The Cloud is great. It allows you to deploy applications and never have to worry about infrastructure again. But, never worrying about infrastructure again is not the same as never worrying about application availability again.
Numerous recent cloud outages and the impacts that they have had on production environments go to show that now more than ever companies should be thinking of ways to maintain compliance with standard application availability practices that have become common in on premises data centers.
However, with the ease of deploying in the cloud some application developers and cloud ops teams sometimes forget that they still need to abide by some common availability principles. Moreover, as applications continue to scale and developers deploy workloads faster than ever how are administrators supposed to maintain compliance with cloud availability practices.
Turbonomic, allows companies to automatically set and abide by compliance and availability polices without the need to constantly reevaluate deployment architecture and strategy. Thus allowing developers to continue to innovate at cloud speed.
On Premises Redundancy
In the world of the physical datacenter redundancy is fairly straight forward. As an administrator you can run planning scenarios to determine if you are able to support workloads in a failover situation.
The output of the plan answers the question of what will my utilization look like after a failover and how much capacity to I need in order to support workload demand.
When it comes to cloud we no longer need to worry about whether or not a datacenter has capacity to support our workloads. Instead the question becomes is the application running in enough availability zones in order to achieve 99.9% uptime?
Application Geography Redundancy
Where do you deploy workloads in the cloud today? Do you have a favorite region, a favorite zone? The majority of customers today place workloads in the closest region to them geographically. However, AWS guarantees 99.9% uptime if you are deployed in more than one availability zone.
Turbonomic allows your organization to define how many and which regions or availability zones an application should be deployed across in order to maintain compliance with geographic availability policies.
Determine which data centers applications should reside across:
Setup region redundancy:
Receive compliance decisions in order to maintain redundancy:
By defining policies per app, you can be sure that mission critical production applications remain geographically distributed, while less important development workloads can reside geographically close to the end users.
Application Level High Availability
Geographic redundancy and failover capacity are two ways of making sure that you are meeting HA requirements both in the cloud and on premises. Once the infrastructure is solved the real question becomes is the application itself highly available.
In addition to specifying datacenter redundancy requirements Turbonomic allows you to specify application high availability requirements.
By understanding application availability constraints Turbonomic is able to make sure that there are always enough instances running to be able to support a minimum level of workload. If the number of application servers drops below the specified minimum Turbonomic will give start commands in order to assure performance.
Putting it All Together
High availability is thought of as a critical piece of on premises datacenters, unfortunately “the cloud does it for us” mentality has taken over as companies shift IT operations to AWS, Azure, and Google. Yet you need only look at recent events to understand why HA should be taken more seriously in the cloud.
Using geographic redundancy in conjunction with application level high availability Turbonomic is able to ensure that your applications are always up and running in a highly available fashion. Using these compliance best practices in conjunction with SLA adherent scaling assures application performance while maximizing cloud efficiencies.