Is It Time To Look Beyond Mainframes in the Hyperscale Era?

If you say the word ‘mainframe,’ plenty of people think of the word dinosaur – big and powerful technology that is more-or less-extinct. They think of technology designed and built in the late 1960s that’s still carrying on today, even though its death has been predicted since the 1990s. Whereas, if you say the words ‘cloud‘ and ‘hyperscale,’ many people immediately think of the future. They think of the ability to access their applications and enable remote working. And they think of how easy it will be to scale up or down to suit their business needs, to be able to store, manage, and access growing amounts of data, and to utilize large amounts of compute infrastructure and applications whenever they’re needed. You can almost picture the sunlit uplands and the swelling musical backing as they picture the scene.

So, let’s take a look at these dinosaurs and see what has kept them in use at 70% of the Fortune 500 companies. The first big myth about mainframes you hear people say is that they are old technology. It is often assumed that a mainframe from the 1960s would look pretty much the same as a mainframe from 2021. That’s a bit like saying a car from the 1960s would have a similar engine to one from 2021 or a jet plane would be exactly the same. Not only has mainframe hardware improved all the time – IBM typically announces a new model each year with new features – but so has the software. The only thing that has remained the same is that older programs – the applications – still run on the latest hardware.

Learn More: How to Make the Mainframe PCI Compliant 

Solving the ‘Legacy’ Dilemma

In many ways, mainframes were the first cloud computers. This shocking statement is borne out by the fact that a centralized mainframe computer site could run applications and store the data for any number of users who could remotely connect to the mainframe whenever they needed to. This is very similar to running applications from a cloud provider and storing data in the cloud – just 50 years earlier.

I can almost hear you say, “yes, but they didn’t scale, did they?”. That’s the thing, they could, and they can. If an application needs more processing time, it gets it. And if more storage space is required, it can be accommodated.

You may hear people say that the mainframe is not using the latest chips to process work. This may be true for some mainframes at some sites. Still, researchers Arcati Opens a new window assert in the Arcati Mainframe Yearbook 2021 that “typically, users have upgraded on a regular basis to the most recent hardware to take advantage of capacity increases and cost benefits.” So, it would seem that the majority of mainframe users are using a chip in their mainframe that is three years old or less – the z14 or z15 microprocessors.

Another argument against the relevance of mainframes is that they are islands of computing that just can’t connect with mobile, edge, cloud or any other kind of modern computing platform. That again isn’t true. There was a time when companies pretended that they were the only platform in town, but that hasn’t been true for a long time. 

Mainframes use Java and RESTful APIs (Application Programming Interfaces), making it very easy to connect mainframe applications to mobile or anything else. Mainframes are an important part of the API economy and can be a part of new applications that are constructed using parts of existing applications on different platforms. So, mainframes are no longer isolated servers; they are part of a mission-critical IT infrastructure matrix. 

Many people shopping online through a browser on their phone or an app are using a mainframe-based application at the back-end. Anyone who has ever used an ATM to get money from a bank has probably used a mainframe. It’s estimated that 90% of ATMs link to IMS running on a mainframe.

Learn More: Mainframe Mayhem: Here’s Why Government Agencies Need to Pursue Modernization 

Setting the Record Straight

IT security is a big issue – perhaps even more so with so many people working remotely. Mainframes are at the forefront of security with the introduction of pervasive encryption a few years ago. It means that data is encrypted while at rest as well as in flight. This level of encryption is not available with cloud computing, no matter how hyperscaled it may be. With z/OS® Encryption Readiness Technology (zERT) software, it becomes possible to check the cryptographic security attributes of IPv4 and IPv6 application traffic being sent using TLS/SSL, SSH and IPSec cryptographic network security protocols. This helps maintain the security of data in transit. IBM also introduced Data Privacy Passports with their z15 processors, which provide transparent, end-to-end, data-centric protection and privacy.  Users can encrypt their data and grant and revoke access to it, no matter where it is.

With cloud computing, there are issues with security and unauthorized people gaining access to data stored in the cloud. There can be issues with insecure APIs used by companies, making them open for hackers. Phishing attacks can lead to passwords being stolen and then to cloud accounts being hijacked. There’s the issue of visibility because cloud-based resources are outside the corporate network. Cloud data is designed to be easily shared – and that could be with a hacker. Hyperscaling an organization’s cloud usage can worsen the security issues associated with it.

Many people assume that the mainframe is still running the same old applications that it was running 20 years ago, or 30 years ago, or even 50 years ago. To some extent that is true. One of the features of mainframe technology is that old applications will still work. But just because old applications do work doesn’t mean that all the applications in use are old ones. The truth is that DevOps is as much a mainframe technology as any other platform. Using agile techniques, new applications can be created by developers and tested by users so that exactly what’s needed can be quickly produced. Also, updates can be produced in months rather than years. IBM subsystems like CICS and IMS are updated quarterly.

Another wrong assumption is that all mainframe applications are written using COBOL or some other programming language that your dad or your grandad would have programmed in. Again, that’s not the case. Many new mainframe applications are written in Java. Similarly, it’s assumed that mainframes are so arcane that no-one under 50 can understand how they work. Again, that’s not true. There’s ZOWE, which is an integrated and extensible open-source framework for z/OS. It offers modern interfaces to interact with z/OS and allows users to work with z/OS in much the same way as they would with cloud platforms. And that means young people, straight out of university, can use mainframes.

Is the End Near? The Answer Is No

If hyperscaling is so good – and mainframes are so good, what’s the future? Both offer the ability to deliver computing as a utility. Both can store massive amounts of data, which is growing all the time. Both allow speedy application development. Both allow projects to expand their capacity. New digital services are likely to rely on one or the other or both. Transaction processing works well on mainframes, and mainframes are used for this by financial services, insurance, transportation, and many other major industries. The big difference is that mainframes are mainly used by larger organizations. And cloud can be used by any company. However, it is probably larger organizations that would need hyperscaling. But, it’s not an either/or decision for many companies; there’s no reason why they couldn’t make use of both computing platforms.

It looks like not only have mainframes not lost their relevance but also they would make a high-functioning part of a best-of-both hybrid design (referred to as a multi-cloud) with hyperscaling cloud environments in the future.

Do you think companies should migrate off mainframes in a cloud-focused world? Comment below or let us know on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d love to hear from you!