I spent a lot of time this week with various storage vendors, including Dell, Coraid, Pure Storage, Brocade, Tintri and more. While the presentations tended to be very technical, some practitioner tidbits managed to make their way to the surface as well and, in reviewing the various sessions, there are some clear commonalities that I saw across the board and they spell good things for businesses that want to refocus their Information Technology efforts more toward bottom line-driven business objectives.
Bearing in mind that the meetings this week revolved solely around storage, the most significant trend that I saw was simple: Simplicity. Of course, every vendor is different, but most of the vendors that we spoke with, when describing the problem they’re trying to solve, understand that IT organizations need to spend less time on touching the infrastructure and more time on the business.
That means that infrastructures simply need to run. We need to either make infrastructure someone else’s problem or implement set-it-and-forget-it systems that don’t require constant care and feeding.
One way in which infrastructure can be made easier to manage is by taking some of the guesswork out of how the environment is operating. One vendor in particular is doing a really nice job in helping businesses gain increasing insight into their environments. Tintri, in my opinion, provided the most impressive look at latency because they did so in and end-to-end way from the perspective of the application. For example, suppose an application begins to suffer from what seems to be a performance problem somewhere. Rather than just tell the administrator that Tintri’s storage array isn’t the problem, Tintri is actually digging deeply into vCenter and pulling enough statistics to be able to allow the administrator to immediately and accurately identify the exact source of the bottleneck. Perhaps the host is overcommitted from a processor perspective or the network can’t keep up. Or, perhaps there really is a problem in the storage array somewhere. It’s easy to identify and because it’s so easy to identify, an administrator can take corrective action much faster than would otherwise be possible. That kind of transparency adds to the simplicity of the solution.
Tintri also provides higher-level informatics to help administrators make better decisions with regard to storage. While most storage devices provide a capacity gauge that slowly dwindles as more and more data is moved to the device, overall storage performance has been more of a mystery. Tintri provide a gas gauge-like performance view so that administrators know exactly how much performance capacity is left on the array while continuing to display how much storage capacity remains.
However, while I was very impressed with Tintri, Drobo—a company with products aimed at the SMB—also didn’t disappoint. Drobo has certainly grown over the past couple of years, moving from a company catering to the enthusiast that needed a lot of direct-attached storage to one that can compete in the SMB space with iSCSI-based shared storage solutions.
For the SMB space, Drobo brings their BeyondRAID technology, which takes the pain and suffering out of disk redundancy planning. BeyondRAID is a proprietary single-disk or dual-disk (user choice) data protection method that Drobo created so that administrators have just one decision to make with regard to data protection—how many disks can I afford to lose? Drobo also allows users to add disks of different sizes to a single BeyondRAID group. From there, BeyondRAID figures out the best way to deploy its 1023.75 MB stripes to protect user data.
Beyond being automatic, BeyondRAID brings to the table some enterprise level capabilities that SMB’s can use… again, with no hassle. These features include thin provisioning and dual disk redundancy.
As is the case with all of the vendors I saw this week, Drobo has also added flash disk capability to its arrays. The company has built specialized processes for ensuring that the flash disks are used to best effect. Again, user intervention isn’t required to try to figure out how to deploy these disks in a way that makes sense. Drobo just does it.
That’s also the case with Tintri. Tintri uses a combination of spinning disk and flash disk to achieve its performance and capacity goals and takes a “write everything first to flash and then figure out what should be moved to spinning disk” approach by analyzing what data is hot and what data is cold. Although the company plans to add new features over time that allow administrators to tweak how things work, Tintri is pretty much an out-of-the-box solution that you simply deploy and use. It really is that simple.
I’d also like to briefly mention Coraid and their ATA-over-Ethernet (AoE) protocol, which may prove confusing to former Age of Empires players. It is also confounding to those that are storage aficionados as Coraid has basically thrown out the rulebook and decided to simply use Ethernet—for all intents and purposes, anyway—as the transport. Their AoE protocol is actually something of a misnomer since storage commands are actually SCSI in nature, but Coraid is able to achieve significant scale at prices that will please many. As is the case with Tintri and Drobo, Coraid also sells simplicity as a part of its solution. However, because of the product’s uniqueness, it may not suit everyone. Check it out for yourself.
Although focused on storage, this week’s meetings proved to me that many up and coming and even established vendors and really looking for ways to make IT components that are more appliance-like in nature and that require less administrative overhead. This helps CIOs implement infrastructures that are more automated and more self-healing than they were in the past. While this could allow organizations to possibly reduce their IT staff in some cases, it’s more likely that smart organizations will choose instead to repurpose some of their positions toward business goals and help IT move from a cost savings mindset to one that adds value to the entire organization.