Software-Defined Storage: What it Means for the IT Practitioner
Software-Defined Storage:
What it Means for the IT Practitioner
CapEx is High and Scaling is Expensive
The rigidity and complexity of existing storage environments leads to inefficient storage consumption and ROI.
To avoid storage provisioning complexity and delays, organizations often resort to requesting more storage than
they truly need. This overprovisioning leads to storage resources being allocated, but not necessarily consumed.
And since storage resources aren’t virtualized, the idle storage can’t be easily reallocated or shared, forcing
potentially more frequent purchases. Consequently, not only do storage systems end up representing a large
portion of the IT budget, but scaling storage resources also happens in chunks by either paying for expensive
upgrades or forklifting existing hardware to buy costly new storage infrastructure on a recurring basis.
A New Approach is Needed
To meet today’s challenges, IT needs a fundamentally more efficient operational model– one that’s aligned
to the business and can deliver agility and efficiency. VMware’s goal is to bring to storage the same efficient
operational model that server virtualization brought to compute. This new model changes how storage is
provisioned, and managed – it’s defined by software, not the hardware. And it starts with the application, not the
storage array. Software-Defined Storage, and the Software-Defined Data Center, enable a more agile and efficient
environment, delivering speed, simplicity and cost-effective solutions.
Software-Defined Data Center
The Software-Defined Data Center is an open architectural approach that extends the principles of virtualization —
abstraction, pooling, and automation — to all data center resources and services. The Software-Defined Data Center is
the ideal infrastructure that enables the continuous development of today’s mobile applications, dramatically speeding
up and simplifying the initial provisioning and ongoing management of the virtualized infrastructure of compute,
networking and storage. VMware’s Software-Defined Data Center approach leverages core data center virtualization
technologies to transform data center economics and business agility through automation and non-disruptive
deployment that embraces and extends existing compute, network and storage infrastructure investments.
How Does it Work?
Software-Defined Storage relies on the hypervisor and two new architectural elements - the virtual data plane
and the policy-driven control plane:
• Virtual data plane – virtualizes physical resources, abstracting them into logical, VM-centric pools of
capacity that can be flexibly consumed and managed. VMware’s implementation is vSphere Virtual
Volumes™ on SAN/NAS devices, and VMware Virtual SAN™ on x86 server storage.
• Policy-driven control plane – acts as the bridge between applications and infrastructure, enabling common
management across heterogeneous storage tiers and dynamic control of storage service levels through
VM-centric, policy-driven automation. VMware’s implementation is Storage-Policy-Based Management.
VM
VM
VM
VM
VM
Policy-driven Control Plane
Virtual Data Plane
Virtual Data Services
Data Protection
Mobility
Performance
Virtual Datastores
x86 Servers
SAN/NAS
Cloud Object Storage
Figure 1: Software-Defined Storage Architecture
W H I T E PA P E R / 3
Software-Defined Storage:
What it Means for the IT Practitioner
Virtual Data Plane
This is a new abstraction layer that pools storage into flexible VM-centric virtual datastores where virtual disks are
natively represented on underlying physical infrastructure. Using Virtual Volumes and Virtual SAN, the storage
infrastructure expresses available capabilities (performance and data services) to the virtual data plane to enable
automated provisioning and dynamic control of storage services levels through programmatic APIs.
Compare this to traditional storage environments that are predicated on rigid infrastructure-centric LUNs
(Logical unit numbers) or storage volumes that essentially are static allocations of storage service levels
(capacity, performance and data services). In this situation, storage provisioning is dictated by the hardware
available, and applications are rigidly assigned resources in a “best fit” scenario with little room for flexibility
or change.
Policy-Driven Control Plane
This new management layer provides common orchestration and automation across all storage tiers whether on
external arrays, x86 servers or cloud storage. Unlike traditional control planes, this policy-driven control plane uses
policies to define different classes of storage and automates the composition of these storage services at scale,
while enabling adjustments at the VM level. As application requirements change, policies are updated and the
hypervisor leverages the virtual data plane and virtual datastores to automatically and non-disruptively provision
the new storage resources. The control plane is programmable via public APIs that can be used to consume and
control policies via scripting and cloud automation tools for self-service consumption of storage.
In a typical storage environment, the control plane is tied to each storage device – each array with its own its
own management tools, generating uniquely defined “classes of service” or static pre-allocations of resources,
difficult to change and largely disassociated from specific application requirements.
The Hypervisor’s Role
As the abstraction layer sitting at the intersection between applications and infrastructure, the hypervisor
occupies a privileged position within the IT stack that allows it to balance all IT resources – compute, memory,
storage and networking – needed by an application. The hypervisor is inherently application-aware and has
direct visibility into each application (i.e. virtual machines and virtual disks) and its requirements. With a virtual
data plane and policy-based management in place, the hypervisor now has the ability to provision storage
resources in a similar fashion to compute resources. Now the hypervisor can:
• Automate the delivery of storage levels (capacity, performance) and data services to applications through a
new control plane.
• Abstract the storage infrastructure into flexible pools of capacity where these storage levels and data
services can be delivered to individual virtual machines.
• Manage and dynamically provision storage more precisely and efficiently through storage policy-based
management templates.
Making Software-Defined Storage a Reality
VMware’s Software-Defined Storage model shifts the operational model of storage from the bottoms-up
array-centric approach of today’s storage, to a tops-down VM-centric model. As a result, storage services are
automatically and precisely aligned to application requirements, saving you time and money.
VMware Virtual SAN
Implementing Software-Defined Storage starts
Please complete the form to gain access to this content
