NetApp's Cloud Czar predicts the death of VMAX
Tier 1 is old-school, says CTO
Blocks and Files NetApp's CTO of cloud is predicting the death of EMC's VMAX and other tier 1 storage arrays. What will kick them off the data centre stage? Flash arrays and storage-class memory, apparently.
Val Bercovici, NetApp's "Big Data and Cloud Czar", sees basically two storage tiers in the now not-so-distant future: flash for performance (IOPS) and SATA disk and tape for capacity (plus tape), with seamless integration between these two tiers. The performance tier could be for working set data running very close to the app on a server.
NetApp has a middle ground between the high-performance and bulk capacity camps and has developed a Virtual Storage Tier VST) architecture to fill that gap.
This virtual tier of storage extends to server hosts, and is data driven in 4KB increments, updated in real time and self-managing. It is optimised for performance and efficiency, has a unified SAN and NAS architecture and is dedupe-aware.
The disk drive array physical back-end features deduplication, compression, thin cloning and provisioning. In the array, Bercovici blogs, "NetApp has been able to drastically reduce the effective cost of NAND flash-accelerated storage via deduplicated and thinly-cloned FlashCache blocks."
Application performance demands will continue to pull flash up the storage stack and into the server. Over time, even storage semantics themselves will give way to persistent (non-volatile) memory semantics enabling simpler and faster high performance, real-time applications. But Storage Class Memory [PDF] and NetApp’s future as a memory vendor will have to wait for another blog.
Bercovici says that EMC may have bought Xtremio to bolster itself against "NetApp’s highly anticipated Goldilocks Scale-Out Virtual Storage Tier."
Goldilocks is a pretty strange product name. Presumably it is a reference to the fussy nature of the young protagonist of the eponymous fairytale. High-performance storage flash would be too fast, high-capacity SATA Disk/tape storage too slow, but NetApp's VST solution would be "just right" for the data centre.
He foresees the death of EMC's VMAX, HDS's VSP and IBM's DS8000 arrays:
When the flash disruption is complete over the next few years (before the Storage Class Memory reverberation is expected) the once lucrative Tier 1 Frame Array market will be literally disintegrated. ESCON/FICON attached arrays will continue to leech off the relatively moribund mainframe market. Fault-tolerant data persistence functionality will move from Tier 1 Frame Arrays up the stack all the way to the application layer.
As we’ve already established, performance capability will move to the server/host layer, leaving data management, protection and efficiency to a shared storage architecture. A capacity-optimised layer will dominate here, leaving precious little room for archaic Tier 1 Frame Arrays.
Bercovici adds: "Being pulled apart from both the Performance and Capacity ends, the era of the all-encompassing Tier 1 Storage Frame Array is rapidly sunsetting."
And what's that rising in the east? Why NetApp of course: "Only comprehensive Unified and Extensible Storage architectures like NetApp’s VST provide a future-ready framework for planning, building and running an infrastructure flexible enough to cost-effectively incorporate new storage media technical trends in a consistent, predictable and sustainable manner."
So, to sum up: NetApp is going to produce storage-class memory products or manage them, and it's going to introduce all-flash arrays that may be located very close to servers.
It needs to. The latest Gartner NAS/unified storage survey, "Market Share: Network-Attached Storage/Unified Storage Worldwide 2011", shows the firm losing substantial ground to EMC, which has seized the lead in that market with 42 per cent of the revenues and an 89.9 per cent year-on-year growth rate – compared to second place NetApp's 36 per cent and 18.6 per cent annual growth. NetApp is getting trashed in its very NAS heartland by EMC. ®
The bears are coming . . .
"NetApp is getting trashed in its very NAS heartland by EMC."
This. EMC may be the storage company people love to hate, but they are moving quickly given their size, while NetApp appears to be flailing around, pushing out incremental updates to Data OnTAP and presenting no concise vision of their future plans.
Seriously, is this sponsored by the Department of the Obvious Department?
"NetApp's Cloud Czar predicts the death of VMAX"
I rank that one up there with "Our initial assessment is that they will all die" from the Iraqi Information Minister during the invasion. From Baghdad Bob to Valley Val.
Re: Lets look at what potentially the best NetApp customer in the world is doing....
Yes, I think you are right. Google has 10,000s of nodes, so if a node goes down they replace the entire server and figure out why it went down after the fact. I assume they don't throw away the entire server but handle break/fix depot as a separate process. It is wasteful in the sense the hardware sense, but not wasteful in the staffing sense. Even if they had PFAs on the servers, it would be such a huge manual effort to fix the server in place instead of just replacing the node at their crazy scale. It could be done more efficiently. The largest problem with Google is the low utilization rates on their servers. I have always thought that Google IO intensive workloads would be well suited to a few rows of large mainframes with Linux. There are definitely problems with the Google model, but I think that is definitely the "cloud" model. Not the EMC/NetApp (as much as NetApp slams EMC, they are basically following the same strategy) approach of high cost hardware, high cost software. NetApp is only bashing the tier one providers, the "big three" of EMC, IBM and HDS, because they can't compete in scale up storage. They have a filer which they have rigged into a SAN, not a SAN which also handles NFS.
All of these proprietary storage vendors, NetApp and EMC, like to talk about "cloud" and bash the other for not being a "cloud" solution. Yet none of the "cloud" companies use anything resembling their model. In the "cloud", neither VMAX or NetApp exist.