P4000 snapshot vs. competition

[righthandsan]

Why is it that P4000, when creating the first snapshot, consume as much space as the volume itself? For example, if I have a fully provisioned 1 TB volume, and take a snapshot, the initial/primary snapshot is also 1 TB. Now I've used 2 TB of the precious space. This is quite different from how other SAN vendors, e.g.. NetApp, EMC, do their snapshots.

What would be the reason behind HP/LeftHand choosing to consume this much space when creating snapshots? Are there real benefits to this approach or is this more of a shortcoming?

Anybody with both P4000 and NetApp (or EMC) experience care to comment on the differences in the snapshot technologies? 

[bryan_hadzik]

It doesn't consume as much space as you think. It is all "virtual" space anyway.

Lets take your anlogy of the 1tb volume. Lets also consider 500gb is actually used up. The instant you take a snapshot, that 500gb gets shifted into the "old data" column, and consumes 500gb. the volume now consumes 1gb(1gb is the smallest it starts at). after a few minutes/hours, that 1gb may grow to 3gbs, so the total consumed space is 500gb+3gb = 503gb. The next snapshot kicks off, and consumes 1gb, then slowly grows to 5gb. 500gb + 3gb + 5gb = 508gb.

It all depends on how you count the space. For our example, the snapshot takes up the most amount of space, because that is where all the data resides. Almost all data is "old" data anyway.

Bottom line: It is not doubling your usage or anything even close. It only has to keep track of the changed blocks since the snapshot. The reports showing consumed data in snapshots are sometimes confusing.