During the last two weeks I was out of the office for some vacation and only in the last couple of hours I’ve been able to read all of the posts that piled up on the blogosphere.
I do agree with Hu that before virtualization was quite impossibile to migrate chunks of data storage between tiers but now it’s technically easy to develop new features like this one. Furthermore I think that in a (maybe or maybe not) close moving data to/from the cloud will be desiderable an cheap but now it’s simply impossibile since there are no standard protocols to do this in an easy way!
But what is the subject of this post? The ease of use of new features (Tiered storage in this case)!
Ease of use, low acquisition costs and ease of implementation of new features in a storage systems means less TCO. One of the most important thoughts in this days of recession is to cut the TCO (Total Cost of Ownership) in IT and, above all, in Storage! Tiered storage going that way but to reduce the costs we need to comprehen the complexity to may hide behind the implementation of0 a feature and how this impact on the rest of the storage or whole IT infrastructure!
One example? EMC will deliver FAST (Fully Automated Storage Tiering) next year but only for the new V-MAX… if you have a DMX, you need to buy a new storage, understand the implementation of FAST and relative constraints (probably FAST will limit some other features…. ), train people on the VMAX and FAST, migrate all data, hire consultants to implement tiering (will it works as expected in release 1.0?????), dismiss your old DMX, 🙁 Uhm!
Storage tiering needs to be simple! if we think about fully virtualized storage systems or a federated pool of storage systems we need to be sure to not spend valuable and scarse resources to implement new feature every time they are released! It is likely to inexpectedly become a pricey affair, since some times, costs are
Compellent has his Data Progression: a fully integrated feature for automated tiered storage. This feature is a softwre license to add to Storage Center systems…. every generation of Storage Center is capable of Data Progression!!! As soon as I enable it I can use it for every LUN that are just a “point and click” away from associating them to the right tiering profile!
so, to recap:
EMC (and other taditional storage vendors): buy new stuff, train (new?) people, hire consultants, migrate data, adopt new features (FAST in this case) and manage new costraints == very high TCO!
Compellent (and other next generation storage vendors): buy a license, use it immediately without specific training, no migration or whatever else! == very low TCO!
If your storage infrastructure TCO is too high you will evaluate Cloud Storage as a viable option for tiering… but if you can lower down the TCO than you will be able to do tiering in your local, fully integrated ease of use infrastructure!!! 😉