DataCore Technology and WHS2011/SBS2011

• Application to MS Server Products: Exactly how could I use your products in MS Server applications? What are the key strengths to your technology?

DataCore software is built from the ground up as a .NET based application running on Windows Server, functioning as a “storage controller.”  It is written to harness the power of  x86/x64 Windows servers to speed up disk response (a portion of the memory on the server is leveraged as disk cache), replicate storage for backup, manage and protect stored data and other advanced functions, while generally managing the allocation of the virtual storage pool.

When it comes to the Windows Home Server and Small Business Server applications, we expect to be providing similar benefits – with the caveat that we’re still working with Microsoft and the user community to determine the final set of features that will be included in our first release. We do know that we need to provide similar capabilities as were delivered in Drive Extender – which is not a challenge for us, as these features have been part of our core software for a long time (drive pooling and data mirroring). There are other features in our enterprise products that could also be of value to home and small business users – we’re evaluating them and the time it would take to fold them into future releases for WHS and SBS.

• How could storage virtualization from DataCore enhance the WHS2011 and SBS2011 products?

We believe our proven technology will bring enterprise-class storage functionality to the home and small business user, combined with user-friendly management tools.  Bottom line: We believe that users of these products will end up with greater ability to effectively manage their storage environment.

• Which technologies do you intend to bring to WHS2011 and SBS2011?  I assume storage pooling and data protection as the big ones, maybe thin provisioning.

Our initial focus is to provide key features such as disk pooling and mirroring.  As we progress we’ll evaluate other functionality such as our block-level snapshot capabilities to determine the value to the users.  As part of this process we welcome feedback from the community.

• Can you give us an overview of how your drive pooling technology works?

Virtual disk pooling is DataCore’s over-arching feature responsible for consolidating storage capacity from like or unlike disk resources.  The pool may encompass a variety of brands and models of disks, effectively creating storage tiers with different price/performance/capacity characteristics.  We’ve heard feedback from users that they would like the flexibility to create different types of pools, our technology provides this capability.  Pooling is fundamental to storage virtualization, enabling virtual (or logical) disks to be rapidly created from blocks of space on the physical devices.

• Does your technology support mixed hard drive sizes/models?  Or do I need to use all the same type, brand and size of drive?

You can have drives of varying sizes and brands in a disk pool, however, for data protection (mirroring) the drives need to be formatted to the same size.

• What hard drive specifications work best with your product?  Can we use normal desktop drives or are enterprise grade drives recommended?

Any type of disk drive visible to the OS can be added to a disk pool (i.e. SCSI, FC, iSCSI, SSA, SAS, SATA, ATA, or IDE).  USB drives are not supported as part of a disk pool.

• Any size restrictions on pools that can be created?

There are no technical restrictions on the size of the pool. We have customers whose DataCore disk pools house petabytes of capacity.

• Can you give us an overview of how your storage protection technology works?

Multiple physical drives can be added to a DataCore virtual storage pool.  Within the virtual pool you create mirrored drives – identical copies of the data you want to protect. The drives maintain their mirrored state through “write acknowledgement,” information sent back to the application once the mirrored block-level data is secure in two separate physical locations – whether in two separate hard drives under your desk or in a completely separate building. (Mirroring is one of the core high availability features used by our enterprise customers to protect important data in the event of natural disasters, power failures, etc.)   A good overview of this (for our enterprise product) can be found at http://www.datacore.com/Software/Features/Synchronous-Mirroring.aspx.

• How much of a storage penalty does your approach imply?  If you are strictly mirroring, then I would expect 50%; if a hybrid RAID, then less storage space allocated to data protection.

Our implementation delivers drive mirroring (RAID-1) within the pool (as well as between pools).  RAID-1 results in 50% usable capacity assuming all of the data is mirrored.

• Can you tell us how DataCore technology handles simple drive failure?  Is my data protected while the drive is being replaced?

Upon a drive failure occurring, the primary path goes into write-thru mode, the idea being to get the blocks directly to oxide.  Your system continues to operate during this period, albeit no longer in a mirrored state.  Upon replacing the drive, the drive pair initiates a rebuild process to resynchronize the mirror.

• Can you give us an overview of how your thin provisioning works?

Sure!  This feature is one of the main reasons that customers use DataCore software.  There is a great overview at http://www.datacore.com/Software/Features/Thin-Provisioning.aspx.  In a nutshell, thin provisioning allows you to pool physical drives into virtual volumes which are then presented to the application.  As the application needs to consume storage, it’s given a 128MB chunk to write to (this is the default size).  As it fully utilizes the allocated space, it’s given another chunk from the pool.  The net result is an efficient use of available capacity through “just in time” provisioning.

• Would you want to implement this in WHS2011 and SBS2011?

Absolutely!

• How do you intend to allow users to manage the virtual storage?  Will another program outside the server console be required, or will you integrate that with the console?

Based on feedback from the community it is clear to us that ease of use is one of the most important requirements for a WHS add-on.  Our plan is to link directly into the WHS console, allowing the user to access the DataCore functionality by simply selecting a tab and making basic choices related to disk pooling and data protection.

• How will home and small business users utilize these powerful enterprise tools?  Simple administration is a core capability of WHS version 1.  Will your implementation be simple enough for this user base to embrace?

Our goal is to bring enterprise-class functionality to the home user in a way that is intuitive and easy to use.  We look forward to feedback from the community as we get closer to the release of our product.

• What specifications will be required to run your products on WHS2011 and SBS2011?

In addition to the OS, the minimum requirements are:  1.4 GHz or greater x86 or x64 CPU, 1GB available disk space, 2GB of RAM, directly-connected disk (i.e. SCSI, FC, iSCSI, SSA, SAS, SATA, ATA, or IDE)

• When do you expect to be able to release your product to OEM’s?  WHS2011 and SBS2011 are expected to ship in the first half 2011.

We are planning to introduce our solution in Q2.

• Do you anticipate making this technology available to self-builders?

Our initial plan is to introduce our solution via WHS OEM vendors, however, based on user demand we are also exploring other routes to market.  Stay tuned!

We value and appreciate user feedback, our team can be reached at whs@datacore.com.  As news becomes available we’ll also be updating our WHS website at http://www.datacore.com/Solutions/storage-virtualization-and-virtual-server-desktop/Microsoft/homeserver.aspx

Summary

Carlos tells me that DataCore intends to be actively engaged with us as they move forward and that they welcome our feedback as they bring this technology to WHS and SBS.

This technology looks very interesting to me.  This is proven enterprise technology, is currently in production and is from a company with a track record in the space.  I intend to keep a close eye on progress here and look forward to getting my hands on a WHS2011 system with this technology loaded.  As far as DE replacements go, I think that this is the most interesting alternative.

What do you think?  Comment away!

DataCore

DataCore WHS Page