IBM Db2 for z/OS and its ecosystem

Digital transformation offers exceptional opportunities for enterprise organizations to become more insights-driven, build stronger relationships with customers and achieve competitive differentiation. In pursuit of digital transformation, many organizations are looking to the hybrid cloud with the promise of greater agility, abundant capacity, and little-to-no up-front investment.

But with hybrid cloud, organizations face increasingly complex data management challenges such as rapid data growth, rising operational costs, and the need to infuse AI into operations, applications, and transactions. All this with growing regulatory requirements and the need to reduce security exposures wherever and however the data is stored.

As organizations embrace hybrid cloud, they are also recognizing the value of a data fabric. This data management architecture approach can help optimize access to data and intelligently curate and orchestrate it for self-service delivery to data consumers. A data fabric can help provide faster, trusted AI outcomes by connecting the right data, at the right time, to the right people, from anywhere it’s needed.

96% of mainframe surveyed organizations are considering a data fabric architectural approach to address data integration, yet...

67% find processing some of their information in the cloud can delay availability¹

The concept of gravityis a key tenet of a data fabric, helping to bring compute to where the data originates or is stored. This approach helps ensure that data is readily accessible and current for applications, analytics, AI, and business process automation. Leveraging data gravity as part of a data fabric can help organizations minimize cost and infrastructure complexity, improve security and data governance, and ultimately help drive better decisions based on real-time insights.

Many of the core systems that you rely on for your existing transactional environment have been built and highly customized over decades to accommodate your differentiating business processes. They have been optimized to reduce cost and execute transactions within very tight service level agreements (SLAs), at times within a few milliseconds or even micro-seconds. Those transactional systems are being adapted to support hybrid cloud approaches in support of digital transformation efforts.

According to a Deloitte study on hybrid cloud and mainframes, the most compelling applications combine aspects of hybrid cloud and data from systems of record. Deloitte believes that it’s not a question of whether to invest in hybrid cloud, it’s what’s the right balance of a hybrid cloud and mainframe investment to achieve the greatest business outcome.²

Organizations are not willing to discard their differentiating core operational systems, so their on-premises mainframe investments are staying largely in place. They need a clear path to modernize their mission-critical systems for intelligent, efficient data serving and AI.

Many organizations are modernizing and building new applications off platform, on the cloud, while their systems of record remain on IBM® Db2®for z/OS®. That’s because IBM Z and Db2 for z/OS are designed to evolve with these changes and support modernization efforts with agility and security.

As part of a hybrid cloud architecture, IBM Z®, Db2 for z/OS and its ecosystem can add great value to an organization’s data fabric. It offers resilient, efficient, no-compromise, agile enterprise data serving for the world’s most demanding hybrid cloud and transactional and analytics applications. Enterprise organizations can access IBM Z data in place as part of their data fabric and infuse AI throughout transactions, applications and operations – at speed and scale.

IBM Db2 13 for z/OS enhancements deliver leading edge innovations:

• Readily integrate AI insights within Db2 applications
• AI embedded in operations and applications
• Enhanced resiliency, efficiency and application stability
• Simplified database management and upgrades
• Enhanced scalability, security and compliance
• Synergy with IBM z16

IBM is embracing autonomics to reduce human intervention and lower cost through efficiency and ensure the highest achievable availability. IBM Db2 AI for z/OS represents steps towards autonomous database management – simplifying database management efforts and enhancing system performance by embedding AI into Db2 for z/OS.

Db2 AI for z/OS leverages machine learning to help improve operational performance and maintain the health of Db2 for z/OS. It simplifies database management efforts through built-in domain expertise to help deliver optimal Db2 for z/OS performance, reliability, efficiency and cost effectiveness – even under the most demanding circumstances.

Underlying all of this is the assumption that you are leveraging IBM Db2 utilities and tools that improve operational efficiency, optimize actionable insights and intelligent automation, and provide greater business readiness and protection. But IBM Db2 Tools also help you modernize legacy Db2 processes and applications, reducing the mainframe developer learning curve, using familiar Visual Studio Code extensions specific to Db2 for z/OS. The Db2 IT professional can integrate Db2 data into an automated, hybrid CI/CD pipeline, while ensuring organizational standards and business rules are maintained and protected. These modern processes and tooling, using Db2 data, help deliver applications to market faster, with less cost and lower risk.

