Blogs

Appliance-Based Analytics and Transactional Processing: Part 1

Integrate expertise for rapid deployment of efficient data management

Rapid business decisions are becoming a mainstay for operations in many organizations. Amid increasing complexities, ever-growing volumes of data, and time-is-of-the-essence scenarios, speed and efficiency in making those decisions are seen as requirements for maintaining a competitive advantage.

IBM® PureData® System appliances compose one of the system groups within the IBM PureSystems® family of pre-configured servers that includes the IBM PureFlex® platform for infrastructure-as-a-service (IaaS) implementations and the IBM PureApplication® system for platform-as-a-service (PaaS) solutions (see Figure 1).

 
Appliance-Based Analytics and Transactional Processing: Part 1

Figure 1. PureData System types in the PureSystems family

 
PureData System is designed to deliver data services for demanding applications simply, rapidly, and cost-effectively. It is available in different models that are optimized for a range of applications and data workloads. For many of today’s organizations, an overview of the PureData System portfolio and its integrated features and tools can be helpful when planning big data analytics initiatives.

A hierarchical architecture

PureData System offers a selective and hierarchical approach to execution within a virtual environment and infrastructure. It effectively adapts and builds in best practices and expertise to handle massive volumes of data with diverse variety and velocity. The PureData System architecture enables organizations to capture and automate the expertise necessary to implement infrastructure patterns and application patterns. PureData Systems are preassembled and pre-configured to minimize deployment time and expense. The three PureData System classifications include PureData System for Analytics, powered by IBM Netezza® technology; PureData System for Operational Analytics; and PureData System for Transactions.

Analytics

PureData System for Analytics is designed for data warehouse applications that load large volumes of data and run complex analytics and reports, such as sales analyses. It integrates database, server, storage, and advanced analytics capabilities into a single, easy-to-manage system. Netezza technology enables the appliance to be immediately operational, and PureData for Analytics offers a built-in analytical and scalable database infrastructure that contains an extensive library of statistical and mathematical functions for analytics tools and parallel programming languages. The Netezza technology also helps avoid the need for complex database management tasks and manual storage administration.

Operational analytics

PureData System for Operational Analytics provides a platform for applications that run complex analytics while handling operational reads and writes for real-time decision making, which is often required for fraud detection. This system offers many preset features to help optimize performance, throughput, and resource utilization of operational analytics workloads. These features include advanced data partitioning, parallel operational access, continuous data ingest, adaptive compression, and enhanced query joins and optimization. PureData System for Operational Analytics also provides expertise into pre-integrated deployment patterns, which helps speed the development and delivery of new services.

Transactions

PureData System for Transactions can be deployed for applications that handle millions of daily online transactions, such as ecommerce order management. It is a highly available, scalable database platform that helps reduce complexity and data management costs. PureData System for Transactions contains server, storage, network, and software resources that are optimized for online transaction processing (OLTP) workloads. The IBM DB2® pureScale® clustered database system is integrated to provide advanced technologies such as a caching facility, shared data, and a two-tier buffer pool.

Data management best practices are embedded into the system. They cover everything from data availability and disaster recovery to lifecycle management, database monitoring, data compression, and database tuning for optimal performance. In addition, the system provides two types of patterns to capture best practices into reusable, deployable packages that can accelerate deployment of data services. First, database patterns can be used to deploy cloned databases within the system, including a default, pre-tuned, and pre-configured OLTP database pattern. Second, topology patterns can be used to deploy database clusters of different sizes to best fit the application’s needs.

Along with the highly reliable uptime, scalability, and client transparency of the DB2 pureScale databases, the system includes a unified management control interface and a set of database tools for database development, administration, monitoring, and optimization that help keep the system running efficiently and at optimal capacity.

Workload optimization

Like the other appliances in the PureSystems family, PureData System for Transactions builds on decades of experience at IBM to provide an elastic, highly reliable database platform. The second installment in this series takes a deeper dive into the tools integrated in PureData System for Transactions.

Please share any thoughts or questions in the comments.

Resources