Veeam schrijft het volgende:
More and more organizations are choosing VMware Infrastructure to virtualize their mission‐critical applications (Active Directory, Exchange, SharePoint, SQL Server) to create a flexible, easily administered virtual infrastructure.
Virtual machines (VMs) and any applications they contain must be protected against failure. Typically, in the virtual world, this is done by performing an image‐level backup of the whole machine (for instance, using VMware Consolidated Backup). This method results in what is known as a crash‐consistent image. Restoring a crash‐consistent image is essentially equivalent to rebooting a server after a hard reset. For operating systems, this has not been an issue, since they can easily handle this type of activity. For database applications as well as for applications featuring replication, however, such a restore will often result in lost data, data corruption, or application failure.
To illustrate this concept, let’s review the consequences of using different VM disaster recovery methods for one of the most common mission‐critical applications: a Microsoft Active Directory domain controller (DC). Depending on the solution used, the following results will be achieved:
Download de PDF hier: VMwareandVSS-ApplicationBackupandRecovery.pdf
ICT-Freak.nl 
A very interesting read indeed. The example of a domain controller is not entirely fair (given the “””limited””” support for virtualized domain controllers, see for example MSKB 888794, which already indicates that domain controllers in virtual environments need to be handled with care). However, it clearly indicates that a complete VSS integration for full virtual machine backups is an absolute necessity if these kind of backups need to be relied upon in disaster scenarios.