Jeff Bezos has always advised to let your customers guide how you develop a product. In fact, one of the core missions of Amazon is to ensure, “every day to make every important aspect of the customer experience a little bit better.” This is clearly what you see when peeling back the onion on the data management company, Rubrik. Like the winged monkeys marching in The Wizard of Oz, Rubrik has charted a course in the data ocean that’s taken them from a scrappy startup with a very intriguing scale-out based value prop, to a clear contender for Enterprise data management needs. While there has been a lot of buzz around features and functionality within the product, I’d like to take a step back to analyze how I feel they hit the mark for enterprise deployments today in the first of a few blog posts focusing on the product directly.
Docker, Serverless, Immutable, oh my.
Most my enterprise customers run their businesses on top of architectures that would be considered legacy by nearly all startups. Game – Set – Match. There aren’t very many enterprises who have standardized their money-makers on containers or even in PaaS for that matter – yet. I’m not saying that they don’t have strategies and buildouts to support these frameworks. In fact, many do and are running applications today using more modern architectures. What I am saying is that, in general, most enterprises are running the apps that drive their businesses with more traditional Microsoft, Oracle, and SAP platforms. While Rubirk has worked to ensure that their current feature set will be able to be molded into supporting newer technical architectures, dues still had to be paid where Oracle and Microsoft dominate the enterprise application deployment model.
As a startup with their fingers on the pulse of the Enterprise market, Rubrik was very smart to work towards making data management of these critical apps paramount to their overall business vision. I don’t need to jump into the technical details here, there are many blogs discussing the new features that are churning out with every Rubrik release, but I will say the we should for a minute focus on what they’ve specifically done to appease Oracle and Microsoft SQL server customers with the platform. Keep in mind because these integrations don’t require any integration with hypervisor components they are fully portable regardless of cloud platform, and that kind of story matters to organizations that need to scale across platforms both on-premises and in the public cloud. Supporting these applications really lays the groundwork for a platform that supports clouds of all shapes and sizes.
SQL Server Heavyweights.
If you’re a .NET heavy shop, SQL server is likely the king and owner of your relational data. While some misconstrue the Rubrik platform as only being capable of providing a crash-consistent image similar to a hypervisor-based snapshot, Rubrik broke out the big-kid gloves to allow for very advanced recovery options for SQL server. We’re talking about table level restores within SQL server and with the help of some slick filesystem tweaking and by leveraging a connector application in the OS you can fully mount older versions of your SQL data on an existing SQL hosts by using the magic of SMB3. There are a ton of technical possibilities that are feasible leveraging this new option for looking at recovering a database, but the thing that’s great for Enterprises is that your DBA doesn’t need to really know anything new, this integrates directly with the restore process that DBA’s are comfortable leveraging today. SQL server in your VMware environment? Check? SQL server, in the public cloud? Check. SQL server running on Hyper-V? Check. They all just get backed up. You get the point, Rubrik can provide a consistent backup platform for any cloud because of integrations here with the OS and application itself.
Oracle’s covered as well.
DBAs rarely like to step out of leveraging their native toolsets, unless they’ve worked to bring on a specific tool to help support their data ecosystem. Rubrik provide an easy button for Oracle DBAs by enabling their platform for native RMAN access. While this isn’t a totally new feature to Rubrik, some of the new features released include the capability to leverage a native Oracle incremental merge to decrease the DBA’s need for taking full backups. All of these features are of course available by the same great RESTful API that Rubrik has leveraged for its product since day one. Just like SQL, because this backup capability is exposed at the application level, the cloud or bare-metal platform hosting the database engine isn’t relevant. With SQL and Oracle, Rubrik is working to be as divested from the underlying hypervisor as possible to enable next generation data portability.
Rubrik’s path to the cloud:
Rubrik, as a startup made some very wise decisions. First, they started with a filesystem developed by engineers who did nothing other than architect distributed filesystems. They combined that model with a scale-out infrastructure platform to drive a tighter line with consumption. In my previous life, as an infrastructure manager of a hosting team, one of the first people to approach me about cloud was our head-bean-counter who said that he’d like to get storage purchases more in line with consumption. With that statement he showed me a pictured very similar to this picture (taken directly from the AWS site):
So many organizations were forced to figure today’s capacity and assume 20% – 30% year-over-year growth to determine how to size a new IT asset purchase. The result of this purchasing model was overspending in year one to support our assumed needs in year three. That was exactly the bean-counter’s message. All of this was a waste compared to just leveraging and paying for what you need today.
Rubrik, by default enables a more cloud-like on-demand consumption model by leveraging a hyper-converged architecture model. When coupling this with a bullet-proof filesystem that automatically expands as you add additional nodes, enterprises can take advantage of scaling both capacity and performance by adding additional Briks on a one-by-one basis.
You say SLA, I say SLO…
Who cares about these words? Enterprises do. Which is why from the outset, Rubrik has taken a position to support the way they do business by simply letting the system do the dirty work of addressing backups and their schedules. By configuring a Rubrik SLA domain, you can then assign objects (maybe VM data, maybe database backups) to these domains and just let the appliance address taking care of the backup scheduling and execution. This feature is a simple nod to a core tenant of cloud consumption that we like to call resource pooling.
When consuming resources within a cloud, we often want to make a decision point for our orchestrations as simple as possible. The more decisions to make logically when choosing where a VM object may live from a network/storage/compute/backup perspective, the more lines of code that end up being written and maintained by your cloud team. By allowing the backup engine to simply address the details of backup and providing a simple API that can be consumed by the orchestration engine of your choice, Rubrik gives us one of the easiest backup platforms to integrate with yet. This is key as the number and complexity of endpoints and decisions for orchestrations only grows.
Groundwork partially done!
So far, we’ve addressed a few high-level components that makeup parts of the overall path for Rubrik’s application data management strategy as it relates to the cloud. In future posts, we’ll get even a bit more cloudy, by jumping into how these decisions to protect core-enterprise applications in a very meaningful way also translate into providing cloud enabled data management regardless of where your application and data make their homes. We’ll further venture into some of the awesome things that the platform is doing to help Enterprises with their cloud strategy and journey as a whole. Thanks for reading.