The IBM Db2 for z/OS database will continue to be the cornerstone that drives your digital transformation and accelerates your business initiatives for both public and private cloud. Db2 for z/OS delivers the high availability, real-time analytics, regulatory compliance and trusted security needed for your enterprise data requirements.

Data relationships and insights are often un-accessed, un-identified and therefore unmonetized. Traditional machine learning approaches do little to draw out this valuable information. To get more value from your existing data, you need a simpler, database-integrated approach for developing and deploying AI insights — with minimal data science dependency. With IBM Db2 SQL Data Insights, organizations can use SQL to embed self-service AI insights within Db2 for z/OS applications. A new feature, Db2 SQL Data Insights, makes this possible by exploiting the IBM Z Integrated Accelerator for AI in the IBM z16 and IBM zIIP processors.

This database-integrated approach extends the traditional relational model beyond exact value matching to incorporate similarities, dissimilarities, clustering, and correlations. For example, similar records do not need to have exact value matches between columns to uncover similarities. The approach eliminates the extract, transform and model building process and offers direct AI integration within a SQL accessible environment. Any Db2 for z/OS application that executes SQL can incorporate these new AI-related built-in capabilities.

That means that organizations can apply AI in new ways. For example, it can be leveraged to identify customers that have similar (but not exact) characteristics for better targeted marketing or to uncover new fraud activity based on similar known fraudulent behavior.

Key to the Db2 for z/OS ecosystem is the IBM Db2 Analytics Acceleratorfor z/OS. Coupling Db2 for z/OS with the Accelerator enables an industry-unique augmented transactional environment that allows you to leverage the high-value, time-sensitive data that originates on IBM Z in real-time without moving it.

Read the article: Is The Cloud Slower For Analytical Insights? -Tom Davenport, Ph. D.

The Accelerator is deployed on IFLs or on an IBM LinuxONE stand-alone system. This means you can integrate your mission critical transactional applications with business critical, time-sensitive, resource intensive processing, such as analytics, in a single environment, without compromise. And this efficient and cost-effective environment benefits from superior IBM Z qualities of service.

Traditionally, there has been a separation between resource-intensive operational and resource-intensive query processing platforms that provide insight. This has often led to significant latency between when data is generated and when insight is available. Insights lose their value the further in time they are away from when they are generated.

The IBM Z platform can process real time analytics on live transactional data without impacting transactional workloads. This is a key differentiator of the IBM Db2 for z/OS approach. It assures industry-leading performance for mixed workloads. Resource-intensive processing, such as analytics, does not degrade transactional workload performance because transactional and resource-intensive queries are each processed within separate compute resource pools.

The Accelerator enables an innovative approach to ensure that organizations can reduce costs and meet SLAs in batch and online–no matter what the requirement. Organizations use the Accelerator for analytics, batch acceleration and even within the context of resource-intensive, online transactional applications.

Read the case study: How can you boost the speed and efficiency of mainframe analytics processing?

As organizations deploy new applications to the cloud, the Accelerator can provide ready access to Db2 for z/OS data. For applications that require significant data aggregation and transformation, the Accelerator acts as a high-performance data transformation engine to satisfy these new data requirements.

When data originates outside of Db2 for z/OS, for example in IMS, VSAM or non-IBM Z data sources, the IBM Analytics Accelerator Loaderfor z/OS can provide data to the Accelerator to support queries that combine data originating within and outside of Db2.

Along the same lines is an exciting tech preview leveraging replication from IMS, VSAM and other data services. Using these data sources with change data capture technology incorporates real-time changes into Accelerator-Only tables. This means you can quickly join DB2 for z/OS and data originating from other sources in a single high-performance query.

You can leverage Db2 for z/OS for its transactional capabilities and drive analytics through your data fabric from your operational data without having to move the data off-platform. That means less complexity, less latency, and better security, governance and decisions.

To gain agility, your organization needs an approach that accelerates application development and deployment while working within existing hybrid cloud infrastructure investments. Recent hybrid cloud innovations support digital transformation while taking advantage of mainframe infrastructure, applications and data to support key initiatives such as compelling new end user applications, analytics/AI and data fabric – without impact to existing workloads.

Through a data fabric, organizations can access governed data across hybrid cloud resources to help deliver trusted, business-ready data and insight. This means lower costs, greater reusability and faster time-to-insight for better decisions and better results. Data fabric does not make existing data warehouses and data lakes superfluous; it leverages them wherever they might be – on-premises or in the cloud.

