Saturday, February 6, 2016

1.2 Summary

For readers who do not have much time this chapter gives an overview of the most important thoughts of the blog.

Current challenges in grown system landscapes

  1. Today's default system architecture paradigm is client-server with dedicated servers for applications and databases.
  2. Historical limitations in computing power led to fragmented, complex system landscapes with artificially inflated volumes of duplicated and inconsistent data
  3. Data is exchanged over point-to-point interfaces or enterprise application integration-infrastructures (EAI) mostly in asynchronous batch runs.
  4. Operations and maintenance of these landscapes is extremely expensive and they are not prepared to support the business in the digital era.
The following picture illustrates how today's IT landscapes look like from a high-level technical perspective. The systems in the top row represent applications with their respective servers. "Blue" systems are dedicated reporting systems (read only). The systems in the bottom row represent database servers. I have not included mainframe systems in the picture which we nevertheless still find frequently. The arrows in the picture represent data exchange processes between the applications.


 Proposed solution

  1. Introduction of a new system architecture approach with a Corporate Data Pool based on HANA as single point-of-truth.
  2. Protection of investments into application logic while shutting down all isolated data silos.
  3. Step-by-step transition without interruption of business and with immediate benefits.
The illustration below shows how the example landscape from above will look like after a complete transition. The isolated database servers have been eliminated and now all data is stored in exactly one HANA database with no duplicates anymore. By eliminating the data silos we subsequently got rid of our interfaces and batch runs for data provisioning. We decommissioned the dedicated reporting- and in this example also one operational system and integrated their functionality in the remaining three operational systems with embedded analytics. We kept all required application logic with minimal adaptations.

Sounds easy - still not many enterprises are doing it yet. Why?! Because here I introduce HANA into an environment where it is not at all used today: running, productive non-SAP systems. This has amazing and unexpected effects which offer new options for users.

No comments:

Post a Comment