The model is based on drives rather than capacity, and the changes to the engine give customers confidence they can use maximum sized drives without concerns about restore times.
Today, storage software vendor StorONE is making a pair of announcements. First, they have revamped their S1:Engine by remaking their storage I/O, auto-tiering, and erasure coding in their 3.8 release, to allow customers to use high end storage without concerns about long rebuilds. Secondly, they use this increased efficiency to support a new pricing model that enables customers to scale for free.
“There are two big things in this announcement, the new tech release and the industry’ first scale for free pricing module,” said George Crump, StorONE’s CMO.
StorONE emerged a decade ago as a unified storage vendor, and that is still a key part of their message.
“The consolidation message is still there,” Crump said. “The typical vendor trying to consolidate storage needs to eliminate three or four storage systems at once, and the other unified storage vendors tend to be third or fourth in the category. What we do is compete as a leader in each category. We are incredibly successful in backup because of our efficiency. The same software works with databases, at the other end of the spectrum. This single engine can drive a variety of different use cases.”
Crump reviewed the changes to the engine.
“We rewrote and collapsed the storage IO stack – so theres no 30-year old code that slows down today’s hardware innovations,” he said. “This allows us to deliver for the customer the maximum flash performance utilization. 12 drives is about 1 million IOPS.”
They also rewrote the autotiering.
“Legacy hybrid systems are based on the false premise that SSD will fill up, then trickle down to the HDD,” Crump said, “Real data centres have peaks and valleys where you design for the peak. We look for moments of lower utilization and use them to move data from the flash tier to the HDD tier at those times. We try and keep space available for the next peak, so we intelligently use predictive algorithms.”
The third major change was rewriting the erasure coding for high performance.
“This is something that we can do since we own the IO stack,” Crump said. “We want customers to use HDD, not the mid-tier storage that our competitors use. They are concerned that with HDD failures, rebuild times will be long, so we rewrote the erasure coding for high performance. RAID can’t do this. We also now deliver very fast rebuilds, so we can bring it back in three hours rather than three days. It enables the use of 20 TB HDD safely, without worrying about a weeks’ long rebuild.”
The new engine facilitates the new pricing model, which replaces capacity-based licensing strategies with a scale-for-free model that leverages the full benefit of the innovations in media density.
The model encourages the use of high-density drives, since it is not based on capacity pricing.
“This means that there is a lower cost upfront, and after four or five years we literally take out the existing drives and replace them with what is current at the high end,” Crump said. “There’s no extra charge for the software licensing. This is why we can say that we scale for free.” The only limitation is that there is a minimum number of drives that needs to be updated at a given time.
“It’s a new model, that came from a few directions, and which enables is to do what is right for the customer,” said Gal Naor, StorONE’s co-founder and CEO. “We are the only company who can enable it because we are the only one who can maximize the drives this way.”