IBM Cloud Pak for Data makes the concept of a data fabric possible. It is available via software as-a-service on public and private clouds as well as on x86 and the IBM Z platform. IBM Cloud Pak for Data is an enterprise insights platform that simplifies and automates data collection, organization and analysis of data and accelerates the infusion of AI throughout a business.

With its capabilities to connect data everywhere, run workloads anywhere and to build, deploy and manage AI at scale in hybrid cloud environments, IBM Cloud Pak for Data is the enabler for business digital transformation.

Build a trusted, integrated data foundation with data fabric

Complementing IBM Cloud Pak Data is IBM Db2 for z/OS Data Gate (Db2 Data Gate). It can provide a synchronized copy of Db2 for z/OS data hosted within any supported IBM CloudPak for Data environment. It offers a modern hybrid cloud approach for IBM Z customers to deliver data originating in Db2 for z/OS. Organizations can readily support the increasing demand for analytics and the dramatic increase in modern high-volume read-only applications.

IBM Db2 Data Gate can help you reduce the cost and complexity of hybrid cloud data delivery. Instead of building and maintaining costly custom code, IBM Db2 Data Gate does the work.

Evaluate the benefits of using IBM Db2 for z/OS Data Gate technology versus a do-it-yourself data extraction and synchronization approach.

Not all your IBM Z data originates in Db2 for z/OS. Regardless, when you are developing applications that require data from traditional non-relational IBM Z data sources,you can still access that data where it originates, without specialized skills, and reduce the time and resources used to combine data from multiple systems. The time and resources you save can translate into cost savings. This functionality is delivered by IBM Data Virtualization Manager for z/OS.

You can cost effectively provide current system-of-record data to modern cloud-based applications and a data fabric without the complexities of an “assemble-it-yourself” infrastructure–and without moving data off-platform. You can simplify the development of modern applications from traditional non-relational IBM Z data sources such as VSAM and IMS, facilitating access to business-critical data that is readily consumable.

The IBM Z platform and data can play an essential role in today’s hybrid cloud and data fabric architectures as both a primary source of valuable enterprise data and a high performance, resilient platform to deliver actionable, real-time data and insight where and when it is needed.

As you implement traditional supervised learning approaches to AI, you need to consider the best approach: on-premises, in the cloud or a blend of both. You must decide where to build machine learning models and where to deploy them.

A data fabric infrastructure is often used to develop machine learning models. IBM Watson® Machine Learning for z/OS® gives you the flexibility to train your models in many frameworks, and on any platform including IBM Cloud Pak for Data, then readily deploy those models on IBM Z to gain the unique advantage of inferencing in-transaction, in real time. Through low latency, high performance inferencing you can readily embed models in mainframe applications and score every transaction without significant impact to SLAs or cost.

Watson Machine Learning for z/OS also supports inferencing with deep learning models. Deep learning models converted to the ONNX format can be imported into Watson Machine Learning for z/OS. The IBM Deep Learning Compiler generates an executable program for model execution to be readily integrated within IBM Z applications. Scoring processes are also often zIIP eligible, allowing for greater cost effectiveness and efficient execution with minimal impact to operational processes. And with the latest IBM Z infrastructure, AI software can take advantage of the IBM z16 Integrated Accelerator for Artificial Intelligence that both accelerates performance and reduces the cost of deploying deep learning model inferencing on IBM Z.

So, when can inferencing in-transaction, in real-time have an impact? Examples can include customer cross sell and upsell, customer attrition, risk mitigation and fraud detection. By scoring every transaction as it is taking place, more opportunities can be uncovered and quickly acted on to help increase revenue, decrease cost and addressfraud in real-time.

IBM now offers the IBM Watson Machine Learning for z/OSOnline Scoring Community Edition. This lightweight version of the Watson Machine Learning for z/OS scoring service provides a no-charge option enabling you to easily download and try the in-transaction scoring approach for your deep learning models.

According to a recent article from Tom Davenport, PhD, “There is also going to be more appetite for timeliness of analytical insights, particularly during business crises but during normal times as well. Hybrid cloud modernization and a data fabric approach can offer exceptional opportunities to become more insights-driven, build stronger relationships with customers, and achieve competitive differentiation. But excessive data movement can impede both data usage and time-to-insight and can undermine some of the potential value of some modernization efforts.”³