August 9, 2016 | Written by: Namik Hrle
Categorized: Mainframes | Real-time analytics
Share this post:
Whether you use data analytics for fraud protection, for delivering better customer experiences, for digital marketing, for the Internet of Things, for operational data lakes or more — your business gains competitive advantage by having the right answer, at the right time, at the right place.
A common practice in the past was to copy data to some other platform where the analytics then took place. But this is no longer efficient for modern applications and solutions. This practice not only generates many versions of the data and associated versions of the truth, it also results in latency between data creation and data consumption. This is simply no longer viable in a world where real-time analytics provides a major competitive advantage.
In order to drive immediately actionable insights from business-critical operational data, the IBM z Systems teams developed a new workload-optimized subsystem that combine the best of both worlds: superior transactional processing and real-time analytics.
Now you can design for a cognitive business. Now your self-learning systems can bring you new insight and new potential.
A mighty hybrid data processing engine
Transactional and analytical processing patterns are quite different. IBM z Systems use unique hybrid technology to combine separate engines, each dedicated to different kind of workloads. IBM DB2 for z/OS performs transactional processing and IBM DB2 Analytics Accelerator for z/OS performs analytical processing. The choice of the engine is automatic, based on an optimizer assessment about the type of processing required. Users do not need to be aware of the hybrid characteristics of the system because they only connect and communicate with DB2.
I like to compare this experience to driving a hybrid car: the driver has only one set of controls (steering wheel, accelerator, brakes) but there are actually two engines under the hood. The engines engage automatically and transparently.
All other solutions rely on a single engine to execute both transactional and analytical processing patterns, and there is simply no single engine that can match the best characteristics of a set of the dedicated specialized engines. A single-engine solution has a major scalability problem once the transaction rates and the number of concurrent analytical queries reach enterprise proportions. The hybrid orientation provides unique heterogeneous scale-out, where each of the workloads scale independently of the other.
Data visualization and correlation patterns
Since the hybrid orientation eliminates the movement of data off of the z Systems platform, this enables the delivery of timely information from the most accurate and trusted data available.
It allows an enterprise to deliver instant analytics and insight because a user can read the data in real time. From risk and fraud protection to delivering better customer experiences, the hybrid orientation provides valuable information immediately and helps organizations visualize patterns and make correlations within the contexts of location, transaction, time and space.
Many enterprise customers have built large, mission-critical applications on and around z Systems. These are operational applications that automate core processes and generate the most relevant business data to capture the essence of the business itself. This data is a gold mine for analytics. IBM z Systems provides the most reliable, available and secure IT infrastructure and data processing technologies available for the simultaneous hosting of business applications and a very large number of concurrent users.
Is your IT infrastructure ready for the cognitive era? Take our assessment here.
To learn more about the promise and challenge of cognitive across your systems visit us here.