IBM Db2 with BLU acceleration is a group of technologies that are integrated with Db2 and is designed to increase the speed, efficiency, and adaptability of the software architecture. It will speed up the performance of analytical workloads by answering the queries twenty times faster.
Why the Speed?
The range of innovations included with column organized storage, actionable compression, data-skipping, CPU acceleration, and in-memory performance made the BLU acceleration to achieve the speed.
Column Organized Storage
The traditional way of storing Db2 data which is in row organized tables has been replaced by BLU acceleration where it enables the data to be stored in column-organized tables. However, the BLU maintains row relationships but to store the data it does by column.
With the column organized storage, the storage compression efficiency gets improved that leads to greater analytic performance and that lets you to achieve over 100 times faster processing of OLAP queries compared to row-based query.
Actionable Compression of Data
BLU acceleration enhances the existing Db2 data retrieval algorithms with order-preserving frequency compression and page-level compression efficiency that will have many values for the same column on one page. BLU Acceleration stores the number that is frequently retrieved at higher level than those numbers that appear less often. This helps you to compare the data in compressed state avoiding the heavy workload of decompression, reducing data on disk and in memory.
Optimizing the Performance with In-Memory Enhancement
BLU Acceleration comes with in-CPU memory optimization where every aspect of BLU Acceleration has been designed to minimize access to RAM and maximize processing time in L3 and L2 caches, which operates at faster rate compared to RAM.
BLU Acceleration uses more in-memory processing techniques compared to Db2 in order to optimize access to RAM as required. If the data size exceeds the RAM capacity it will automatically store the intermediate results to disk.
To achieve better CPU performance, the memory access is optimized as majority of access is via a CPU cache rather than RAM so that the CPU is kept busy minimizing the latency.
Data Skipping without Fine-Tuning
By skipping the data that do not qualify for a query enables BLU acceleration to save CPU, RAM, and I/O resource. The user intervention is not required to tune this as the BLU automatically records the metadata.
CPU Acceleration Through SMID
To process multiple points of data simultaneously and retrieve the results from multiple data elements from the same register, BLU Acceleration uses low level CPU instructions in the single instruction, multiple data processing (SMID) hardware feature.