From the course: Cloud Complexity Management for Multicloud Deployments

Abstraction

From the course: Cloud Complexity Management for Multicloud Deployments

Start my 1-month free trial

Abstraction

“

- [Instructor] So let's start with the data domain, and the concept of data abstraction. Data abstraction is ultimately the ability to deal with complex data that typically exists within the enterprise with the on-premise systems, and perhaps on a public cloud, such as Amazon Web Services, Google, or Microsoft. Typically this is around heterogeneous data. The ability to leverage different databases to solve issues at the time leads us to a huge portfolio of different databases that are around, both in the cloud, and on-premises. The ability to abstract the data or basically create versions of the data that exist only in memory are ultimately where the solution resides that's going to solve the complexity issue. So we deal with tools such as data virtualization tools, which are able to hide the physical schemas from us, enable to simplify the way in which we access very complex data sets that are on very complex platforms. So an example would be some sort of data service manager. So ultimately it is the jumping off point for us to access the information. From there we deal with abstractions, or virtualization tools, and we may have a master data management system, we may have other things that are in essence bringing us a version of the data that's represented in a structured schema that may only exist in memory. Then we also have the physical data, the ability to deal with the array of different database technologies, built in different generations and times, storing information in different ways, and hiding the complexity from you by layering these abstractions. So, in order to solve the problem, we have to leverage virtualization tools that allow us to deal with abstract data. There are examples of these including Denodo, TIBCO, and Red Hat Jboss Data Virtualization tool, just to name a few. There are actually dozens and dozens on the market. And as part of this process, you're going to go out there and find the tool that's right for you in terms of your ability to abstract the complexity of the various systems. So ultimately you have to keep a few things in mind. Number one, the tools are really where the value is going to be in your data abstraction technology, therefore you have to look for them very carefully. Ultimately you have to consider other operational aspects of this, such as the impact on performance when dealing with abstraction tools. Obviously you're changing things, there's processing that's occurring between when you're abstracting the data, and when the data is being presented, and that needs to perform up to expectations. Ultimately security should be considered systemic to all of this. If we're dealing with data, we're dealing with compliance issues, we're dealing with regulations, you know, such as HIPAA, and other laws, and other constraints, and other policies that surround the use of the data. Those have to be enforced. And ultimately we have to consider how guardrails are going to be placed around the data. How we're going to govern the data long term, in terms of changing the information, how it's being displayed, and how it's bring processed.

Contents