Many organizations assume that data moved to PaaS (Platform as a Service) platforms in the cloud does not require any administration, because “it’s managed by the cloud.” In reality, only a small portion of database administrator tasks are automated/included with a PaaS managed solution.
Setup
It is vital to consider the options involved when setting up a database as a PaaS solution. Some of these options cannot be changed once they are defined. It is important to ensure that configurations are set according to industry best practices, as well as organizational standards.
Security / Access
In the current age of data breaches, the security relating to your data on a PaaS solution is more important than ever. A strong policy (strictly enforced), is vital, both from an organizational perspective and from the compliance perspective. Cloud IAM (Identity Access Management) policies and roles control access to the database metadata, but not to the database itself, which still needs to be applied using database commands, preferably with a proper approval process in place.
Monitoring / Maintenance
While today’s cloud provides all provide a minimum amount of monitoring for PaaS solutions, this is usually focused on infrastructure-type metrics (CPU utilization, memory, etc.). To truly understand how your platform is performing, some data-specific monitoring needs to be used. Some examples might be total connections, running queries, average query time, etc. While some of these can be enabled/configured with the built-in cloud tools, others are not so simple. Think active statements over time. (For example: what was running at 7am this morning, when we received a service alert?)
Continuous maintenance is also required to keep your platform running optimally as it grows and changes over time. While backups and high availability are usually built into the platform, most other maintenance tasks need to be configured and maintained by hand. Some providers provide notifications of recommendations, such as adding indexes. These cannot just be automatically applied, they need to be investigated to determine if they’re truly necessary, or if they may do more harm than good. Automated systems don’t do this analysis particularly well. For code-first applications, this also requires manual intervention to get the recommendations applied.
Migration
How can an organization migrate its data into a PaaS platform? How do they guarantee accuracy?
Some cloud providers include migration services that can be used to migrate data seamlessly between on-prem and cloud platforms. Or so they claim. It is important to validate migrated data, and in some cases, such as views or stored procedures, may require code changes to work effectively. In the case of extremely large data sets, how do we keep the PaaS platform up-to-date? Do we need to shut down the application in order to migrate the data? These are concerns that keep application owners up at night, when considering moving to a PaaS solution.
THE BOTTOM LINE
Ascend Technologies’ Data Management team has the knowledge and experience to help with all these tasks. We can work with your development teams, business owners, and IT staff to ensure a seamless migration of all functionality, with minimal impact to existing processes.
Written by Andy Maser, Data Architect