IBM z16 Day 2: Predict and Automate; Compliance and Security

Introduction

IBM has just announced its new IBM Zsystems server: the z16.  This new server is based on a new microprocessor (the Telum processor) – a processor that continues IBM strengths in secure database and transaction procession – but now extends its processing capabilities into the world of Artificial Intelligence (AI).  With Telum, IBM z16 can now provide on-chip inferencing capability (see this Clabby Analytics report for more details on what this architecture can do – and the types of workloads it will enable IBM zSystems users to process).  And, with on-chip inferencing (the heart of AI), IBM will be able to open a whole new world of applications and services for its customers.

The Marketing Message

IBM believes that its new z16 represents the future of computing – as its new marketing message now reflects: “built to build the future of business.”  This future will make systems and applications more intelligent (using AI as the catalyst to do so).  To support its marketing message that IBM z16 represents the future of business, AI has not only been woven-into IBM’s z16 system hardware – but also into the company’s infrastructure product offerings, particularly in security, resilience, and operations.  Further, IBM has broadened its infrastructure offerings on IBM zSystems by helping its independent software vendors (ISV) partners to AI-enable their infrastructure offerings on IBM zSystems.

Additionally, IBM positions its new server as ideal for extracting real-time business insights – as IBM puts it, the new architecture “infuses AI in real-time into every business transaction”.  IBM products that support this claim include the company’s new Db2 13 for z/OS product with its AI Powered SQL for Data Insights. feature.  Other examples include Watson Machine Learning for z/OS; IBM Cloud Pak for Data; AIOps (to assess, diagnose and resolve incidents); and the venerable Db2 Analytics accelerator for z/OS.  Third -party products that support AI and Open Source AI Frameworks that are optimized (to take advantage  of the IBM zSystems SIMD support and the z16 Integrated Accelerator for AI on the Telum chip) and) for IBM zSystems include ONNX, Spark, XGBoost, Anaconda, TensorFlow, Keras, PYTORCH, and many others.

What this all means is that, with z16, IBM has an integrated AI architecture, with predictive infrastructure that can aid in a variety of systems functions (security, resiliency, management), while serving (in a highly efficient manner) a new generation of AI-based applications.

What’s Going-on with AI Agility on IBM z16?

AI agility, in this sense, refers to how quickly and easily AI “infusion” can take place.  To AI-enable a system, infrastructure component, or application/database environment, it is necessary to:

  1. Import Data (cleansed, non-redundant data);
  2. Model and prepare that data for inferencing;
  3. Train that model;
  4. Deploy the trained model; and then
  5. Predictive behaviors can take place.

This process is being used to AI-infuse a wide range of AI infrastructure products; it is being used with new applications; and it can be used with existing IBM zSystems applications.  For instance, consider existing COBOL applications.  CICS COBOL applications can invoke ONNX modes using standard CICS commands (ONNX is an open format built to represent machine learning models).  ONNX along with IBM zSystems’ Deep Learning Compiler provides features for optimal exploitation of the new Telum Integrated Accelerator for AI.  IBM’s WMLz can then be used to manage the deployments of the new model scoring service – thus enabling both new applications as well as COBOL programs to quickly and easily become AI enabled.

Yes, it’s nice that zSystems infrastructure products have been AI enabled; and yes, it’s nice that new applications as well as existing COBOL applications can also be AI-enabled – but the real gem in IBM’s database portfolio is AI-enhanced SQL.

With the SQL Data Insights feature of IBM Db2 13 for z/OS, AI-enhanced SQL can be used to mine some of the enterprise’s most valuable data – its “system of record” and “system of engagement” data.  By mining this data, enterprises can find and monetize hidden insights; identify similarities, dissimilarities, and correlations; use a single model for query to answer multiple questions all WITHOUT data science skills.  With the industry shortage of data scientists and analysts, this represents a simple and straightforward way to glean important insights from existing mainframe data.

What’s Going-on with Resilience?

When we last covered a major IBM zSystems announcement (the z15), we described the actions that IBM was taking to automate capacity shifts to ensure operations through disaster – or to support recovery testing.  IBM’s z16 now fully automates shifts in capacity between IBM z16 systems at different sites – allowing production capacity to be moved dynamically almost imperceptibly (in seconds).  New to zSystems resiliency is an IBM capacity service called IBM Flexible Capacity for Cyber Resiliency.  This offering enables clients, under their own control, to proactively reduce the impact of downtime by shifting workloads to an alternate site to ensure business continuity.  It allows for proactive outage avoidance while ensuring business continuity compliance (in disaster recovery and disaster recovery test scenarios).  IBM allows clients to take advantage of this offering up to 12 times per year and for the capacity to remain at the alternate site for up to one year.

When we last described IBM’s System Recovery Boost – we described how it allowed for faster shutdown and startup; faster recover on Sysplex; how it allowed for faster GDPS automation; and faster elimination of backlog after planned or unplanned downtime.  This turned out to be a wildly successful IBM offering – with over 95% of IBM eligible systems running System Recovery Boost to recover quickly – and to take advantage of additional processing capacity.  New in System Recovery Boost is 1) faster middleware restart (35% faster); 2) faster SVC core dump recovery (30% faster); and faster Hyperswap configuration loading.

What’s Going-on in Compliance?

One of the most interesting new developments in compliance is IBM’s new IBM Z Security and Compliance Center – an application environment chartered with helping ensure that clients can achieve continuous compliance readiness with over 300 pre-built goal validations.  Evidence of compliance is automatically collected and validated against these goals.  This offering helps clients optimize resources; assess their compliance posture; and identify compliance drift (moves that take an enterprise off its compliance path).

The most exciting aspect of this announcement is that it can help reduce the number of resources required to prepare audits by over 40% — reducing audit preparation time from one month to one week!

Summary Observations

During the first day of analyst briefings, IBM announced its new AI-enabled z16 system; explained how IBMz16 fits into the company’s overall strategy; and IBM also described it approach to ensuring its systems are quantum-safe (see this report).

In this second day, IBM further explained its zSystems strategy — and IBM also described the activities in agility, resiliency and compliance that it had undertaken since its last major IBM zSystems announcement.  The key take-aways are:

  • IBM has not only announced a new, powerful AI-enabled hardware platform – but also related infrastructure and ecosystem products.
  • The company has a clear roadmap for helping its customers exploit is new platform, using AI to automate functions, modernize older programs, and deploy new AI-enabled programs. Additionally, IBM has made it possible to do analysis without having to be a data scientist)..
  • The company has made huge strides in helping IT managers and administrators deal with the complexities and time requirements of compliance audits. And,
  • The company’s resilience products have improved – as has its position on making use of recovery services – and of testing recovery scenarios (IBM’s new capacity service).

For decades one of the biggest challenges for IBM zSystems has been to capture new workloads.  And over those decades zSystems have captured much of the world’s transaction business; made huge strides in the capture of business applications; repositioned itself to serve cloud markets; captured new cloud applications; changed the system personality to perform business analytics more efficiently – and has now, positioned IBM z16 to capture next generation AI applications.  No wonder this architecture – which critics suggested would be replaced by distributed servers in the 1990s – has continued to survive and prosper, constantly staying at least one-step ahead of the competition.