Veeam will cause additional IOPS on the datastore of the source VM (and a target VM, in a replication job) potentially impacting other VMs in the same datastore.
In some cases this may cause the performance of other neighboring VMs (on the same datastore) to decrease during this heavy I/O period.
This heavy I/O may be either the result of a Veeam backup or conversely, may negatively impact the performance of a Veeam backup. If some VM(s) (not being backed up) is monopolizing the datastore with heavy I/O this may cause Veeam to have small slice of the disk I/O available during backup processes.
This is not specific to Veeam and the issue can be addressed by VMware by use of Storage I/O Control (SIOC).
SIOC is actually using a congestion algorithm to determine a normal latency for storage devices in a datastore and then spread this evenly across all VMs to assure an equal share of the available IOPS for each device.
Note: NFS storage is now supported in vSphere 5.x.
SIOC protects all virtual machines from unwarranted negative performance impact due to I/O-heavy virtual machines. SIOC prevents a single virtual machine from monopolizing the I/O throughput of a datastore. Storage I/O Control (SIOC) was introduced in vSphere 4.