Figure 1 below shows conceptually how we have traditionally done computing, prior to the Cloud Computing generation.It is obvious sharing of workloads across servers was not a common proposition unless some form of clustering was applied. Individual servers had to be scaled to meet workloads. Virtualization provided a method to scale workloads across several server resources as depicted in Figure 4.
However deployment through virtual machines is not the only way to provision computing environments in the Cloud. As we see in Figure 5, one can also automate the provisioning process bypassing the virtualization layer completely. This is typically used by workloads needing more resources and a dedicated server enables more computing horsepower.
“Bare Metal” Computing gives you all the advantages of quick provisioning and flexibility, while bypassing the virtualization layer. A special/hybrid case of Bare Metal is a “Dedicated Server”, in which case a user requests a group of servers with full server capacity dedicated to their computing needs.
AT&T’s Cloud Architect offers a full solution set to cover all workloads – multiple virtualization platforms, bare-metal and dedicated servers including inter-operability between different types of computing resources.
What is your computing workload and profile ?
-Do you need Virtual Machines ?
-Do your Virtual Machines have to be on a certain Virtualization Platform or Hypervisor ?
-Do you need Dedicated Servers ?
-Do you need Bare Metal Instances ?
-Do you have a hybrid need that goes beyond these architectures ? We look forward to your answers.