Need to be very accurate with this as we're quite late on delivering Beta. Sorry it's A LOT of work. In a nutshell: we'll do log-structured file system (something VMware got with Virsto) except unlike Virsto and ZFS we don't use log to accelerate random writes, we have a solid log-structured file system (read - no background process stealing your IOPS moving data from vLog to vSpace). So no more random writes with StarWind. Also LSFS will incorporate new version of deduplication (current one has issues with being used as a VM primary storage). We'll be specially flash optimized as a primary storage and we'll be using flash as an L2 cache and one of the tiers (yes, automatic storage tiering will come either with flash caching or just a bit later - depends on other tasks). True async replication. Massive scale-out (beyond 3 nodes, definitely not SMB scenario). Closer integration with vCenter and VMM (we're already done with SMI-S provider and we'll continue integration so you'd be able to manage storage and VMs from one place). Cloud gateway thing. That's it... Well, and a new re-worked and simplified GUI of course
Not sure what did you find in Nexenta... There are other guys like Oracle who have actually the original team and rights and the most current codebase for ZFS they sell and support. Also there are other guys who like Nexenta did not write ZFS but they are not trying to sell it to you giving free of charge (various FreeBSD and Linux based appliances).
robnicholson wrote:Hi Anton- I'm both excited and slightly disappointed at the same time. Unless you tell me that the roadmap delivers this within the next three months because right now, our continued use of StarWind is been challenged by other vendors, specifically Nexenta. I'd much prefer to stay with a product that we understand and know works, because we've used it for three years but shiny new toy is always a risk.
Cheers, Rob.
