Monday, November 24, 2014

Part 4 - vRealize Operations Manager Deployment Models!

Now that we are aware of the new features and the architecture of vRealize Operations Manager, with this post I will talk about the various deployment models which are possible while deploying vROps 6.0 in your infrastructure. I will be writing about the options available to those who are already running on vCOps 5.x for moving to vROps 6.0 in the next part, however with this part I am focused on talking about the deployment architecture of the solution by considering various deployment scenarios.

While talking about the Cluster Architecture of vROps, I also pointed out to the fact that the scalability and performance of vROps 6.0 comes form the in-memory database - "GemFire" which allows you to access hot data from the memory layer making it fast and also allows you to stripe this data across multiple nodes.

Here are a few design principles which we need to consider while choosing the deployment model:-

  • vROps 6.0 Cluster Nodes are required to be in a Local Area Network.
  • For collection from remote vCenter Servers you would require a REMOTE COLLECTOR
  • Deploying a Cluster Architecture is not a compulsion. It is done only when scalability & resiliency is required.
  • For 100% resiliency it is important to protect the MASTER NODE by using a MASTER REPLICA.
  • Each node in a cluster has to be of the EXACT SAME configuration (CPU, MEMORY & STORAGE)


Now let's look at some of the deployment scenarios and what would be the architecture of vROps.


Deployment Example 1 - Multiple vCenter servers in one datacenter location!


With a single datacenter with one or Multiple vCenter Servers you would deploy either one or more than one node in the cluster depending on how you want to scale the cluster for monitoring more resources and also increasing the resiliency by enabling the HA feature. All the users login to this one site.


Deployment Example 2 - Multiple vCenters in one location & vCenters in many other different datacenter location.



With the above scenario, you can either choose to have an individual vROps instance (clustered or non-cluster) in each site or you could have one centralized vROps instance in the major location, while you can have remote collectors in each of the other sites which have vCenter Servers in them. If the centralized deployment is used then all the users will have a unified view through the centralized datacenter. With vROps instance in each site, the users will login to the site they wish to access.


Deployment Example 3 - Multiple vCenters in two main locations.


In this third and final example, each datacenter has a number of vCenter Servers hence you can choose to deploy a separate instance of vROps cluster in each datacenter. In this case, the users will login to their local vROps instance to access the resources they are monitoring.

With the next part, I will take you through the steps to install & configure the vROps appliance and then we will look through the various options available within the solution.

Till then..


Share & Spread the Knowledge!!




2 comments:

  1. hi Sunny, your blog on VROPs cleared all my doubts, i am deploying vrops in 8 node cluster, can you suggest me the best place to deploy vROPs, we have different environments like dev, Prod and etc

    ReplyDelete
    Replies
    1. Glad the posts are helpful Pratap.

      You can deploy it anywhere... Most of the organizations treat their monitoring systems as business critical and hence deploy in Production.

      Delete