The cloud be damned, momentum for mainframes continues.
You read that correctly.
As organizations look to run compute-intensive artificial intelligence (AI) applications, the mainframe is often where they turn. This is due in part to the fact that the mainframe can support huge numbers of transactions, as many as 30,000 a second.
Even in a cloud-centric world, mainframes have retained their relevance because of their transactional throughput, according to Steven Dickens, CEO and principal analyst of HyperFRAME Research. “These are uniquely designed systems; the processor is uniquely designed for high-transaction, low-latency types of workloads, and there really isn’t anything else like that,’’ Dickens said.
Even though cloud usage is expanding, that doesn’t mean there is n0 opportunity for the mainframe to expand as well, he said, adding, “It’s not an either-or.”
IBM is the leader in the mainframe market, followed by Fujitsu, Unisys, BMC, and Broadcom.
Where mainframes shine
Mainframes continue to be used in sectors including banking, insurance, retail, and government. “For large financial services organizations that must process millions of transactions per second, … the mainframe is the only place where they can address that scale at this time,’’ said Forrester Research senior analyst Brett Ellis.
“The mainframe remains important because it provides what businesses require most: reliability, security, scalability, and performance,’’ said Elpida Tzortzatos, IBM Fellow and CTO of AI on IBM Z and LinuxONE. “These features are essential in sectors that require high-volume transactional workloads, real-time processing, and strict compliance standards.”
In healthcare, mainframes enable real-time claims processing while adhering to strict data protection rules such as the Health Insurance Portability and Accountability Act (HIPAA), he noted. “Retailers utilize them to optimize inventory and supply networks on unprecedented scales.”
At a time when cyber risks are on the rise and organizations have increased business expectations, “The mainframe continues to evolve by combining known strengths with cutting-edge technology such as artificial intelligence,’’ Tzortzatos said.
What sets the mainframe apart is its ability to evolve, he said. For instance, the IBM z16 integrates on-chip AI acceleration with the Telum processor, enabling real-time fraud detection and credit risk analysis directly within transactional systems, he said.
“This all happens while maintaining…security and availability,’’ Tzortzatos said. “With features like confidential computing, sensitive workloads are protected within a secure boundary, ensuring data privacy and compliance.”
The numbers attest to its continued worth: IBM reported 6% growth in its mainframe business in the second quarter of 2024. A September Forrester report found that 61% of global infrastructure hardware decision-makers said their firm uses a mainframe. Of those still using mainframes, 54% indicated their organization would increase that usage over the next two years.
The mainframe market was valued at $3,478.6 million in 2023 and is expected to grow at a CAGR of 7.8% from 2024 to 2030, according to Precision Business Insights.
The modernization of the mainframe
The Porsche 911 and Corvette are both the same age as the mainframe, Dickens pointed out, yet “If you were to look at the latest models, you wouldn’t say ‘that’s a car from 1964’.” IBM is keeping the hardware current with the Telum 2 processor, “which will be as cutting-edge as anything launched next year,’’ he said.
The mainframe’s operating system has been containerized to parse applications into smaller, more manageable microservices. In terms of applications, “that’s where it becomes a customer story,’’ Dickens said. “You’ll hear a lot of people say the apps are old; that’s the case of any app launched 20, 30, 40 years ago. If an enterprise hasn’t kept the app up to date, it’s not the underlying platform’s fault.”
There is a misconception that modernizing means moving to the cloud, Dickens said. That’s the right move for some applications, but not all. “It’s a case of looking at the workload, looking at its characteristics and non-functional requirements and saying, ‘What’s the best platform for the workload?’,” he said. “The more transaction-heavy, the mainframe is the right answer.”
Echoing Dickens, when it comes to applications, Tzortzatos said, “It is not about modernizing the mainframe, but about enabling clients to modernize their applications and fully utilize the platform’s advanced capabilities.”
To support this, IBM’s watsonx Code Assistant for Z is a generative AI-assisted product designed to accelerate modernizing mainframe applications at a lower cost and with less risk than today’s alternatives, according to Tzortzatos. “It supports the end-to-end application developer lifecycle with capabilities for application understanding, refactoring, and optimization, as well as code explanation and transformation.”
Privacy is paramount
The mainframe is also the go-to for many organizations that prefer not to keep particular types of data in the cloud, due to concerns about security, compliance, and data ownership. To meet regulatory requirements such as Europe’s General Data Protection Regulation (GDPR) or HIPAA, sensitive data—such as personally identifiable information (PII), healthcare records, and financial transactions—must be kept in a highly protected and controlled environment.
Additionally, data sovereignty and residency laws may require that sensitive information remain within specific geographical boundaries, which can pose challenges for cloud providers that lack data centers in certain regions, making the mainframe a more viable solution for such requirements.
“Maintaining complete control and avoiding risk exposure are major concerns for organizations that use data to gain a competitive edge,’’ Tzortzatos said. “The mainframe is designed specifically to manage this type of mission-critical data, providing unrivaled security, dependability, and scalability.”
Drawbacks
Mainframes do have their shortcomings.
For one thing, they are costly to purchase, operate, and maintain.
Further, systems running legacy applications are typically less flexible than newer technologies, wrote Jet Lü, customer solutions at Amazon Web Services, in a LinkedIn post.
“Making changes to mainframe-based applications or integrating them with modern software and platforms can be challenging and time-consuming,’’ Lu said. “This lack of flexibility can hinder the ability of government entities to adapt quickly to new requirements or take advantage of innovative technologies.”
There is also a shrinking pool of professionals with the skills to operate, maintain, and develop applications for mainframe systems.
Migrating data is more difficult than it sounds, due to privacy and security concerns and interdependent workloads, said Forrester’s Ellis. “Ultimately, if a company is still using the mainframe in 2024, it is because the workloads hosted there are strategically important to the business, and migrating them to another platform introduces significant risks.”
As those risks are addressed in the market, organizations will migrate important workloads, Ellis said, “though I think there will always be a place for a scale-up hardware platform like mainframe, though in 20+ years that platform might look very different than it does now.”
Esther Shein is a freelance technology and business writer based in the Boston area.
Join the Discussion (0)
Become a Member or Sign In to Post a Comment