Last Updated: 7 July 2026


IBM z/VM 7.5 Announcement


 

Overview

Availability date - September 18, 2026

IBM z/VM 7.5 is the latest release of the premier hypervisor for IBM Z and IBM LinuxONE servers. It includes critical enhancements that help empower enterprises modernize their IT infrastructure through improved security, enhanced memory scalability, and simplified management, while continuing to support Linux, hybrid cloud, and containerized workloads. z/VM 7.5 provides a highly secure and scalable foundation for Red Hat OpenShift, IBM Cloud Paks, and traditional workloads.

In addition to enabling access to the Telum microprocessor on-chip AI accelerator, z/VM 7.5 is designed to allow Spyre AI accelerator adapters to be dynamically managed as part of a virtual machine configuration. z/VM enables operating systems that support AI platforms to fully exploit AI hardware accelerators. These environments include Red Hat OpenShift AI, IBM watsonx Assistant for Z and other offerings that complement and exploit IBM Z and LinuxONE infrastructure supported by z/VM.

Building on the robust foundation of z/VM 7.4, z/VM 7.5 introduces key features and enhancements that enable clients to fully exploit the latest IBM Z and IBM LinuxONE hardware innovations.

Key highlights include:

  • Enhanced security, ensuring that software used during IPL is authenticated and trusted through Host Secure Boot.
  • Improved memory management and scalability, which can help clients optimize resource utilization and prepare for future large-memory workloads (memory overcommitment control and real memory scalability infrastructure).
  • Simplified system configuration, reducing administrative overhead (fully integrated RACF, Removable Media Services, and z/CMS the default CMS).
  • Networking enhancements, providing consistent connectivity for core system guests (System Infrastructure VSwitch).
  • Modernization to align with industry standards (multithreading enabled by default for eligible systems, semantic versioning, and “memory” terminology).
  • Deprecation of support (removal of 3380 system-attached device support).
  • Enhanced GDPS Support, enabling customers to extend their disaster recovery scenarios (HYPERSWAP for Guest HYPERPAV).

Virtualization is fundamental to delivering infrastructure as a service (IaaS), which provides the foundation for cloud operations models. IBM continues to invest in z/VM technology to deliver advanced leading-edge virtualization capabilities. This strategy can meet the needs of IT organizations to deliver the foundation for user satisfaction with various types of workloads:

  • Cloud-native workloads based on Red Hat OpenShift Container Platform and IBM Cloud Paks or other container technologies;
  • Noncontainerized workloads deployed using cloud and traditional operations models, including IBM WebSphere, open-source such as MongoDB, or software from independent software vendors (ISVs).

Together with IBM Cloud Infrastructure Center, which helps manage the lifecycle of virtual infrastructure, z/VM provides the foundation for private cloud on an IBM Z and IBM LinuxONE server as part of a hybrid cloud approach.

z/VM virtualization technology is designed to enable the deployment of more virtual servers in a single IBM Z and IBM LinuxONE server than can be achieved on any other platform through legendary scalability, system management, and performance. z/VM supports the following guest operating systems: Linux, IBM z/OS, 21CS VSEn, IBM z/TPF, and z/VM itself, as well as Red Hat Enterprise Linux CoreOS as a part of the Red Hat OpenShift Container Platform. In addition, the z/VM hypervisor helps ensure continuous infrastructure availability by tightly integrating with IBM GDPS resilience-focused offerings and with built-in Live Guest Relocation (LGR) support.

With z/VM 7.5, IBM continues to deliver enhancements to z/VM advanced virtualization technology on IBM Z and IBM LinuxONE servers using the z/VM continuous delivery (CD) model. This CD model offers clients timely support for new technology throughout the life of a z/VM product release using a delivery mechanism that is familiar and time-tested. The z/VM CD model allows schedule flexibility as IBM partners with the z/VM community in the design, development, and delivery of new functions that are not centered around a specific release date.


 

Description


The objective of z/VM 7.5 is to strengthen the proof points that support digital transformation, particularly those related to scalability and efficiency. z/VM is designed to enable the deployment of thousands of Linux servers on a single IBM Z or IBM LinuxONE server.

