Moving Toward the Software-Defined Datacenter
The Software-Defined Datacenter is the next stage of evolution for the datacenter. The move toward the Software-Defined Datacenter began with the rise of virtualization as a mainstream technology. While virtualization started as a handy technology for setting up development and test environments, the advent of bare-metal hypervisors like VMware’s ESX Server boosted the performance of virtual machines to be comparable to physical servers.
May 29, 2013
The Software-Defined Datacenter is the next stage of evolution for the datacenter. The move toward the Software-Defined Datacenter began with the rise of virtualization as a mainstream technology. While virtualization started as a handy technology for setting up development and test environments, the advent of bare-metal hypervisors like VMware’s ESX Server boosted the performance of virtual machines to be comparable to physical servers.
Now, with VMware’s vSphere and Microsoft’s Hyper-V virtualization platforms both capable of delivering in excess of one million IOPS for a single VM, it’s clear that virtualization can handle virtually all production workloads. Server virtualization has become a fundamentally evolutionary technology for IT. The abstraction of the virtual machine (VM) from the underlying server hardware allows the VMs to be deployed, cloned, live migrated, and replicated in ways that a physical server never could be. This server-level abstraction provides the key for the other changes that are required to unlock the path to the Software-Defined Datacenter.
The next steps on the path toward the Software-Defined Datacenter are network and storage virtualization. Just as server abstraction frees the server and application workloads from the underlying physical server hardware, network and storage virtualization abstract the logical network and storage components from the underlying physical networking and storage infrastructure. Network virtualization enables you to define a logical network that your application uses, which is actually a layer on top of the physical network.
Network virtualization is typically implemented using the Network Virtualization using Generic Routing Encapsulation (NVGRE) for Microsoft virtual networks and Virtual Extensible LAN (VXLAN) for VMware virtual networks. A key advantage of network virtualization is the fact that it enables your virtual network to span both on-premises and cloud infrastructures. This enables you to seamlessly move VM’s between on-premises and the cloud with no need to worry about whether the cloud and your on-premises network are using different subnets.
Storage virtualization works in much the same way. Storage virtualization enables end user provisioning and automation without users needing to know the details of the underlying storage infrastructure. For instance, EMC offers a storage virtualization technology called ViPR. ViPR is storage virtualization software that separates upper-level storage management functions (the control plane) from lower-level storage access functions (the data plane). ViPR can work as a storage controller for newer storage arrays and can also work with existing VMAX, VNX, Isilon, and Atmos products. ViPR can also integrate with cloud providers using REST APIs and HDFS access methods.
The Software-Defined Datacenter builds on server, network, and storage virtualization to create what is essentially a virtual or software-defined datacenter. The key point is that the infrastructure is no longer defined by physical systems and networks. Instead, the datacenter is a logical entity that’s entirely defined in software. It’s important to understand that the software defined datacenter is not the same as the private cloud. The Software-Defined Datacenter can span the private, public, and hybrid clouds.
The benefits for the Software-Defined Datacenter are much like the benefits you get from server virtualization, but it’s expanded beyond servers into the surrounding IT infrastructure, as well.
The Software-Defined Datacenter provides increased flexibility and agility, as well as the ability to scale up and scale out on-demand. Different layers of the datacenter are surfaced as services.
Decoupling the hardware from the infrastructure enables increased automation as changes can be controlled using policies and scripting. The end result is that IT can be more responsive to changing user and workload demands.
About the Author
You May Also Like