After NetApp and Pillar,  3Par and HDS have now their “space guarantee” programs, but not EMC!

In the last days I saw a lot of chattering about this argument from many sources (you can find some links here, here, here and one from EMC here).

Space guarantee programs are only marketing campaigns to lure customers with fireworks and not with real substance.

Some of these programs have a lot of fine prints and clauses that look like this: “…The 50 percent guarantee is for the raw storage capacity for migrating from third party RAID-1 source environments to dynamically provisioned RAID-5 target environments. For any third party RAID-5 environment, Hitachi Data Systems will guarantee a 20 percent storage capacity reduction. …”  (here the complete press-release from HDS).

So if you migrate your DB LUNs from RAID1 to RAID5 thin provisioned you can save space but no guarantees about performance! it seems more a well packaged and legalized fraud than a real opportunity for the customers!

Now I would like to show how my space (and performance) guarantee program works, a recent success case with Compellent.

My Space (and performance) guarantee program:

Compellent Storage Center with its Automated tiered storage (Data Progression), Thin provisioning (Dynamic capacity), space and performance efficient snapshot (instant replays) features may save you a lot of space, grant IOPS and big savings without any more or less hidden clause.

One of our Compellent’s customers (SCM Group) acquired their first Compellent Storage Center more than one year ago and they migrated some of their data to it, one year after (it took quite a bit to test the whole solution in production about support, real different workload, management, etc.) they then decided to move all their contents from an old HDS AMS 500  to Compellent!

The HDS was configured with a total of 1 controllers tray + 8 disk trays as follow:

• 2 trays  300GB/10K : 6.6TB allocable capacity;
• 4 trays 146GB/10K : 6.5TB allocable capacity;
• 3 trays 146GB/15K : 4.9TB allocable capcity;

All the disks were configured in RAID5 (6D+1P, 1 spare per tray).

TOTALS

• RAW disks capacity (excluding spares) is about 21TB;
• Theoretical allocable capacity was 18TBs;
• Really allocated for usage was 14TBs, due to shadow images (LUN clones);
• Really used was 12,5TB!!! (near an half of raw);

We did a complete analysis of their needs and summarized the findings in these two following graphs (click on pictures for a full size view):

iops

spazio

You can easily draw two important conclusions:

1. The space used is high: nearly 80% of usable space;
2. The real iops deliverd are not a lot (99th percentile is 4059): no wide striping.

Well, with these information and forecasted growth in our mind we designed the best solution for the customer (not on the basis of a unrealistic marketing program):

The results are the following:

• 135 FC disks were replaced by only 38 300GB/15K FC DISKS: less than one third!
• To achieve new space requirements (present + 30% forecast growth) we added 16 1TB/SATA disks!
• + 7000 measured IOPS from the Compellent upgraded (thanks to fast track and data progression), + 72% than HDS!
• 22.7TB of allocable space, much more usable than what was allocated on HDS, thank to Instant Replays!
• The Array is SSD ready (added more backend loops and some slots are free to accomodate some SSD disks in near future)!
• Space usage down to 12 Rack Units from 27!
• Huge energy savings (4 trays compared to 8 trays + 1 controller), something like 6KW (including cooling)!

And last, but not least: if we break up the cost the customer had to face for the whole operation into what covered the same space/performance they already had and what covered the increase in space/performance, well, the first part was the very same needed to just sustain the old storage (maintenance + power/cooling dissavings), all in all what they payed more for was only what they got more!

This is the best NON Marketing space (and IOPS) guarantee program I know, what do you think about it?