I have three words for you, Automation, Automation, and Automation. This is where true disruption in the industry is happening and it’s happening right now. The fundamental question is why is “Automation” so important for the future of our Industry and more importantly when you scale that question back, only encompassing your world, how will I benefit from Automation?
This is not going to be a history lesson on where we came from or how we got here in the Virtualized Data Center, we’ve heard that more than enough at this point. This post simply addresses all the technical areas that Virtualization companies, more-so VMware, are looking to automate.
The buzz word or Acronym I should say is now SDDC. For those of you that don’t know what this is, it stands for Software Defined Data Center. The name itself implies why we must move all mgmt points from hardware devices into the virtual stack, but why? Well, we’ve heard everyone say we are taking the same steps with Storage and Network as we did with the Compute layer several years ago. VMware’s focus over those years was first and foremost addressing scale to enable customers to virtualize even the most mission critical, heavy resource, low latent workloads and this goal has been realized with the release of vSphere 5.5.
Now VMware’s focus is to apply those same principles to the Network and Storage stack to tackle an even greater goal. The answer is simple, all hardware stacks of the data center MUST be managed by software in the virtual stack to realize the ultimate goal. Well, what is the ultimate goal. In my view the ultimate goal, and this may be a bit shocking to some, but it’s been my belief for the last few years that Data Centers will eventually run themselves. The only way to get there is through intelligent software.
I do believe that we are a ways off from that, though we are getting closer and closer each year. In order to build that intelligence, possibly AI, it’s going to take time. In the meantime, we are starting to begin the automation journey and right now I believe that for us IT professionals that implement Cloud/DataCenter solutions we need to be a part of that journey and redefine ourselves or we will not hold the same value we have today.
So, you’re probably asking yourselves what the hell does this have to do with the title of this post? Well, i’m going to take a few steps back. First there are levels of automation from creating scripts in PowerCLI to building workflows in Orchestrator, to configuring BluePrints in vCloud Automation Center. What do all three of these have in common? You guessed it, they all still require human interaction for them to be any use to us, so they are simply a tool to solve a problem, hence the reason whey I believe we are a long ways off from Fully Automated Self Healing, data centers, “Skynet” if you will.
Policies, however, make it simple for us to guide actions. Now I don’t need to tell you that in any modern data center today, there are several tiers of resources, whether that’s compute, network, or storage and those resources are getting much more difficult to manage. What if you could logically pool all those resources, to include storage and network, and even security, where they are application aware and delivered by policy-based automation?
Let’s take for example storage, as my example will be based on vSphere 5.5 storage policies.
You have an array that provides several “Tiers” of storage and you have application workloads that require specific storage Tiers. Wouldn’t it be easier to pool all of your storage resources into a logical pool, assign policies that define the Storage Tier and when you create your application (VM) assign that Virtual machine that storage policy?
Now think about this at the automation level. Now you have a Private Cloud and you have several types of networking, storage, etc that you want to abstract and pool and have automation make decisions for you at the Virtual Machine level. You couldn’t do that without a policy, hence the requirement. For example, you want to deploy a new DB server that requires high performance compute, network, and storage. In vCAC that would be defined as a blueprint, but at the vSphere layers underneath, what constitutes “high performance?” See where i’m going with this? There has to be something in place to allow a specific course of action to take place based on specific criteria and that is the Policy. Assigning Policies to applications allows easier methods of assigning the proper resources to your application.
Stay tuned for more on Automation, Policies, Operations Management as I change focus into Cloud and Operations.