Netvault database size (4044097)

Explanation of the storage requirements for entries in the NetVault Database (nvdb).

For reference Netvault only takes 71 bytes to track a file/directory assuming an average filename size of 12 characters - note this is filename size not complete path). So if you have 100,000 files and 2,000 directories the database would use 7.3 MB to keep track of that backup. If you backup every day, and keep track of all backups (regardless of full/incremental), this amount to 7.3Mb * 365 = 2.6 GB per year.

Here are the sizing calculations for the constituent parts of the Netvault Database (nvdb) & for the temp directory usage for backup indexes.

Index Size 

The size of the index is 71 bytes + file name not including path information for every file backed up.
The filename size consumed will be 2bytes per character of the name.
There is also a one off overhead that is dependent on the number of directories backed up, this is minimal.
As we record the entire topography of the data set specified within each index for restoration purposes the full & incremental indices are the same size.
The index flags each file that is contained within the associated backup save set as having been backed up.
Therefore the difference between the full & incremental backup indices is the number of files flagged as backed up.

Temp space usage

For an incremental we copy the previous index into the tmp directory, and the incremental pre scan takes up 24 bytes + filename not including path for every file in the selection. Plus we write the new index into the tmp dir as well.

Note: incremental indexes are the same size as full indexes.
We do not include the pathname in the filename, it is just the size of the filename that is added to the 71Bytes.
The pathname is implicitly stored within the nvdb index as part of a tree structure it is created within the backup job.
Each individual branch of the tree is backed up as a directory entry into the tree & has the same storage requirements as a file entry in the index 71Bytes + size(directory_name).

For example:
If you backup the file C:\PROGS\NETVAULT6\BIN\NVGUI.EXE then you create the following index entries:
C:
PROGS
NETVAULT6
BIN
NVGUI.EXE

When you come to restore this file then the index tree is searched for the location of the file & the full pathname is generated from the index.

=> The index storage requirement for the file is as follows
(71Bytes + size(C:) ) + (71Bytes + size (PROGS)) + (71Bytes + size (NETVAULT))
+ (71Bytes + size(BIN)) + (71Bytes + size (NVGUI.EXE)) = total index storage required.

NB: some index entries such as C: have a 1 to many relationship to files & this single entry will be reused as part of a tree structure therefore this equation can be used for the calculation of individual overheads but is not scalable for the entire index.