The following enhancements are planned for inclusion with z/VM 7.5:

  • An Architecture Level Set, requiring an IBM z16 or IBM LinuxONE 4 family, or later, server.
  • Semantic Versioning for CP:
    Introduces a four-element semantic version identifier for the z/VM CP component (M.N.O.P), eliminating mod-level reporting and enabling clearer dependency and compatibility identification. This enhancement will align z/VM with the current industry standard, and allow for easier communication with external teams not deeply involved with the mainframe.
  • Several packaging changes:
    • Removal of the requirement to apply user modifications to DMSNGP/DMS390NG because CMS initialization code now handles that function.
    • Removal of the ability to perform upgrades and second level installations from an FTP server. A CMS minidisk can still be used for these installations. First level installations can still be performed from an FTP server.
    • A new System Infrastructure virtual switch is setup during a traditional installation and then configured later when IPWIZARD is run after the installation. During an upgrade, the System Infrastructure VSwitch is defined but not connected to an uplink port, allowing customers to configure the VSwitch manually in the system configuration file when they are ready to exploit it.
  • z/CMS is the default when IPLing CMS, allowing applications that require z/Architecture to run seamlessly in the CMS environment, without requiring special handling. Users who need ESA/390 CMS must explicitly IPL CMS390. While CMS itself is functionally unchanged, defaults, install behaviors, and packaging will now assume a 64-bit execution environment.
  • RACF/VM enablement has been simplified to use CP Exit-based invocation rather than CP nucleus modification. The optional enablement of RACF no longer requires an IPL, and RACF configuration elements that previously required modification to RACF modules in the CP hypervisor now reside in the system configuration file. Security Manager enablement and configuration are now dynamic, and third-party ESMs can be integrated using standardized exits.
  • HYPERSWAP for Guest HYPERPAV lifts the restriction that prevented guest-exploited HYPERPAV devices from participating in HYPERSWAP. Guests using 1-END minidisks can now benefit from both HyperPAV performance and continuous availability mechanisms that are enabled by HYPERSWAP.
  • z/VM Host Secure Boot introduces cryptographic validation of SAPL during list-directed IPL using signed binaries that are verified by the hardware loader. Secure boot stops execution when verification fails, preventing the execution of untrusted code. After successful verification of SAPL, which is z/VM's stand-alone loader program, the machine loader passes control to SAPL, which validates each module it subsequently loads. This provides foundational support for a subsequent z/VM 7.5 feature pack that will introduce cryptographic validation during IPL of a signed CPLOAD MODULE, ensuring it also is loaded securely.
  • Memory Overcommitment Control adds systematic detection, reporting, and enforcement of memory overcommitment thresholds. Overcommit behavior can be controlled per operation (LOGON, AUTOLOG, XAUTOLOG, VMRELOCATE, DEFINE STORAGE).
  • Real Memory Scalability Infrastructure introduces foundational changes, including a 4 GB minimum real memory requirement and a reserved 2 GB CP region, to support future large-scale memory exploitation.
  • Multithreading will be enabled by default for partitions that are configured with 40 or fewer IFLs and are using vertical polarizationwhen no explicit MULTITHREADING statement is specified, z/VM will automatically enable multithreading using a validated configuration designed to benefit Linux workloads on IFLs.
  • This change improves out of the box performance for supported LPAR configurations for most workloads while preserving existing behavior for systems that do not meet the enablement criteria or that explicitly disable multithreading.
  • Memory Terminology Simplification transitions z/VM product terminology from “storage” to “memory”, providing parallel aliases. Commands, messages, documentation, and configuration support both during transition.
  • System Infrastructure virtual switch automatically deploys a default system virtual switch during installation to support internal infrastructure guests (e.g., TCP/IP). This virtual switch will provide consistent connectivity without requiring clients to create a bespoke configuration.
  • Integrate DFSMS into CMS packages the Removable Media Services (RMS) functionality as part of CMS to simplify automated tape library setup and configuration. This removes the requirement to perform a separate installation for core DFSMS use cases.
  • With z/VM 7.5, support for DFSMS Space Management within Shared File System (SFS) file pools has been removed. As a result, SFS MIGRATE and RECALL functions are no longer supported. This change includes the removal of support for the DFSMS startup parameter within the file pool server's DMSPARMS file, which was previously used to enable migrate and recall processing.
  • Customers currently using DFSMS SFS Space Management must issue DFSMS RECALL to recall migrated SFS files before upgrading to z/VM 7.5. Before issuing the DFSMS RECALL command, customers should ensure there is enough space available to the filepool to accommodate the recalled files in their uncompressed state. The command DFSMS REPORT can be used to list migrated SFS files.
  • Products or solutions that rely on SFS based DFSMS MIGRATE and RECALL functionality will no longer operate when running on z/VM 7.5.
  • Default Changes adjusts several defaults related to minidisk caching, abend handling, multithreading, and performance posture to better reflect modern workload operational expectations, and best practices.

The following additional changes are planned with z/VM 7.5:

  • Fence 3380 from System Use enforces the policy preventing 3380 DASD from being used for system-attached functions (IPL, paging, CP-owned volumes). Guest-dedicated use continues to be supported. Directory support for 3380 minidisks is maintained for Single System Image clusters that include older z/VM members that support 3380s.
  • z/VM 7.5 no longer provides restricted source upon request or access to program listings via VPL (View Program Listings).


