As most of you have seen by now, IBM has entered into an agreement to acquire Storwize. Being an 'insider' I obviously can't disclose anything about the transaction or the roadmap of the technology inside IBM but what I can tell you is that the Storwize technology will be transforming the data storage industry.
In a webex performed by Storwize and IBM a few months ago, John Power discussed how the trend in the disk drive business, the $/GB being cut in half every 18 months, is not possible to sustain, at least for the foreseeable future. Even the folks in Almaden labs haven’t been able to crack the code. Disk drive can’t spin any faster, and you can’t put any more drive heads on disk actuator arm to get more data on a disk drive.
The storage industry evolves such that when internal technology can’t evolve any further, external forces need to come into play. External forces (appliances) allow vendors to test theories in an effort to prove which new technologies are most effective and give vendors time to figure out how to integrate the new technology into the device. This IS Storwize. Bloggers have written that Storwize doesn’t stop the deluge of data at the source, the users or applications that create it. Bu t until there is no longer a need for data, then we will keep on having a need for primary storage. Bloggers have also said that by having an appliance that sits in front of your data, the only way to get it back is by having the appliance there to decompress the data. I have news for all of you, if you want to optimize your data, no matter what you choose ,you are going to need the keys to the castle at some point to open the front gate. The difference with the Storwize appliance is we don’t care what storage (CIFS / NFS) sits behind appliance, we are storage (vendor agnostic).
As a refresher; Storwize is an appliance that sits in front of primary storage and compresses data before it lands on the storage array. The unique Storwize IP not only compresses data, it also compresses it such that it preserves the file envelope and all the meta data about the file. By preserving the file envelope, all data that lives on the other side of the appliance looks like the data as it normally would in the storage cache and file system. Because of this unique feature, whatever the compression ratio is, is the factor by which the cache is increased. Imagine 4:1 compression giving you 4x the amount of storage cache. Cache is one of the key factors in increasing storage performance. If compressing data can give you better performance, then Storwize is a key that unlocks not only additional capacity, but has the ability to increase performance. (Compression is how Oracle’s Exadata gets better performance.) If we all believe that compression (or optimization technology) will be embedded in arrays of the future, in order to gain capacity and performance, then the Storwize appliance is step 1 in the grand scheme of where optimization may one day be. Until these technologies are embedded, Storwize has the lead in optimizing storage without compromising any of the existing storage characteristics.
The key questions the press / analyst were asking during our call the other day were:
- How does the technology work really work? How is the Storwize compression different?
- How does the Storwize technology work with data deduplication?
- What is the future of the Storwize ‘block solution’? - As far as this goes, what I can tell you is that the same RACE engine that compresses file data, compresses block data so we have many PB of 'testing' done on the engine. We also have customers who are running the product in Alpha stage today. As far as the new roadmap, I cannot comment, but needless to say, block compression is just as important as file compression so I am sure it will be a high priority.
Furthermore, Storwize customers and partners will greatly benefit. Being a part of IBM will unlock all of the potential of the Storwize IP. By allowing customers today to deploy an appliance and optimize their storage, they are giving customers the ability to optimize their storage, save money, and build a green data center.
The real question is when will "optimized" storage become the norm? What is on the horizon for making storage even more efficient? There is a startup out there that has the technology, I can feel it.