Best Practice: HANA Migration

HANA has brought a new paradigm to enterprise-level computing. However, with robust capabilities, companies need accurate methods of benchmarking to ensure the highest ROI possible. Some vendors make grand claims, or some benchmarks fail to provide insight or context into the actual value.

In the past, limitations of a database created separate worlds of OLAP, OLTP, unstructured data management systems, and event-processing. The benchmarks mirrored the separate silos. Each silo had its own benchmark category. Now HANA overcomes these divides to provide a more useful, insightful, and accurate benchmark.

Benchmarks are best when they measure core features that are long-lived and typically quite complex. They must integrate innovation rapidly, in real-time, and bring value without disruption.

More accurately, a value in decision-making is derived through the capability to make those critical decisions within the correct time-frame. With Avantra and HANA, we believe benchmarks go beyond measuring speed; they must measure value as well. Thus, benchmarking at the migration stage sets your company up for the highest ROI possible, and you continue to see measurable, definable results.

You can look at five key areas that comprise benefits and costs. We seek to optimize each of these areas:

1. Going deep – unrestricted query complexity
2. Going wide – unrestricted data variety and volume
3. Within the window of opportunity – rapid response
4. In real-time – utilizes the most recent data
5. No pre-processing data – data preparation

Your company depends on industry-based benchmarks to analyze the viability of new software and hardware. This requires more than measuring the speed of various activities. Essentially, with Avantra, the benchmarks measure accurate real-world performance and reflect the true value that you can realistically expect from the deployment.

Your next step in the HANA migration process is to cleanse your data. This is a crucial activity to perform prior to moving your SAP systems into HANA. You will derive three primary benefits from cleansing the data:

• A smaller data footprint can also reduce the infrastructure, hardware, and licensing costs
• A decreased data size enables you to conduct the technical migration with a decreased business downtime
• By eliminating old, redundant, and non-relevant data, your SAP HANA will operate at higher levels of efficiency

Depending upon your internal processes and business needs, you must ensure an appropriate data cleansing method. You may archive aged data to lessen the size of the database.

This is also the time to review your data quality management process. In general, you should adopt six steps in maintaining quality data within your database. This will help ensure the smoothest move to HANA possible.

1. Identify your data assets to clarify where data storage exists within the organization, how it moves, and how users consume it. This will give you deeper insight into the systems and tables containing the data.
2. Define the following: data policies, standards and quality requirements. This will assist you with setting objectives for your data quality management strategy. This may also be an opportunity to define common business terminology to create a glossary to prevent misunderstandings and miscommunication.
3. Assess the data quality that currently exists based upon your defined parameters and validation methods you previously set. Use quantifiable analysis to apply to data quality for a streamlined data cleansing process.
4. Perform a thorough analysis to identify the root cause of bad data, and identify the negative influence bad data produces further downstream in your systems such as the data warehouse staging area. This analysis into bad or low quality data is crucial to understanding where it is stored and subsequently moved to the data warehouse through ETL.
5. Improve data quality by pre-defining the appropriate data cleansing strategy. After identifying the root cause of bad data, decide on when, where, and how to take action. This can include setting a data quality firewall for applications and systems to prevent the entry of new data that fails to meet requirements, which can be real-time entry of data by users, transactional data, batch loading of external data, or customer self-service.
6. Monitor data quality continually. Apply data validation rules on a regular basis to obtain historical trend analysis and identify changes in data quality over the long-term. This will assist you in quantifying the benefits of data cleansing by demonstrating quality data scores statistically. Be prepared to be proactive with any slight dip in data quality scores to ensure continued satisfaction from employees, suppliers, and customers.