 

Statement of Direction


Stabilization of the z/VM Performance Toolkit:
z/VM 7.5 is planned to be the last z/VM release to support the z/VM Performance Toolkit feature. Future z/VM releases will continue to support and enhance the performance Data Pump dashboards previously provided with the Performance Toolkit feature.

Stabilization of z/VM support for the IBM z16 family
z/VM 7.5 is planned to be the last z/VM release to support the z16 family of servers, which includes IBM LinuxONE 4. An IBM z17 family server, which includes IBM LinuxONE 5, will be the required minimum level for future z/VM releases.

Full-part replacement
z/VM 7.5 is planned to be the last z/VM release to provide CMS update files as part of service. Instead, full-part-replacement files will be provided. These full-part-replacement files will no longer contain sequence numbers, which are unique character record identifiers in the last 8 columns of every line of a fixed format file. z/VM source files will be encoded using the CP1047 code page, unless stated otherwise. Tooling in the form of DIFF and PATCH for CMS is provided with z/VM to assist clients or ISVs who still have a need to update z/VM source files.


 

Description of Function Delivered in the z/VM 7.4 Service stream


Timely, client-driven function, previously delivered in the service stream of z/VM 7.4 positions z/VM 7.5 with improved:

  • Hardware with IBM z17 and LinuxONE 5 Support and Exploitation:
    • Network Express Adapter EQDIO Support within the z/VM VSwitch allows customers to configure the VSwitch to take advantage of lower latency and higher bandwidth provided by networking EQDIO devices within their datacenter. EQDIO VSwitch includes Link Aggregation, Multi-VSwitch Link Aggregation, and provides QDIO to EQDIO translation, llowing guests that do not support EQDIO to directly take advantage of it. Also provided is the ability to define simulated EQDIO network interface cards (vNICs) on the virtual switch.
    • Power Consumption metrics provided within z/VM monitor data include entire CPC and LPAR specific level power consumption information. This information includes power metrics for CPU, I/O, and Memory usage. Consumers of z/VM monitor data, such as the z/VM performance datapump, can be enhanced to calculate/approximate guest level apportionment.
    • Spyre Accelerator Adapter support offers scalable capabilities for complex AI models and generative AI use cases.
  • Usability:
    • New support in GetShopz provides for direct to host transfer of service packages from IBM ServiceLink. GetShopz currently has the function to securely transfer z/VM service packages directly from Shopz. This enhancement will provide the same ability for orders from ServiceLink. In addition, GetShopz is enhanced providing support for z/VM Installation media enabling GetShopz to decompress ZIP files containing installation media. Installation media is ordered through Shopz, so clients using GetShopz can make use of Direct-to-Host transfer in the same way they would for service packages.
    • Diff and Patch Tool for CMS is an implementation of DIFF and PATCH using the unified diff format. The DIFF utility generates a diff file from two input files, while the PATCH utility applies the changes from an input diff file to an input source file. Diff files can be distributed and applied to other copies of the original source file to apply these changes, much like the XEDIT update facility.
  • Hypervisor Efficiency and Scalability:
    • Active Drain for PAGE Volumes allows actively moving all pages from in-use PAGE slots on a volume marked DRAIN to other, non-draining, PAGE volumes. This allows PAGE volumes to be fully drained and removed without a system outage.
    • Control-only Access for Guest Crypto Domains allows APDED virtual machines to be given usage+control and control-only access to crypto domains. Without this feature only usage+control access to crypto domains is permitted. The new control-only designation is made per domain and applies to the domain across all APs assigned to the virtual machine. Providing control-only Access for Guest Crypto Domains enables virtual machines to submit key management requests to those domains.
  • Security:
    • CERTMGR enhancements provides expanded certificate management capabilities in z/VM, including the use of stash files in GSKKYMAN to eliminate repeated password entry and enable easier automation. New functions such as receiving and renewing certificates have been added to the GSKKYMAN interface, while CERTMGR itself now supports receiving, renewing, importing, and exporting certificates.
    • z/VM HCD has been enhanced in accordance with IBM's Security and Privacy by Design Principle to support TLS in the communication between HCM and HCD on z/VM.
  • Simplification:
    • A new system configuration option has been added to allow a z/VM system to IPL without adding crypto resources to the shared pool allowing all crypto resources on the system to be available for dedicated use.
  • RAS:
    • RAS updates to z/VM FICON support are provided to mitigate impactful hardware and software issues and improve diagnostics. z/VM implements the link-recovery-threshold facility on a system where this support is available. This prevents a channel path with a flapping link condition from tying up system resources by disabling and re-enabling the channel path repeatedly, which can potentially cause a system hang condition. Instead, the facility makes the channel path not operational, thus avoiding unnecessary work for the system. In addition, a new error message will inform clients about why the path is not operational, providing suggestions about what might be wrong on the FICON link and what needs to be fixed.
  • Resilience and usability ease of use enhancements for the z/VM Systems Programmer. Enhancements provided in the area of:
    • Soft Abend Improvements
    • RACF Utility Enhancements
    • Simplified DirMaint RACF Connector
    • Systems Management Enhancements for renaming a userID
    • SMAPI and DirMaint Enhancements

For a detailed description of function delivered in the z/VM 7.4 service stream, see the z/VM New Function page.


