Change is good, and often necessary. Many IT organizations recognize that they need to adapt and become more responsive and agile for their business in what we call the New Style of IT.
But not every IT organization needs to—or should have to—change in the same way all at once. While implementing Database as a Service (DBaaS) enables agile interactions that help transform IT organizations into service providers, it also allows them the flexibility to adopt the right level of automation given their operational circumstances.
I think of this in terms of a spectrum of different service patterns: from the traditional, totally human-based process of calling a Database Administrator, to the completely automated service pattern that empowers a LOB to patch its own database. Let’s take a look at how a range of transformational service patterns can be enabled by DBaaS.
Different service patterns for different scenarios
Figure 1 represents an Enterprise DBaaS implementation. In its steady-state, DBaaS will have a Service Catalog filled with standard database configurations that can be ordered by Application Teams or IT Administrators. As a simple example that demonstrates the value, consider that there are these three offerings in the catalog:
Oracle OLTP for Dev
Oracle OLTP for UAT
Oracle OLTP for Production
Fig 1: Enterprise DBaaS personas and components
When fulfilled, each of these offerings will execute detailed OS and RDBMS provisioning automation—in this case, to provision Oracle for an online transaction processing application. Through the simple utilization of REST APIs, DBaaS provides access to these standards from the right tool for the right environment. A few examples:
Developers — They may interact with a Self-Service portal to order the first offering above. The subscription they now own is for a simple “standard” OTLP database to which they have direct access.
Testing —On the other hand, for testing purposes it might be desirable to obtain a subscription for the second offering programmatically as part of the UAT’s automated test runs. Subscriptions would come and go much more quickly in these environments.
Production — More stringent production processes are often captured in runbooks. If autonomic, the runbook would utilize similar programmatic methods to provision as UAT, but in a more human-based form, would just trigger the Application Team to obtain a subscription to number three via the portal.
Patching production database estates
Once DBaaS has provisioned the database, it naturally must be maintained. In previous blogs I’ve talked about DBaaS patching and the concept of the Database Estate, but let’s revisit it from the perspective of transforming IT service patterns. When the DBaaS automation services provisioned the RDBMS, they established its presence in the DB Estate. The DBaaS Operator can now view the entire DB Estate and remediate it when necessary. A key example of remediation is the need to patch production systems to the latest vendor level. For IT shops that struggle to keep anything patched, DBaaS operator patching facilities are a pretty nice first step.
However, it is not really transformational because the DBaaS Operator is still a central cog in the organization’s change control processes. The pattern for transformation is then the same as in the case of the provisioning services: create a standard set of offerings that can be accessed during a more automated change control process. The trick here is that DBaaS patches the system based on the subscription regardless of how the subscription was obtained. The DBaaS patching service can therefore be accessed programmatically from a runbook or via the application team pushing a “Patch It” button. Extending this even further, DBaaS can help the IT organization consolidate operations by offering services to patch the entire subscription from OS thru RDBMS.
As you know, to truly transform IT Enterprise DBaaS will need to offer similar services for database code release and test data management. Hopefully this exploration of DBaaS provisioning and patching reveals the patterns that are possible. To find out more about the HP DBaaS Solution: