Serial Attached SCSI (SAS) has in its latest version, SAS 2.0, moved beyond being a simple drive connect interface and now has the potential to function as a basic storage networking technology. It, along with a host of other protocol options such as iSCSI, ATA over Ethernet and NFS, can be a viable shared-storage option for entry-level SAN customers. So what does SAS 2.0 bring to the table that makes it different than these other technologies? And does it make sense for you to position SAS for the first-time shared-storage customer?
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
As discussed in a recent SearchStorage.com article, "Is SAS-2 a viable alternative to Fibre Channel drives?" while it's often thought of primarily as a new drive interface, potentially replacing Fibre Channel hard drives, SAS 2.0 is also a sharable technology. The specification allows for the use of expanders: Think basic switching that allows multiple servers to attach to a single array and share access to volumes without the use of any storage virtualization software. And SAS 2.0 drives' 6 Gbps speed and price advantage make them an attractive alternative to Fibre Channel drives.
As a default, with SAS 2.0 all the attached servers can see all the volumes that an array has provisioned. In the past this was not something you typically wanted to do. But in today's virtualized environment, it's a good thing to have multiple servers able to see the same volume, since that enables features such as live migration of VMs. The virtualization software's file system -- VMware's VMFS, for example -- manages the locking between the hosts.
Of course, in most nonvirtualized environments, you don't want multiple servers to see the same volume. To address that, SAS 2.0 has the capability to provide a form of zoning, similar to SAN zoning. This allows a volume of shared storage to be separated from other servers to ensure data integrity. Expect customers to want a mixture of free-access volumes and private volumes. This need could be commonplace in data centers with a growing server virtualization infrastructure that still have a number of standalone, nonvirtualized applications to maintain.
SAS' primary limitation is that its level of granularity is at the drive level. Unlike storage virtualization solutions and SAN RAID solutions, which will allow you to configure multiple volumes that share a set of drives grouped in a RAID set, SAS will allow only one volume per drive group. For example, if you created a RAID group from four 500 GB drives, with storage virtualization or a typical SAN array, you could create four or five logical volumes of varying sizes across that set of drives. With SAS, you could have just one volume that would be either shared with a group of hosts in a virtualized environment or allocated to only one server. This creates the potential for capacity waste since multiple servers would be unable to share the RAID group's capacity.
The second limitation of SAS as a shared-storage option is that the software included with SAS-based arrays is basic. Most will only allow the creation of RAID groups as described above. Advanced features, like data protection, snapshots, thin provisioning or replication, aren't yet available. You'll need to determine whether your customer really needs all of those features. If so, most OSes and virtualization software include basic snapshots, and you can supplement the array with replication, thin provisioning and data protection via additional software. (This provides further opportunity for you to add value through integration work.)
SAS 2.0 benefits
For these two disadvantages, the customer gains a system that is incredibly easy to use because it basically acts like local storage. The only complexity is the need to know how the sharing will work. To some extent, the lack of the software stack mentioned above -- thin provisioning, replication, etc. -- further simplifies operation. Sometimes features get in the way. And while adding features such as thin provisioning and replication to the system may introduce more complexity, it can now be done selectively and on a gradual basis.
The most significant advantage, though, is cost. SAS-based systems are dramatically less expensive than their shared-storage counterparts, in some cases by 50% or more. The expanders needed for sharing are inexpensive. Consider also that more and more server-class systems are including a built-in SAS interface, and the cost of the systems becomes even more affordable. It's also important to know that these cost savings come without a performance loss. SAS drives are quickly becoming Tier 1 drives in almost all tiered storage platforms. The 6 Gbps transfer speed of SAS 2.0 drives and reduced cost is attractive to suppliers, integrators and end users.
Does SAS 2.0 replace your current SAN/NAS offerings? Of course not. But it does provide customers with another shared-storage option when they either don't need or can't afford the capabilities of the traditional entry-level SAN. As customers' needs become more sophisticated, storage virtualization software could easily be layered on top of the SAS infrastructure for scalable growth and capability. The bottom-line benefit of a shared SAS network is that your customer gets a high-performance shared-storage option at a reduced cost -- and the customer needs your expertise to plug in storage services they may want.
About the author
George Crump is president and founder of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. With 25 years of experience designing storage solutions for data centers across the United States, he has seen the birth of such technologies as RAID, NAS and SAN. Prior to founding Storage Switzerland, George was chief technology officer at one of the nation's largest storage integrators, where he was in charge of technology testing, integration and product selection. Find Storage Switzerland's disclosure statement here.