Today I got to demo advance capabilities and various possibilities in & around vRA 7.2 (vRealize Automation) to an audience. This involved auto-scaling, on-demand load balancers, application authoring and containers. So, if you are a cloud admin or infra admin and wondering what features you can implement using vRA, then this post shows few of them.
If you want to build a cloud environment with a lot of flexibility then vRA is the tool for you. The power of vRA comes from the customisation capabilities through vRO. In vRA 7.2 it had been taken to another level with NSX integration and container integration through Admiral.
The infra that I am running is on nested environment. Every disk is a vDisk, even the VSAN that you see is a VMDK. So performance is pretty slow. Also I have NSX on top of NSX (VXLAN on top of VXLAN) so I have limited capability in terms of network connectivity. Please bear with me for all that. The plus point, it will show how robust VMware environment is.
Before I start I wanted to answer few questions regarding Auto-Scaling. Seems this is the most in-demand feature for any cloud. Customers mostly expect this feature to work as it works in public cloud. Please understand, what you see in public cloud is an end result. Cloud providers like AWS and Azure spent considerable time and effort to build those. Also they do have their limitations as well. So, when you build a cloud solution and act like a provider, you need to spend some time and effort to do the same. It is wrong to expect every feature out of the box. For example, you can build a building using Lego Blocks, but expecting it OOTB is wrong? What you need to check is, whether you have the right tools to help you to build those. vRA along with vRO gives you those exact tools.
Saying above, in future I expect this particular feature to be available OOTB in vRA. Currently, you already have Scale Out/In option out of the box.
Another point, without knowing whether auto-scaling will help them or not everyone wants it. For example, it is very easy to scale-out or scale-in a VM. But if the application in the VM is not clustered, then there's no point with auto-scaling.
The opposite is also true. A lot of times people ask me that auto-scaling at infra level is fine but what about application? Understand, in auto-scaling the major task is spinning a VM or container based on certain parameters. If the application is cluster capable, it will be able to take help of auto-scaling. Application may be anything like SQL cluster, Oracle RAC or Apache Tomcat cluster. If clustering and runtime node addition or removal is supported, then auto-scaling is for application. If the underlying infra supports scaling, then put any application on top of it, and scaling is done for it.
As a provider, the most important question is, how exactly the underlying infra is scaled and which parameters are supported.
What is demoed for vRA 7.2?
Mainly I am showcasing the following:
- How you can build 3 tier/complex blueprints in vRA
- How you can get on-demand load balancer and security groups through NSX integration
- Auto-Scale Out/In with varying inputs
- Different mechanism for auto-scaling input trigger
- Explain and demo application authoring in vRA using Software Components
- A demo to show vRA to host Container Host and deploy containerised application in the container Host
Interested to see the demo? Since I am showcasing a lot of use cases, the video is 1 hours 11 minutes long. I promise it will be worth the time :).
This video should give you a pretty good idea about advance capabilities of vRA 7.2. Specially if you are interested in Auto-Scaling. For more details please refer to VMware Site. Hope you will find this post useful.