 

z/VM Continuous Delivery model


IBM will deliver most future new z/VM 7.5 functions as Feature Packs in the service stream. When z/ VM 7.5 becomes available, z/VM 7.4 will, with a few exceptions, receive corrective service only. The most notable exceptions from the provided corrective service for z/VM 7.4 will be required features for hardware toleration and security compliance. Similarly, when a new release is introduced after z/VM 7.5, Feature Packs will be delivered on that release going forward and z/VM 7.5 will receive mostly corrective service only.

The z/VM Continuous Delivery News web page is the primary vehicle used by IBM to describe new functions that are planned for z/VM. It is the recommended way to become aware of future development and support plans for z/VM. IBM recommends subscribing to this page to be notified of changes. To subscribe, click the Notify Me link in the left-side navigation bar. Additional instructions are included on the VM Site File Change Notification web page.

For a detailed description of z/VM capabilities, see the z/VM website.

IBM has a long history of working with clients to deliver capabilities to improve z/VM. IBM will continue to:

  • Enlist z/VM clients as "Sponsor Users" to advise IBM throughout the design process for many z/VM development projects. These clients can also test early versions of the new support prior to its delivery to the marketplace to make sure the user experience matches their expectations.
  • Invite clients to become involved with the z/VM Continuous Delivery process. IBM finds the Sponsor User relationship to be beneficial to delivering new functions for z/VM and continues to encourage z/VM clients to participate in the continuous delivery process. To learn more about the z/VM Sponsor User program, see the z/VM Sponsor User program web page. For details about z/VM client communication and collaboration, see the z/VM Council web page.
  • Publish information about many of its z/VM development projects as they unfold to help users decide if they want to volunteer as Sponsor Users for selected development projects and help the community at large plan for the introduction of new z/VM functions. This level of communication between IBM and the z/VM user community facilitates discussion regarding implications of the planned support, such as operational incompatibilities, software vendor impacts, and changes to system behavior.


 

Product positioning


z/VM is the leading platform virtualization product for IBM Z and IBM LinuxONE servers and was built on and expanded from the fundamentals of virtualization to help optimize IT environments.

Virtualization is the entry point to build flexible and cost-effective environments for cloud and traditional operations models that can differentiate service delivery to provide an enhanced client experience. In essence, it can become a key control point in the service delivery value chain. The ultimate choice of hypervisor and its management capabilities heavily influences follow-on decisions about products that are selected to support implementations of cloud. The objective is to continue to provide a launching point with z/VM for a new technology base and a deployment and consolidation platform for Linux and Red Hat OpenShift.

With z/VM, together with IBM Cloud Paks based on Red Hat OpenShift and additional container technologies based on Linux, IBM Z and IBM LinuxONE provide a highly scalable, secure, and efficient on-premises private cloud infrastructure. z/ VM also continues to provide these same capabilities for noncontainerized workloads in a traditional operations model that may include products from select vendors.

z/VM 7.5 can help extend the business value of IBM Z and IBM LinuxONE technology across the enterprise by integrating applications and data, while providing exceptional levels of availability, security, and operational ease. This applies not only to private cloud deployments of containerized and noncontainerized workloads, which can be part of a client's hybrid cloud approach, but also to world-class virtualization technology that is offered by z/ VM, which can provide the ability to host large numbers of virtual servers running different operating systems on IBM Z and virtual Linux servers running on IBM LinuxONE.

Typical workload scenarios are the new/additional deployment and large consolidation of:

  • Cloud-native workloads based on Red Hat OpenShift and IBM Cloud Paks or other container technologies
  • Noncontainerized workloads that require high throughput, such as IBM Db2 Warehouse, IBM Db2 LUW, Oracle database, MongoDB, PostgreSQL, and others
  • Noncontainerized applications deployments, such as those based on WebSphere, ISV applications, and others, including open source.

Because virtualization is the entry point to a cloud operations model, the hypervisor choice that clients make heavily influences their implementations of cloud throughout the stack. The IBM strategic objective is to continue to offer IBM Z and IBM LinuxONE servers as the core of trusted digital experiences, together with z/VM as part of the IBM open hybrid cloud approach, as well as for large-scale Linux deployments.