Innovative and modern don't often correlate with mainframes, but IBM hasn't given up on big iron yet. Over the past year, the vendor made several major developments to its IBM z mainframes in an attempt to integrate new technologies such as containers and deep learning and accommodate emerging workloads. Still, the future remains unclear as the IBM z mainframe sales decline and the company rebrands itself to focus on cloud-based services, AI and blockchain technologies.
Is the mainframe market dead?
The mainframe market isn't going away yet, despite the best efforts from cloud computing initiatives to lure prospective mainframe users away from the hardware. Mainframes still have a major role in worldwide banking, with roughly $1.5 billion in daily transactions processed on the technology.
IBM continues to push forward with mainframes, as the z14 release marks another move from the company to entice users to stay on the mainframe for enterprise practicality and move to the cloud for future development at lower costs. Increased cloud security played a major role in z14 updates, as IBM doubled down on security to compete against cloud opposition.
Although IBM saw a one-year growth in z Systems revenue by 4% in the fourth quarter, the reasons behind that financial climb aren't necessarily tied to an increase in z14 mainframe usage. How IBM pushes the envelope on its mainframe will play a major role in what happens to the technology in the future.
SCRT sounds like saving money
There's no arguing that it's expensive to run IBM z mainframes, particularly when it comes to power costs. However, there are ways to shave dollars off your mainframe bill.
IBM's latest release of the Sub-Capacity Reporting Tool (SCRT) enables users to rent IBM z mainframe software on a CPU-time usage basis, not total processor capacity. The pricing model is based on IBM's Monthly License Charge (MLC) Software and charges users monthly by examining the four-hour rolling average (4HRA) of top millions of service units (MSU).
Users input type 70 and type 89 System Management Facility records into the SCRT to find the products that they use in hourly intervals. Then SCRT will calculate the MSU 4HRA for each processor and create comma-separated value files for users to evaluate and send to IBM for billing. The SCRT allows users to identify costly and lengthy applications that rack up CPU time and look into ways to cut down on costs.
IBM z mainframes and machine learning
On the surface, Machine Learning for z/OS is purely beneficial; it enables users to perform analytics directly on the mainframe. But potential customers shouldn't overlook some of the technology's major pitfalls.
Scheduling services, user interface and development generally run on z/Linux. The rest of the IBM Machine Learning services run on z/OS, and nearly all of the software is open source. However, it's often expensive to run this software on z/OS, because it requires high amounts of CPU time and can dramatically increase users' MLC expenses. IBM plans to use Spark to avoid this jump, but users still need to be careful of how they run the software.
Data access could also be a problem with machine learning on IBM z mainframes because there isn't a simple process to apply updates to the database. Before you jump on board, wait and see what IBM's plans are to expand machine learning in its mainframes.
Encryption boosts security in IBM z mainframes
Encryption keys protected by cryptographic hardware are at the forefront of IBM's latest effort to bolster security features in its mainframe systems. IBM z14 with Pervasive Encryption, which was released in July 2017, targets enterprises that need extra security layers for sensitive data.
The strategy shift, which was announced nearly two years after the release of the z13, targets enterprises that might not currently own a mainframe, but want to increase data protection and thwart possible hacking attempts. It also targets enterprises that integrate cloud, AI or other advanced technologies, along with large organizations in need of better data security.
Container management on IBM z mainframes
LinuxOne Emperor II adds more security on top of the pervasive encryption in z14, as well as container management. The new system can manage up to 2 million Docker containers, can handle Java workloads and incorporates a host of new capabilities to appeal to a broader range of enterprise users.
Cost and scope narrow LinuxOne Emperor II to large-scale deployments for the time being, but the expansion into the container market allows for IBM to investigate user demand in the previously untapped area. Improvements in blockchain capabilities with pervasive encryption are also at the center of IBM's push to a wider range of use cases in the enterprise. The move showcases IBM's increased security efforts when compared to other security competitors in the market.
LinuxOne Emperor II's Secure Service Container can manage and secure both on-premises and cloud-based applications that can receive the security features of the new platform without affecting the underlying software. This move frees up IT to focus on integral areas rather than develop advanced security features into internal applications.
CA Technologies moves away from mainframes in favor of the cloud
What does a young mainframe developer look for?
IBM's Strategic Imperatives highlight company transition