I feel like I should preface my remarks today: what follows is strictly editorial. My opinion. Please feel free to disagree. It is absolutely not intended as a critique of any particular vendor. Like my remarks yesterday and in Peace, Love, and Revolution they are intended to apply generally.
With that said...
I have been seeing a bit of a trend lately to go to capacity based licensing for backup applications. More worrying however is the type of capacity: ingest capacity, not stored data capacity. Meaning that if I do 20 full backups of a server with 1 TB of data, I will pay for a 20 TB license. (Which I think is pretty counter-intuitive; most folks would assume that I would only need a 1 TB license.)
I think (and I am guessing because I am not privy to the decision process at those vendors doing this or considering it) that the motivation is to simplify the licensing process.
The theory goes like this: one license, pay per view/drink (pick your poison). We don't care how you get the data to the backup storage, how many servers, media servers, storage nodes, specialized application agents, SAN connections, tape drives (virtual or physical), or whatever. Do it however you want, we will support you as long as you pay the capacity based license.
Which of course would be wonderful, if it actually worked that way.
The fly in the ointment? Somehow, licensing costs seem to get inflated along the way. Sometimes by as much as 5 times.
If you would have paid $100,000 for the collection of backup software components you needed before, you would now be paying $500,000. Yipes.
Makes we wonder if motivation #2 doesn't start with a "g" and end with a "reed".
So what is my beef with this?
It is actually not the price. Not directly. Vendors are free to charge whatever they want for their applications. If users don't like it, they won't buy it. They will seek alternatives that are less expensive. No problem. (Barring egregious cases of lock in and monopolistic practices.)
What happens however is an unintended consequence. At least I hope it is unintended.
And it is this: people will change they way they do backup to minimize licensing costs. Not to provide the best service level. Not to minimize backup times, or recovery times. When the price becomes too high, people will alter their behavior to reduce it. And that is putting the cart before the horse.
I actually think, regrettably, that this is fairly likely. Simply because it is such a pain to switch backup applications.
Licensing costs should not drive technology utilization.
Technology utilization should drive licensing models.
I shouldn't be forced to do an incremental backup rather than a full backup just because my licensing is so expensive.
One possible solution is to reduce the capacity charges to a sane level.
Another is to switch backup application providers.
No solution (and the consequences I describe) is not an option.