Thin provisioning is a great reason to love IBM SANs (Storage Area Networks), and another reason why price point comparisons must be explained. The numbers do not tell the whole story.
Last week, we discussed Easy Tier, which is an automatic, smart process that comes standard on IBM SANs. This process allows companies to buy cheaper disk space without spending extra human resources on data management, or suffering slower retrieval times.
Thin provisioning is an even bigger game changer.
Thin provisioning is a more efficient way of allocating storage space, which allows hard disks to be packed with more data. Here is how it works.
Thick provisioning allocates data for servers upfront. The SAN controller segments a large amount of data at the server’s request. Any other servers that want to use that space are totally out of luck. Think of it like this:
When it comes to data, servers are like children at a pizza party. Every kid thinks he/she can eat one whole pizza. You might have seen kids fight over the number of slices they will eat before a single piece has been consumed.
An inexperienced baby sitter might believe each kid and order ten pizzas, but an experienced parent would manage the situation better. IBM is the experienced parent in this picture.
Thin provisioning supports on-demand allocation by tricking each server into believing it has all the pizza it wants. With the other hand, the controller allows other units to use whatever storage they want. The process is managed intelligently, supporting a much fuller utilization of the storage disk’s potential.
Typically, 10% of storage space is used in a thick provisioning SAN, whereas the hard drive can be fully tapped with thin provisioning. You are looking at roughly 1,000% the space, from the controller software, alone.
You can see why price points do not tell nearly the whole story. Just looking at the last two facets of IBM SAN controllers, you could see that the controller adds an order of magnitude to the value of every terabyte of storage, and then some.
Come back next week for a discussion about FlashCopy, which is an IBM feature that falls onto the spectrum of backups and data recovery, but doesn’t exactly replace them either.