Flying Through Clouds: Provisioning and Deployment

By Dan Chilton and Bhavik Desai, NetApp Solutions Performance Engineers


In our last Flying Through Clouds installment, we covered establishing the right performance methodology and compared that to establishing a flight plan. As you think about deploying and provisioning for new and/or existing applications, the key thing to do especially if the task is going to be repeated quite a few times is to standardize. Consistency breeds predictability. Standardizing on a consistent set of practices and processes takes the guess work out of a complex set of tasks like provisioning and deploying applications in a datacenter.


The Southwest Airlines Way


A great example of how standardizing has helped an organization is Southwest Airlines. The airline industry is a commoditized market where most of the major airlines are operating at a loss. Southwest Airlines is the most profitable airline in the US. A big reason of that is they have embraced standardization in pretty much everything they do. They fly one big plane, the Boeing 737. 


This simplifies complexity in their processes and helps them standardize on things like parts inventory, pilot training, and plane maintenance. From the pilot to the baggage handler, everyone is trained on single standard equipment which also helps in knowledge management. Standardizing on equipment and processes helps them drive higher operational efficiency in running their business.


NetApp on NetApp


Provisioning and deploying IT services in a datacenter is a complex task.  Let me share with you an example of how NetApp IT standardizes these tasks. For each service platform that Dan identified in the previous blog, NetApp IT has a set number of storage controllers and disk shelves that fit into one complete rack or half rack. 


So when the time comes to deploy a new service platform due to reaching the thresholds of performance or capacity of the existing platform, they just deploy a new service platform with the standard configuration and add it to the existing environment. Leveraging the flexibility of clustered Data ONTAP, they can then provision new applications or move existing applications to the newly deployed platform seamlessly. Since the hardware is standardized, even provisioning of aggregates and volumes for every service platform is automated using OnCommand Workflow Automation and follows the same process every time a new platform is deployed.


Making the Transition: From Incubator to Live Nodes


Managing the transition of applications from 7-Mode to clustered Data ONTAP can seem daunting at first.  NetApp is NetApp’s customer #0 for clustered Data ONTAP and we have been going through this transition ourselves for several years. When moving workloads from 7-Mode to clustered Data ONTAP, NetApp IT follows a set of procedures that we would love to share with you now.  The clustered Data ONTAP setup has 2 nodes, which serve as incubator nodes for incoming workloads from 7-Mode. The workloads go live on the incubator nodes and stay there for 24 hours where they are monitored, with the help of OnCommand Insight.  After checking the water levels, the workloads are moved to a service platform that can provide the SLO required for the application.  This ensures careful consistent deployment of workloads on clustered ONTAP followed by easy migration to the appropriate service platform.  It’s working well for us and we believe it can work for you, too.

Having a process in place like what I described above, combined with the power of Clustered Data ONTAP and tools like Workflow Automation and OnCommand Insight, allows NetApp IT to deliver superior services to stakeholders and helps IT become an enabler to the business.