mehta said. “You monitor the responses coming back from the mainframe and from Google Cloud to see if the Google Cloud instance is working just like the mainframe. And then at some point, you switch to making Google Cloud the primary and the mainframe the secondary.”
Dual Run, which is in preview, has been developed for industries including financial services, health care, manufacturing, and retail, as well as public sector organizations. Approximately 90% of the top banks still use mainframes, Mehta said, as do 23 of the 25 largest U.S. retailers.
“All of these companies want to modernize their applications that are running on mainframes and bring them to the cloud for all the benefits from security to scalability, cost-efficiency,” he said. “But because these systems are so mission-critical – and mainframes are especially unique in how long they've been around and how much legacy technology there is in [them] – they perceive a lot of risk, and they don't bring it to the cloud.”
Approximately 90% of the top banks still use mainframes, Mehta said, as do 23 of the 25 largest U.S. retailers.
Banco Santander, a Google Cloud customer, in May reported its progress in digitizing its core banking platform, saying it had migrated 80% of its IT infrastructure to the cloud using software developed in-house called Gravity. Google Cloud has exclusively licensed that technology, and its engineers worked with Santander during the past six months to optimize it for end-to-end mainframe migrations for customers in multiple industries.
“Their use case was very limited,” Mehta said. “We have substantially increased the relevance of the solution to any mainframe customer. It's a big deal for anyone running mainframes to have this option. It may not use unique hardware or anything that is dramatically different than what other vendors might have access to, but it's how we constructed the pieces together. One has to think through every step of that journey and carefully engineer it.”
A presentation deck for Dual Run.Image: Google Cloud
Google Cloud partners including Accenture, Capgemini, and Kyndryl will help deploy Dual Run for customers.
Google Cloud is billing Dual Run as a “first-of-its-kind” service. AWS Mainframe Modernization, meanwhile, launched in June, and supports two primary migration options: replatforming and automated refactoring.
What’s novel about Dual Run is that its workload-dispatching technology and software and data synchronization allow both systems to run on the same data and logic, according to Hauke Heier, who leads Accenture’s Google Cloud business group in Europe.
“In addition, Dual Run for the first time offers Google Cloud as one of the parallel run avenues,” Heier said. “This, in turn, enables extensive observability in one platform, which is critical for heavily regulated industries [such as] banks. As workloads are moved over from the mainframe to Google Cloud, the bank and their regulatory stakeholders are able to systematically demonstrate equivalency of the new run avenue and the mainframe, gaining confidence in Dual Run’s robustness.”
Dual Run also can reduce migration risk by enabling partial workloads/application modules — which are certified to be moved to Google Cloud — instead of full applications, according to Heier.
“For other industries, Dual Run’s ability to handle partial migrations may offer better ROI on modernization projects,” he said. “In any case, Dual Run will address the scarcity of skills for modernizing and maintaining mainframe-based digital cores, since operations on Google Cloud are much simpler.”