Otimização de Mainframe
IBM z17™: Advancing AI-powered innovation, resilience and growth
We’re proud to join IBM in announcing the availability of the IBM z17™ single frame and rack mount systems, representing the next advancement in the IBM Z platform.
As organizations continue to scale their use and implementation of AI, the need for a trusted, secure, and high-performing platform has never been greater. IBM z17™ is designed to help enterprises harness the full potential of AI while maintaining the reliability, security, and scalability required for mission-critical operations.
Building on the strength of the IBM Z platform, IBM z17 empowers organizations to unlock new opportunities for innovation, streamline operations, and protect their most valuable assets in an increasingly complex digital landscape.
Accelerate growth with AI at the core
The AI story on IBM z17 begins at the processor level. The IBM Telum II™ processor helps clients leverage composite AI techniques and a multi-model approach for real-time inferencing within a transaction. With this foundation in place, AI becomes an embedded capability powering faster, smarter decision-making at the moment it matters most.
The new processor design delivers an impactful increase in performance. In benchmark testing, IBM validated “With IBM z17, process up to 450 billion inference operations per day with 1 ms response time using a credit card fraud detection deep learning model.”1 The IBM Telum II processor on the z17 is also designed to seamlessly scale peak AI workloads, as each core on the chip can access each of the 8 integrated accelerators for AI.2
For enterprises running the world’s most demanding workloads, this performance opens powerful new possibilities. The enhanced AI capabilities of the IBM z17 enable organizations to unlock new Agentic AI services and modernize faster to free up development resources. By integrating AI directly into the core of the enterprise, organizations can move from experimentation to execution and transform insights into measurable business value.
Transform operations through intelligent automation
Operational efficiency remains a top priority for organizations seeking to maximize performance and agility. IBM z17 is purpose-built to drive operational efficiency through advanced automation and intelligent capabilities.
Running an AI-infused OLTP workload on OpenShift Container Platform with IBM z17 requires up to 5.6x fewer cores compared to a comparable x86 platform.3 This level of efficiency significantly simplifies scaling across the mainframe environment. IBM z17 helps organizations slash operational complexity, expedite insights, and eliminate the blind spots that slow decision-making. The result is a system that expands what’s possible while reducing the overhead required to sustain it.
These capabilities enable organizations to improve productivity, increase responsiveness, and build a more resilient operational foundation for the future while maximizing the value of every infrastructure investment.
Protect critical data with enterprise-grade security
As cyber threats grow in sophistication and regulatory requirements tighten across industries and geographies, organizations require a platform design to help safeguard sensitive data and maintain trust. The robust security architecture of IBM z17 helps enterprises strengthen cyber resilience, increase transparency to reduce risk, and simplify security and compliance operations.
Embedding security throughout the platform, from the hardware to the application layer, continues to be a focal point for the IBM Z ecosystem. For industries like financial services, healthcare, and government, where trust underlies every transaction and interaction, IBM z17 delivers the enterprise-grade security that modern organizations need to move forward with confidence.
Clients benefit from Software AG commitment to IBM Z
Software AG is excited about the opportunities IBM z17 creates for our clients seeking to accelerate AI adoption, modernize critical workloads, and strengthen operational resilience. As a long-standing member of the IBM Z Community, Software AG’s commitment to the IBM Z platform ensures that enterprises that rely on mission-critical applications built on the Natural application platform and Adabas database can immediately utilize new IBM z17 models.
Early adopters of the IBM z17 Enterprise class, many of our clients have experienced first-hand the exceptional performance for workloads that are the backbone of their operations. The new IBM z17 single-frame and rack-mount models give Software AG clients additional ways to make more effective use of their data center space.
Among our IBM Z users, enterprise AI as a topic of conversation has grown and transformed: experimentation has given way to execution. The question is no longer whether AI belongs alongside mission-critical operations; it’s how to use it reliably, securely, and at the scale those operations demand. At Software AG, we support these ambitions with CONNX., a data access and virtualization layer built specifically to make operational enterprise data — Adabas, VSAM, IMS, Db2, and other systems of record — reachable by modern AI workloads.
CONNX exposes these data sources through the Model Context Protocol natively, preserves source-side governance across hybrid environments, and is engineered to keep MSUs overhead proportional to the value of the query rather than the volume of the reads. The power of CONNX lies in its ability to provide real-time access to systems of record, which is essential for agentic AI.
The unified vision
IBM z17 represents an important evolution of the IBM Z platform, bringing together AI, automation, performance, and security in ways that help organizations innovate with confidence.
Together with IBM, Software AG remains focused on empowering organizations to embrace the opportunities of the AI era while preserving the reliability, scalability, and trust that have long defined the mainframe. We look forward to helping our clients transform today’s investments into tomorrow’s competitive advantage.
Read the IBM press release aqui
Disclaimers:
[1] Performance results are extrapolated from IBM® internal tests running on IBM Systems Hardware of machine type 9175. The benchmark was executed with 1 thread performing local inference operations using an LSTM-based synthetic credit card fraud detection model to exploit the Integrated Accelerator for AI. A batch size of 160 was used. IBM Systems Hardware configuration: 1 LPAR running Red Hat® Enterprise Linux® 9.4 with 6 IFLs (SMT), 128 GB memory. 1 LPAR with 2 CPs, 4 zIIPS, and 256 GB memory running IBM z/OS® 3.1 with IBM z/OS Container Extensions (zCX) feature. Results may vary.
[2] Each IBM integrated Accelerator for AI on IBM z17 Telum II processor is shared by all cores in the same drawer. This feature enables a core to direct AI operations to any of the 8 AI accelerators in the drawer if AI accelerators are not currently in use by another core. On IBM z16, cores can send work to only one integrated accelerator for AI on the same chip.
[3] Performance results are based on IBM® internal tests running on IBM Systems Hardware of machine type 9175. The MegaCard benchmark was containerized and deployed on Red Hat® OpenShift® Container Platform (RHOCP) 4.17 on Red Hat Enterprise Linux® (RHEL) 9.4 with KVM. 3 RHOCP Compute nodes, each with 16 pinned vCPUs and 32 GB memory, 2 RHOCP Infrastructure nodes, each with 4 pinned vCPUs and 12 GB memory, 3 RHOCP Management nodes, each with 4 pinned vCPUs and 16 GB memory, running in virtual machines. The 3 RHOCP Compute nodes ran 3 MegaCard instances in parallel, each driven remotely from Apache JMeter™ 5.2.1. IBM Systems Hardware configuration: 1 LPAR with 5 dedicated IFLs (SMT), 192 GB memory. Compared x86 Server Hardware configuration: 28 Saphire Rapids Intel® Xeon® Gold CPUs @ 2.20 GHz with Hyper[1]Threading turned on, 1 TB memory. Results may vary.
Most Batch Problems are Peak Problems: The Hidden Cost of Mainframe Bottlenecks
Many mainframe batch issues are peak problems, not capacity limits. Learn how shifting COBOL workloads to zIIP reduces GP spikes and improves performance.
Improve IBM Z Workload Efficiency Without Recoding COBOL
JOPAZ recompiles COBOL into zIIP-eligible Java byte code, optimizing IBM Z mainframe efficiency and reducing CPU usage without changing code.
How JOPAZ Turns IT Infrastructure Optimization Into ROI
JOPAZ modernizes legacy COBOL applications, reclaiming mainframe capacity and maximizing IT investment ROI without rewriting code or changing business logic.