Forum Discussion

GertjanA's avatar
GertjanA
Moderator
9 years ago

ArchiveName, number of items, compressed size, uncompressed size

Hello all,

Every now and then, I run the below SQL.

I only recently noticed that for several archives, the uncompressed size is smaller than the compressed size. This might be due to these archives being around since EV4 (or 5 perhaps), but I thought the compressed size (i.e. in EV) would always be smaller then the compressed size.

Query used:

SELECT A.ArchiveName "Archive Name",

COUNT(S.IdTransaction) AS ArchivedItemCount,

SUM(S.ItemSize)/1024 "Compressed Size (MB)",

SUM(SP.OriginalSize)/1024/1024 "Uncompressed Size (MB)"

FROM EnterpriseVaultDirectory.dbo.Archive A,

EnterpriseVaultDirectory.dbo.Root R,

VaultStoreName.dbo.ArchivePoint AP,

VaultStoreName.dbo.Saveset S,

VaultStoreName.dbo.SavesetProperty SP

WHERE A.RootIdentity = R.RootIdentity

AND R.VaultEntryId = AP.ArchivePointId

AND AP.ArchivePointIdentity = S.ArchivePointIdentity

AND S.SavesetIdentity = SP.SavesetIdentity

And (A.ArchiveName = 'name1' or A.ArchiveName = 'name2')

 

I have as example:

Archivename items compressed uncompressed

name1 2444 238 237

name2 152 11 9

On the other hand, I also have some with massive compression.

85680 items, compressed 8457, uncompressed 14889

I believe the query is correct (as it is a copy/past (from JW I believe)), but I'd like a reverification please. Will the compression be better when there are more items in the archive? did compression improve during the versions? (currently we're on EV11.0.1 CHF2)

Thanks

Gertjan.

 

 

  • What you're probably seeing is things like JPG files being compressed which makes the entire thing bigger.

    But really the problem with this type of stuff is that the space savings really come from OSIS more than compression, but if you have something like a 5KB message sent to three people, you're going to end up with more space used up in EV than you would in exchange

    Each message is converted to the DVS, but also has user information added.
    Each person who recieved the message will have their own DVS file, and their own indexed item and it becomes messy.

    But the hope is that the pay off will be when 100 people archive a 5MB attachment, so you have something like 100 DVS files, 1 DVSSP and 1 DVSCC and saves you 490MB of date through single instancing

     

    Really the uncompressed Size is good for knowing how much storage you would need when restoring back to the mailbox. if you export to a PST file, it will be a lot bigger, but if you simply added up the compressed size and said this is how much space is used on EV storage, it won't be anywhere close

2 Replies

  • What you're probably seeing is things like JPG files being compressed which makes the entire thing bigger.

    But really the problem with this type of stuff is that the space savings really come from OSIS more than compression, but if you have something like a 5KB message sent to three people, you're going to end up with more space used up in EV than you would in exchange

    Each message is converted to the DVS, but also has user information added.
    Each person who recieved the message will have their own DVS file, and their own indexed item and it becomes messy.

    But the hope is that the pay off will be when 100 people archive a 5MB attachment, so you have something like 100 DVS files, 1 DVSSP and 1 DVSCC and saves you 490MB of date through single instancing

     

    Really the uncompressed Size is good for knowing how much storage you would need when restoring back to the mailbox. if you export to a PST file, it will be a lot bigger, but if you simply added up the compressed size and said this is how much space is used on EV storage, it won't be anywhere close

  • Hi JW,

    Thanks. I was wondering more about the other way around. Is the uncompressed size what will be in Exchange when restoring archives to Exchange (2013) mailbox?

    In other words, taking your example:

    I have an archive of 100 people who all received this 5MB attachment. On EV, it takes up 10MB.

    The query above probablly shows something like "1 item, uncompressed 5mb, compressed 4mb" for each of these 100 archives.

    When restoring those archives back to exchange, I expect to see 500MB in exchange (+the indexing overhead)

    right?

    Thanks!