Server virtualization has brought numerous benefits to enterprise datacenters, transforming IT with increased flexibility, higher server utilization and lower costs. However, there are many challenges in managing virtual machines, from providing the appropriate storage for a single VM, the impact of sprawling virtual disks (VMDKs), the quality of one datastore serving many potentially noisy neighbor VMs, to the cost of idle VMs with their cold VMDKs taking up valuable storage resources.
These widespread VM management challenges create significant complexity and costs for enterprises. Primary Data’s DataSphere platform is inherently designed to resolve the common data management problems plaguing virtualized server environments by supporting policy-based storage with existing or future storage purchases, tiered storage with Virtual Volumes, and active data movement to ensure the performance, cost and protection needs of an application are continually and dynamically met.
- Adopt VMware’s Virtual Volumes using existing storage
- Implement VVol’s policy-based storage management across multiple storage types and tiers
- Simplify administration with a universal VASA provider and a single scale-out VVol datastore
- Ensure compliance to VM storage policies automatically as workloads change
AUTOMATE VM-GRANULAR MANAGEMENT
The advantages of VMware Virtual Volumes (VVols) come to life when combined with the benefits of Primary Data’s DataSphere software. With DataSphere, IT can make existing storage VVOL capable, set storage requirement policies on a per VM basis, add new storage types seamlessly, eliminate the cost of over-provisioning, add new data services and automate the detection and migration of inactive VMs’ data to low-cost storage.
VMware vSphere 6 with VVols enables VM-granular management, which allows admins to define storage requirements for each VM, rather than the costlier approach of placing VMs on dedicated LUNs or a volume-based datastore. VMware has defined a set of application programming interfaces called VASA (vSphere Storage APIs for Storage Awareness), that allow vCenter to pass these requirements to a storage provider and to request offloaded VM snapshots and clones.
Primary Data’s DataSphere platform provides VASA 2.0 APIs, enabling automated policy-based VVol VM-granular management across all storage, with additional data management services such as snapshots and clones. More importantly, DataSphere can enable any existing or new storage to be VVol capable. The attributes of the storage, such as performance, cost and reliability, can be mapped to the objectives set within vCenter. And unlike other VVol solutions, DataSphere can non-disruptively move virtual disks between storage devices of different types, such as SAN, NAS and Object resources. For example, active VMs can automatically be moved to fast flash storage, while cold VMs move to lower cost storage according to business objectives. While management tools such as vMotion can move VMs to different datastores, with vMotion, IT must first determine and find which performant, noisy, or cold VM(s) to move, and then manually perform the operation. DataSphere works through vCenter to receive the target objective, automates the data management activity, and continues to monitor for any changes or new demands.
DataSphere is a metadata engine designed to separate the architecturally rigid relationship between applications and their storage to achieve unprecedented improvements in performance, efficiency and scalability. In VMware environments, DataSphere gathers application or VMDK usage telemetry through a local service VM running DSX software. The DSX runs on every ESXi server and presents the VVOL or NFS datastore. The VMs consume VMDKs from the datastore. This datastore is DataSphere’s global namespace, or typically a section of the global namespace. The global namespace is an added benefit for ESXi users, as this means that other applications outside of VMware datastores can access the same physical storage resources without any conflict with the VMDKs in the datastore. In short, there is no need to suffer the overprovisioning cost of dedicating storage to your VMware cluster.
Figure 1 - DataSphere can combine multiple storage tiers into a single universal VVol datastore and automatically place virtual disks on the right tier to meet objectives
CREATE VVOL DATASTORES WITH EXISTING STORAGE
DataSphere allows organizations to create VVol datastores on existing storage devices, saving the costs of expensive upgrades. If more capacity or performance are required, adding storage is nondisruptive and the new resources automatically expand the VVol datastore. Storage types can include NAS, SAN and object/cloud, providing a blend of performance and protection abilities to meet a wide range of needs.
Figure 2 - DataSphere can combine multiple storage tiers into a single universal VVol datastore and automatically place virtual disks on the right tier to meet objectives
Realizing the benefits of VVols and DataSphere is easy, as VVol datastores can be created alongside existing VMFS and NFS datastores, and VMs can be non-disruptively added to the DataSphere global namespace. Using vCenter, admins create VM Storage Policies to define performance and protection needs and apply them to VMs. DataSphere automatically ensures the VM is placed on or moved to the appropriate storage.
AUTOMATICALLY MAINTAIN POLICY COMPLIANCE
When a policy is applied to a VM, DataSphere will place it on the appropriate storage tier. DataSphere continues to monitor VMDK performance and availability. If the storage no longer meets the VM policy requirements for performance, or if the policy changes to satisfy evolving business needs, DataSphere will non-disruptively move the VMDKs to storage that can serve the VM’s needs. DataSphere also protects mission-critical VMs from I/O starvation from noisy neighbors, either by moving the mission-critical VM to faster storage, or by moving the noisy neighbors to reduce I/O load on the shared storage resource.
Figure 3 - DataSphere can combine multiple storage tiers into a single universal VVol datastore and automatically place virtual disks on the right tier to meet objectives
FREE UP EXPENSIVE STORAGE CAPACITY
Using objectives, a VM’s data or VMDK can be moved to an alternate tier based on activity levels. For example, cold VMs can be automatically moved from expensive (usually higher performing) tiers, to lower cost storage such as the cloud. VMs and VMDKs that are moved to a different tier remain available in the VVol datastore. Moving inactive VMs can provide significant cost savings from recovered capacity. Importantly, no admin intervention is required to maintain policy compliance or realize the cost savings of freeing capacity consumed by cold VMs.
RAPID VM PROVISIONING
VMs can be provisioned in seconds from vCenter by cloning a template or existing VM. Cloning is offloaded by DataSphere so it is fast, efficient and uses little additional storage.
BETTER BACKUPS WITH FAST AND EFFICIENT SNAPSHOTS
DataSphere VM snapshots are fast and efficient, integrate with vCenter and enhance existing backup tools by enabling more frequent backups than traditional snapshots without application impact.
SIMPLIFY VM MANAGEMENT, SAVE AND SCALE AGILITY
DataSphere complements virtualized architectures to help enterprises realize the most advanced VMware features across existing and future storage resources. It greatly simplifies VM and datastoremanagement, automatically ensures VM compliance, reduces cost, and increases business agility, without the need to purchase new storage resources.
Use the Primary Data TCO Calculator on our web site to see how much DataSphere might save you.
Connect With Us
Enter your name and email to receive news and updates from Primary Data. Fields marked with an * are required: