5 Things To Know About Reducing Data Movement Costs in z Systems
MartinKeen 1200007VU3 Visits (10030)
Finding hidden trends in operational data can be highly valuable to any business. But one cost that is often overlooked is the cost of moving data from a system of record to an analytics environment. It is estimated that for every online transaction processing database, there are seven or more copies being maintained in other locations for analytics or other purposes.
Here are five things to know about how you can reduce data movement costs in z Systems environments:
1. OLTP systems and analytics systems often live on different platforms
The IBM z System platform is where more than 70% of the operational data generated by banks, retailers, and other large enterprises resides. This operational data is used in analytics environments to gain new business insights. However, these analytics systems often run on distributed platforms. Each day much of this operational data is moved to these separate analytics environments using an activity called extract, transform, and load (ETL).
2. Moving data costs money – a lot of money
A typical IBM mainframe customer moves multiple terabytes of OLTP data from z Systems to distributed servers every day. In terms of MIPS this is an expensive operation, often consuming 16-18% of a customer’s total MIPS. IBM conducted a study to quantify the cost of this intensive ETL activity, finding the cost of moving 1 TB of data each day resulted in movement costs of $10 million over a four-year period.
3. Co-locating analytics with OLTP addresses this problem
Running analytics workloads as near as possible to where the data already are can help you avoid these high data duplication costs. Using the IBM zEnterprise Analytics System 9700 solution allows your operational data to remain in the z Systems environment. Analytics run within the solution using IBM DB2 Analytics Accelerator and its Advanced Loader, which copies data from the operational system to the DB2 z/OS and Analytics Accelerator simultaneously.
4. The Analytics System 9700 solution's hybrid architecture is ideal for operational analytics and deep analytics
Operational analytics enable high-volume, short-running, highly concurrent, OLTP-like queries. These run best on z Systems mainframes. Deep analytics involve fewer but more complex and long-running queries and are used by power users such as data scientists and statisticians. This second type of analytics runs extremely fast in DB2 Analytics Accelerator. With the Analytics System 9700 solution's hybrid architecture, you can run analytics in the environment most suited to the task.
5. You can reduce your data management costs by around 88%
IBM conducted an additional study to measure the cost of copying 1 TB of operational data from DB2 11 for z/OS into the Analytics System 9700 solution. It showed a cost of $1.2 million to copy 1 TB of data daily from the operational system to the Analytics System 9700 solution, compared with more than $10.2 million to copy the same data in the earlier study. That’s a saving of 88%, or $9 million.
Find out more in the IBM Redbooks Point of View publication Redu