Rock-Solid Business Analytics
Business analytics on IBM System z streamlines well-informed decision making to gain a competitive edge
Many organizations realize that in today’s business environment, the key to remaining competitive is gaining access to ever-growing volumes of data that are expanding at a significant rate. There is increasing demand for gaining insight at every level of the business—and providing wide access to analytics information has become a top priority for chief information officers (CIOs) in multiple industries.
Business analytics is the heart of acquiring business insight and is now being recognized as a critical strategic asset in various enterprises. A number of industry studies have shown that organizations that invest in business analytics and use that insight for business decisions can achieve enhanced business performance and competitive advantage.
Business analytics software, solutions, and services provide everyone in an organization with the ability to spot and analyze trends, patterns, and anomalies. Armed with these tools, organizations can predict potential threats and opportunities, and change course when necessary to help improve outcomes.
Analytics brought to the data
A significant proportion of the data used for analytics originates on IBM® zEnterprise® System mainframe platforms. Running business analytics on zEnterprise System can build on the strengths of zEnterprise through easy consolidation, high availability, streamlined management, and simplified governance. In many organizations, the highly critical data can reside on this platform.
Deploying business analytics on zEnterprise System provides an end-to-end solution that accurately and securely collocates data warehousing, business analytics, and transactional data and significantly reduces data movement with reduced latency—almost in real time. IBM DB2® Analytics Accelerator (IDAA) software can dramatically enhance query and response time with rapid acceleration of complex queries. And real-time scoring brings in-database scoring functionality to online transaction processing (OLTP) applications.
Tremendously high availability performs at full capacity with prioritized critical queries and workloads. Integrated disaster recovery provides uninterrupted business analytics features for planned and unplanned outages. And a single, centralized server with scalable and flexible infrastructure offers very high resource utilization for mixed workloads.
Business analytics elements
IBM business analytics software on IBM System z® platforms helps organizations enhance their business outcomes. It requires several key elements, beginning with fundamental components that are required for a solid solution. These elements include the IBM DB2 for z/OS® database, an accelerator such as IDAA for complex queries, and information processing and interfacing requirements such as IBM Cognos® Business Analytics software, IBM SPSS® predictive analytics, and integration with existing systems.
The DB2 for z/OS database offers an important source of data for business analytics run on System z platforms. It is designed to provide high availability and high performance for mixed workloads. Many client organizations implement DB2 for z/OS to isolate ad hoc queries without impacting Customer Information Control System—IBM CICS® application servers—and IBM Information Management System—IBM IMS™ hierarchical database management system (DBMS)—workloads. DB2 for z/OS also provides access to both OLTP and data warehouse–based data in a single system.
IBM DB2 Analytics Accelerator powered by IBM Netezza® appliances provides cost-effective, high-speed deep analytics. When combined with DB2 for z/OS, it offers a single workload-optimized system for OLTP, data warehousing, and business intelligence (BI) along with application transparency and reduced, manageable data movement. Other key highlights of IDAA include the following:
- Dynamic routing for highly efficient, fit-for-purpose execution architecture
- A single, optimized environment for operations merged with data warehousing
- One environment for security, logging, backup, and recovery
- A single programming access point for business data
The latest tooling support makes possible directly loading IDAA with nonrelational data such as IMS, Virtual Storage Access Method (VSAM), and sequential data or outside relational data such as from an Oracle database. The tooling enables consolidation of all enterprise analytics data on IDAA no matter where this data is generated.
Cognos Business Analytics provides a unified workspace for analysis, reports, dashboards, ad hoc queries, real-time monitoring, and collaborative and mobile BI on a leading-edge platform hosted by zLinux.
Interactive visualizations can amplify mobile BI with advanced, innovations that enable line-of-business users to quickly pinpoint trends in data. New requirements for visualizations constantly emerge, which is challenging because each requirement is unique. Cognos on System z can provide a wide variety of visualizations with the means to extend and customize.
Line-of-business users can gain a comprehensive view of the business and act quickly off of insights with expanded support for big data sources, including the Apache Hadoop framework, analytic data stores, and real-time streaming data. It combines data in motion and data at rest. General Apache Hive support to access Hadoop, Cloudera, Hortonworks, and Amazon Web Services Elastic MapReduce (AWS EMR) is available, along with other distributions offering optimized access to the IBM InfoSphere® BigInsights™ platform through Big SQL and SAP HANA.
IBM SPSS® predictive analytics technology drives the widespread use of data in decision-making processes through statistics-based analysis of data and the deployment of predictive analytics in the decision-making process. Predictive analytics includes the following characteristics:
- BI technology that predicts what is likely to happen in the future by analyzing patterns in past data
- Delivery of predictions in the form of scores generated by a predictive model that has been trained on historical data
- A predictive model that has been trained on specific data for assigning these predictive scores
- Learning from the cumulative experience—that is, data—enabling enterprises to take action by applying what has been learned
CICS and IMS transactions call DB2 for z/OS scoring User-Defined Functions (UDFs) with a ported SPSS scoring algorithm on z/OS. DB2 scoring UDFs contain code ported from SPSS and callable by any SQL statement.
In addition, SPSS provides the Predictive Model Markup Language (PMML) XML-based file format. PMML offers several advantages in DB2 scoring, including enhanced speed and accuracy of scoring to drive accurate business results, easy incorporation of scoring into applications, and operation on current data—last committed data—and scoring in a single SQL statement.
Real-time analytics transactional scoring
Transactional analytics that occur in real time, such as credit card fraud detection, compute intensive neural network calculations that are required to be offloaded on alternative hardware. Overnight batch runs with latency costs of offloading can negate compute advantages. The optimized on-board floating-point architecture of System z enables re-hosting a real-time analytics transactional scoring application on z/OS while helping avoid network latency delays and adding value to OLTP transactions.
The sooner an act of fraud is detected, the more cost-effective a solution can be for organizations, which is made possible by the SPSS scoring infrastructure integrated into DB2 for /zOS UDFs, IBM ILOG® decision optimization integration of COBOL and Java business rules, and ILOG CPLEX® optimization on zOS.
Collaborative business analytics initiatives
By deploying these key components, enterprises can be ready to move forward with business analytics initiatives on System z (see figure). Keep in mind the following four considerations for running business analytics on System z:
- Organizations are using analytics to outperform their competition.
- Increasing numbers of line-of-business users across the organization want access to business-critical analytics.
- Business-critical analytics demand low latency and high qualities of service and performance. Infrastructure should be designed to be scalable, available, and reliable. Data governance and security must be effective, and analytics need to be timely and accurate.
- Analytics components spread across multiple departments can increase data latency, cost, complexity, and governance risk such as data security, integrity, and maintenance. Bringing analytics components where data originates helps improve data governance while minimizing these consequences.
Running comprehensive business analytics on System z
A business analytics initiative is neither strictly an IT project nor a business project; it requires a collaborative approach from both IT and business. Please share any thoughts or questions in the comments.
|[followbutton username='IBMdatamag' count='false' lang='en' theme='light']|