Skip to main content

IDAA v7.1 Enhances Performance, Scalability and More

IBM Db2 Analytics Accelerator (IDAA) v7.1 is now a Docker container with new database engine, new hardware and two deployment options.

Circle with right-facing arrows on the right, and gears on the left

IBM has created a new generation of IBM Db2 Analytics Accelerator (IDAA) v7.1. It’s now a Docker container with new database engine, new hardware and two deployment options.     

Architecture and Deployment Options 

Unlike IDAA v5.1 which was only available as a physical appliance built on Netezza connected to IBM Z through Open Systems Adaptor (OSA) cards, IDAA v7.1 allows for two deployment options. Both options provide the same functionality, same implementation and same API. This is possible with Docker technology. Below you’ll find the deployment options and details:

1. IBM Integrated Analytics Systems (IIAS)—physical appliance
  • Physical appliance connected via OSA cards with dedicated 10g network cable 
  • Built on Power hardware (P-series)
  • Available in one-third rack, two-thirds rack, full rack, double rack and quad-rack
  • Designed for massively parallel performance
  • Powered by Red Hat Linux
  • All flash storage
  • 4x threads—database operations can be run concurrently across multiple threads
  • 4x memory bandwidth
  • 4x more cache at lower latency handling larger chunks of data at higher speeds
  • Balanced and designed for Massive Parallel Processing (MPP)—optimized for performance 
See Figure 1, below, for different IIAS configurations.

Screen-Shot-2019-07-16-at-9-59-15-AM.png

2. IBM Db2 Analytics Accelerator for z/OS (zIDAA)—software appliance
Software appliance running on IBM Z on Integrated Facility for Linux (IFL)
  • IFLs and runs on a new LPAR type: Secure Service Container (SSC)
  • Software stack packaged into an IBM Secure Service Container
  • Fully self-managed appliance running in a SSC LPAR that can be deployed in minutes
  • Integrates seamlessly into the Z environment and leverages LPAR, memory and CPU management procedures, including call home support for enterprise hardware components
  • Uses customer-provided storage to hold the accelerator-side data 
More on SSC LPAR:
  • Not for general purpose use, only appliances can be deployed
  • Speeds up deployment with an integrated installer
  • Ensures integrity of downloaded appliance
  • OS included in the Docker runtime
  • Network and storage control including encryption
  • All management interactions with an appliance are via Web UI/REST APIs
  • System monitoring: admin/user controls; view messages, events; network management; user management; disk management; view appliance status; create/obtain dumps/logs; apply service, updates 
For zIDAA hardware considerations, see Figure 2.

Screen-Shot-2019-07-16-at-9-59-41-AM.png

The Docker Container  

IDAA v7.1 continues to deliver query acceleration, data warehousing and ETL capabilities as IDAA v5.1 did, but now it’s in a Docker container.  This allows for an automated, faster ease of deployment. 
 
The accelerator Docker container includes:
  • An accelerator server—establishes the connection with Db2 z/OS and manages accelerator tasks
  • A database engine
  • Components required for high availability, monitoring and security
  • A Linux OS  
Outside of the container:
  • A server that includes multi-core CPUs
  • A large memory
  • Shared filesystem to persist the data
  • Docker supported Linux OS used to launch the Docker

The Engine

IDAA v7.1 is all new technology powered by in-memory columnar database engine, Db2 with BLU Acceleration making it a software acceleration technology rather than hardware acceleration.
 
There were some challenges with the IDAA v5.1 PostgreSQL database engine that no longer exist with the new Db2 engine (i.e. datatype support, correlated subqueries, Db2 functions). Hence, the new database engine enables more Db2 capabilities such as:
  • Improved SQL compatibility
  • Improved throughput and performance
  • More Native SQL compatibility         
  • More data type support

Db2 Analytics Accelerator Loader for z/OS  

The Db2 Analytics Accelerator Loader is a separate licensed tool that enables the capabilities of loaded non-Db2 and non-Z data into IDAA. This allows access to the non-relational (i.e. VSAM, IMS, SMF, IDMS, etc.) and relational (i.e. Oracle, SQL server, etc.) data through SQL queries. This is possible through data virtualization.   
 
The Db2 Analytics Accelerator Loader extracts the data into memory, virtualizes and converts the data into a Db2 load format using zIIPs, and passes it to the batch loader to load into IDAA. This reduces elapsed time and CPU consumption, and simplifies processes while using zIIPs. 
 
The capabilities of the Db2 Analytics Accelerator Loader and IDAA are enormous, and can be used to:
  • Load and store SMF data for capacity planning, forecasting, trending, troubleshooting and dash boarding
  • Extract-Transfer and Load (ETL) data—run the ETL process on the accelerator and pass the data to the downstream systems
  • Load the accelerator from registered Db2 image copies (i.e. Cross System Load, Group Consistent Load)
  • Load multiple accelerators in parallel with the HA LOAD 
To see the Db2 Analytics Accelerator Loader Infrastructure, see Figure 3, below.

Figure3.png
Figure 3: Db2 Analytics Accelerator Loader Infrastructure

Unique Benefits 

In conclusion, Db2 Analytics Accelerator v7.1 provides two deployment options with one API and one implementation. Each deployment option provides benefits:

1. Physical appliance:
  • Integrated Power hardware and software
  • Out-of-the-box experience
  • POWER-grade Quality of Service
  • Scalable by scale-out
2. Software-only appliance:
  • Deployed on IBM Z hardware and storage infrastructure
  • Download-and-go experience
  • Scalability by using Z expansion of IFLs and memory
  • CA/DR support based on GDPS (active/passive or active/active configuration)
Regardless of the deployment, once adding an accelerator to a Db2 z/OS subsystem, it expands into a hybrid environment, enabling it to continue to process online transaction processing, expanding it into an analytical environment for faster reporting and data archival. This allows for the integration of transactional/analytical workloads of your mission critical applications.
 
However, IDAA isn’t just for query acceleration. It has tremendous potential and can be leveraged as a data server to retain system logs or SMF data or to enhance ETL processes while reducing MIP consumption, enhancing speed or complementing an existing data warehouse.
 
To learn more, visit the IDAA YouTube channel.

Delivering the latest technical information to your inbox.