Real Time Replication for SAP HANA
As I mentioned in my announcement this week, in 15.7.1 SP100 release, the most prominent feature is real time replication support for HANA DB as a replication database. SAP Sybase Replication Server has been known in marketplace for its log-based replication technique and solid performance for the past 23 years. We have extended these well-known characteristics to support in-bound HANA replication from following source databases for custom or third party applications:
a. SAP Sybase ASE
c. Microsoft SQL Server
d. DB2 UDB on Linux, UNIX and Windows
In addition, SAP Sybase Replication Server also supports heterogeneous direct load materialization from above databases to HANA DB. Materialization, a.k.a. initial load, is a process of creating and activating subscriptions and copying data from the primary database to HANA DB. Before you can replicate transactions continuously from a primary database, you must set up and populate replicate database so that it is in a state consistent with that of primary database. Direct Load is optimized for many tables with many rows by massive parallelism. Multiple parallel threads can be configured to load source data from one primary table to its corresponding replicate table. Direct load materialization is enabled by the direct_load option of the create subscription command.
Figure 1 SAP Sybase Replication Server for HANA
To provide high performance real time replication into HANA, Replication Server leverages native HANA ODBC driver’s bulk capabilities (such as bulk inserts and deletes etc.) by using ExpressConnect for HANA DB (ECH) technology. ECH is a library can be loaded by Replication Server dynamically. It doesn’t require a separate process for starting up, monitoring or administering, unlikely traditional Enterprise Connect Data Access (ECDA) based technology.
One of unique features of Replication Server is preserving transactional integrity and serialization between source and target databases. Transaction serialization refers to transactions at target database are committed in the same order as the primary database. Transaction integrity refers to replicated transaction is applied as single transaction at target database. This is the foundation to deliver trusted enterprise applications. Replication Server extended this feature to HANA replication. Therefore, mission critical transactional replication for HANA can be relied on SAP Sybase Replication Server.
To ensure consistency of distributed data, Replication Server also provides Data Assurance utility to compare row data and schema between two or more databases, report and rectify discrepancies. HANA is one of the latest databases supported in this release, in addition to Oracle, SAP Sybase ASE, and SAP Sybase IQ. Data Assurance is a scalable, high-volume, and configurable data comparison utility to run comparison jobs even during replication by using a “wait and retry” strategy that eliminates any downtime. In case of any machine or user errors to cause data discrepancies between source and target databases, there is no need to re-initialize, which is a time consuming process. Data Assurance can be to identify and correct any discrepancies without interrupting production system and inducing downtime.
Figure 2 SAP Sybase Replication Server – Data Assurance
In many HANA deployment scenarios, multiple database sources will need to be replicated into multiple instances HANA database. It is essentially an M to N replication topology, either configured for round-robin load balancing or better simulation in development and quality assurance environment with production quality data. With Replication Server, flexible M to N replication deployment topology without any geo-graphical boundaries is supported natively by Replication Server as illustrated in figure 1.
SAP Sybase Replication Server, a versatile transactional database replication tool, engineered for demanding and complex deployment environment brings real time data loading capability to SAP HANA, with high performance and low impact.