How to deal with data in a multicloud environment

Linh Le

Follow these three best practices to keep the use of multiple databases across multiple clouds from confounding your deployments

You are setting up a quick report around sales for the year, and it needs to use three separate databases. One is an object database running on Amazon Web Services. The second and third are relational databases running on Microsoft Azure.

You have a few issues to address in this multicloud scenario:

  • Finding the single source of truth among the three databases, such as customer information.
  • Dealing with the different data types and structures native to each database.
  • Dealing with complex security layers on each database and cloud provider.

The fact of the matter is that data is not centralized. Indeed, databases these days are purpose-built, and so you have different databases for analytics, operations, and multimedia management.

A multicloud approach brings an extra layer of complexity due to the fact that you have security at the database layer (such as encryption), at the platform/operating-system layer, and at the cloud layer.

Working with multiple databases on a single cloud is difficult enough, so just wait till you work with multiple databases across two or three clouds at the same time for the same application use cases! This is a wall that many enterprises are running into now, and it will get worse in 2019.

So, what are the emerging best practices? Here are a few to try:

First, consider a federated database, which is an abstraction layer that exists between the source databases on the multiple clouds, creating a combined virtual database with a structure and data that you can create custom for your use cases. The federated database software, cloud or not, handles the difference in the structures of each database, including security, and operational issues. So, you’re accessing a single virtual database—instead of two or more directly—that you can customize for your exact needs.

Second, do some advanced planning on databases you’re going to use. Use cases are typically not surprises, so I’m often taken back when the what, when, and where they need the data for is often incompatible with the databases and cloud platforms the organization has selected. A bit of thinking ahead will eliminate most of these problems.

Third, implement a master data management approach and tool set. This won’t solve all issues, but it will get you to be honest around what the right source is for the right data. Yes, it works with cloud-native databases.

I suspect that I will be dealing with this problem more and more at multicloud makes things more complex. But, with a bit of forethought this shouldn’t be that big of a speed bump.

Share the news now

Source : https://www.infoworld.com