Monday, August 24, 2009

EMC Clariion enhances its virtual experience and joins the spin down ranks

EMC announced today several enhancements to their Clariion product line ranging from optimizing the Clariion CX4 virtualized experience, to increased connectivity and scale in the virtual environment to the introduction of an energy efficient spin down feature.

This announcement details a number of innovations that raises the competitive bar for EMC's midrange competitors particularly in the virtualized storage space which will no doubt inject energy into this platform. Guess the close relationship with VMware does have tangible advantages. I appreciated the toggling feature of RecoverPoint as innovative and offers a neat solution to demand spikes.

My first impressions is that there is little unique innovation with the "green" spin down feature which has a look of catch-up rather than breaking new ground. All major vendors already have this capability with EMC being late to the party. I was looking for more innovation from the leader in online data storage such as displayed by the second tier player Nexsan. There is a white paper that is due in September that will hopefully give the detail necessary to enable a sensible analysis of this EMC MAID-like feature. I will revisit this post when I can digest the content of this paper.


Key points in the announcement were.

  • Virtualization Aware Navisphere.
    - Providing integrated views that bridge the management gap between the VMware Administrator and the Storage Administrator.
    - Providing end-to-end virtual to physical mapping.
    - Example given was a 10 ESX server environment with a total of 100VM’s; the pre-enhancement case featured a very time intensive task in recording information from 200 screens onto a spreadsheet which was in contrast with a post-enhancement task of two screens that contained the information to enable a quick determination of the latest virtual to physical mapping. Time required for the task dropped from hours to seconds
  • Connectivity and scalability
    - 8Gb/s FC was a Q2 announcement with 10Gb/s Ethernet iSCSI featured in this announcement
    -FCoE support targeted for 2010.
    - Claim to have increased the number of VM’s supported on a single platform by 8x.
  • Enhanced RecoverPoint protection.
    - One view of physical and virtual protection that shows the replication status of each VM in RecoverPoint interface
    - Interesting feature is the dynamic toggling between synchronous and asynchronous. - This switching mode toggling is also supported on DMX4 and V-Max, Celerra with a Clariion backend and non EMC storage.
  • Replication Manager Integration.
    - Designed to bring instant image-level replicas of VMs to EMC’s full line of storage arrays.
    - A stand alone product and not a plug-in for vCenter or Navisphere but communicates to Flare via public API’s..
  • Drive spin down.
    Being a MAID bigot I was particularly interested in their spin down implementation. Unfortunately not much detail was made available in the announcement which causes me to reserve judgment other that to applaud the effort to join the ranks of meaningful energy efficient online storage.
    However what is known:
    i. The feature will be standard in all arrays.
    ii. Three of the five beta sites are now in production.
    iii. The drives come to a full stop and do not appear to have any intermediate reduced speed options such as with Nexsan.
    iv. The spin down policy is set at the RAID group level and automatically spins down a RAID group when it has been idle for 30 minutes.
    v. Kudos however on having a number of “Spin Down Aware” applications ready to go including Networker, Documentum Content Manager, DiskXtender for Windows and the version for Unix, General Purpose NT Filesystem, CIFS/SMB File Shares. Work is in progress with Celera FileMover, Avamar, SourceOne, MFST SQL, MFST SharePoint and MFST Exchange.
    vi. Because all drives can spin at any one time this approach does not disconnect maximum storage capacity from power provisioning. If all drives can spin then planning must be for maximum potential power requirement. Only COPAN with its native MAID architecture enables thin power provisioning.

1 comment:

tomlassen said...

Bill,

Nice enhancements to Clariion, but they only work on Clariion hardware. This just perpetuates the old model of building silos of storage with dissimilar architectures that don't interoperate and have separate management utilities.

Truly open solutions like Nexenta's ZFS-based product runs on industry standard hardware and supports any host, any O/S, any network and any storage. Run it on bare metal or as a guest in a virtual infrastructure. In fact, NexentaStor has very tight integration with VMware, Xen and Hyper-V to simplify provisioning as well as the moving of images and storage without disruption, with a single command.

Finally, it's really, really low cost...about 70% less than any software from a hardware vendor's street price.

Think of the savings... Keep your existing storage; cancel the S/W licenses and install NexentaStor to manage not just your Clariion environment, but all of your NAS and SAN storage resources. Instant savings! Then when the lease expires, replace the expensive storage with low cost-no frills storage at a fraction of the price. More savings!

We have all witnessed open standardization and convergence in the host & O/S world as well as in the network realm. Standardizing and opening storage is way, way overdue and open solutions like Nexenta's are leading they way.

The age of proprietary storage is over.