By Sowmya Kameswaran and Jens Müller
As we all know, Db2 for z/OS has one of the largest footprints in the enterprise database world. Your organization may have all or most of its' business critical data on Db2 for z/OS (you are not alone and we wouldn't recommend you to change a thing about that!). We realize though that many organizations are experimenting with or, rearchitecting and extending their infrastructure to take advantage of hybrid cloud.
With the changing environment and importance of hybrid cloud, cloud-based applications need easy, secure access to this data for modern applications. IBM Db2 Data Gate for z/OS makes data from Db2 for z/OS readily accessible on the IBM Cloud Pak for Data platform for business users and application developers.
We have embarked on an exciting era of compelling modern application development. There is a surge in both reporting applications requiring read-only access to transactional data as well as data-intensive analytics applications requiring access to historical data. Organizations are developing new compelling applications for differentiating services delivered to their customers. Since much of this data originates in Db2 for z/OS, many organizations have built custom ETL (extract, transform, load) jobs to extract and load this data into other databases to support their application needs. While this approach may work in the short term, some of the problems with this approach are:
- Expensive to create and maintain over the course of time (due to complexity, and the costs of synchronizing source and target databases, and ensuring transactional consistency if necessary)
- Data security concerns once data is moved from where it originates
- Increased operational processes and cost on IBM Z
Why Db2 Data Gate
- It is an integrated solution to securely access data from Db2 for z/OS on the cloud without the need for direct access to Db2 for z/OS.
- Avoids significant investment in building and maintaining custom ETL solutions to move Db2 for z/OS data.
- Provides better data currency via the Integrated Synchronization feature that replicates data from Db2 for z/OS to IBM Cloud Pak for Data.
- Significantly reduces (operation) cost of data replication on mainframe since 96 % of the underlying data synchronization technology is zIIP eligible.
- Enables modernization and transformation in your enterprise's Journey to Cloud.
- With Integrated Synchronization, the data availability for applications accessing data from the source is not affected (source tables are fully online for reading and writing) while data is replicated to the target.
- HA/DR is built directly into IBM Cloud Pak for Data
Db2 Data Gate is based on Db2 (row store) or Db2 Warehouse (column store) as the target databases within IBM Cloud Pak for Data. This model makes it suitable for supporting applications that require row level access as well as analytical applications that benefit from column based data store. Only one Db2 for z/OS database can be used as the data source. The key aspect of the architecture is the Integrated Synchronization feature which is optimized to replicate data from Db2 for z/OS to Db2 running under IBM Cloud Pak for Data.
With Db2 Data Gate and the target database running on the IBM Cloud Pak for Data platform, the solution works wherever the platform is able to run – private, public or hybrid cloud implementations, thus making relevant data readily available to application developers and business users where they need it.
Lab performance benchmarks
Db2 Data Gate boasts unrivaled performance when compared to any other data synchronization tool synchronizing data from Db2 for z/OS to Db2 (Warehouse).
With IBM Cloud Pak for Data and Db2 Data Gate installed on Linux on IBM Z, using Db2 Warehouse as the target database using hostPath data storage:
- Peak load performance (for making initial copy): 2.1 TB/h
- Peak synchronization performance: 200k rows/sec at 1.2 secs peak latency
Db2 Data Gate does not:
- Replace traditional transactional systems since it only provides read-only data access to source data
- Serve as a replacement for other data replication technologies because it supports only one source and one target and also does not support bidirectional replication or data transformation
- Support data versioning and hence is not an operational data store
- Guarantee data currency for high-volume transactional workloads. It is not a carbon copy of source data. Meaning, applications requiring absolute currency should access data at its point of origin.
User Interface snapshots
Db2 Data Gate provisioning
Select the target database type and deployment, resource allocation and network routing to proceed with creation of the Db2 Data Gate instance.
Setting up source
Once the instance is created, the first step is to point the Db2 Data Gate instance to the Db2 for z/OS subsystem to be used as data source.
Select and add tables from source to target
The next step is to select the tables to synchronize data from source to target.
Db2 Data Gate dashboard
Overview of status and activities associated with the provisioned Db2 Data Gate instance.
Db2 Data Gate in action
The video below demonstrates the Db2 Data Gate end user experience. One of the key aspects to note is that when the source tables are added, loaded and setup for synchronization with Db2 Data Gate, there is no impact to concurrent workloads executing on the source tables. The source tables are fully online for reading and writing while Db2 Data Gate makes the copy and starts synchronizing.
- Video: https://ibm.box.com/shared/static/uglbdy6xpwybon8ny54f03j1vb724116.m4v
Our next blog ...
Next time we will look at Db2 for z/OS Data Gate and Watson Knowledge Catalog integration.