Introduced in vSphere 6.0, VVols is a new integration and management framework that abstracts and virtualizes SAN/NAS storage with a software-defined storage approach, based on SPBM:
The VVols architecture has five major components which are described as follows:
- VVol object: It can be config-VVol (metadata), data-VVol (VMDKs), mem-VVol (snapshots), swap-VVol (swap files), other-VVol (vendor solution specific).
- Protocol Endpoint (PE): Although storage systems manage all aspects of virtual volumes, ESXi hosts have no direct access to virtual volumes on the storage side. Instead, ESXi hosts use a logical I/O proxy, called the PE, to communicate with virtual volumes and virtual disk files that virtual volumes encapsulate. ESXi uses PE to establish a data path on demand from virtual machines to their respective VVols.
- Storage Container (SC): Unlike traditional LUN and NFS based vSphere storage, the VVols functionality does not require preconfigured volumes on a storage side. Instead, VVols uses an SC, which is a pool of raw storage capacity or an aggregation of storage capabilities that a storage system can provide to VVols.
- VASA provider (2.0): A VVols storage provider, also called a VASA provider, is a software component that acts as a storage awareness service for vSphere. The provider mediates out-of-band communication between the vCenter Server and ESXi hosts on one side and a storage system on the other.
- Array: This implements VVols features; note that not all storage vendors implement VVols in the same way, so it really depends on the maturity of their solution.
For more information, see KB 2113013—Understanding Virtual Volumes (VVols) in VMware vSphere 6.0 (https://kb.vmware.com/kb/2113013).
With the new vSphere 6.5, there are some improvements:
- Support for storage replication (VVols replication enablement)
- SPBM for availability
- Support for Oracle Real Application Clusters (RAC) on VVols
But other features will be added to VVols. During VMworld 2017 the roadmap for the next release of vSphere (probably vSphere 6.7) was announced:
- Support for Microsoft Cluster Server (MSCS) or fail-over cluster): Actually, it's implemented using RDMs, but VVols can provide more flexibility in the future.
- Performance parity with VMFS: Also, if the I/O data paths and is multipathing between a host and VVols, there are some minor changes; for example, to the path going through a PE to get to the VVols. Anyway, it seems that not all is already tuned both at storage and host side, so some speed improvement is still possible.
- IPv6 support: VMware vSphere already supports IPv6, but not yet on the VASA provider service.
Today, the bind operation happens out-of-band using the control path through the VASA provider, and the goal for the future is to bring it in-band to the data path through the PE instead. Still, some features, such as NFS v4.1 support and in-band binding, are not yet present (and will probably be implemented in the next major release).
There are a lot of papers and documentation on how VVols could be useful in a vSphere environment; for example, the recent IDC research VVols Provides Powerful Application-Aware Management for vSphere Environments which can be found at URL http://idcdocserv.com/US42988017. But there is not much data on the real adoptions in production; that seems still limited. Also, it is a technology that depends too much on the storage vendor's implementation (VMware provides just the framework).