If you have been following my blogs, you know that I’m an automotive enthusiast. I appreciate all types of automobiles from trucks to supercars, combustion or electric, whether advanced or simple like my very first, a 1967 MGB (which, out of necessity, is how I learned to work on cars.)
Today’s storage systems are akin to a Formula One (F1) race car. F1 represents high performance and cutting-edge technology and is a proving ground for advanced materials, telemetry and packaging that once proven often makes its way into the cars we drive every day.
Did you know that 14 people are involved in changing the tires on an F1 car?
As with data storage, in F1, downtime loses races, so they’ve mastered changing tires in two seconds during pit stops. Like a storage system with a failed drive, you can drive on a flat tire but you’re not going to win the race – you better get it fixed quickly.
During an F1 race, even the lightest contact with another car or obstacle will end the race, or at 230 miles per hour, much worse. Even with $100 million plus budgets and big teams to tend to its every need, everyone still has their fingers crossed that nothing goes wrong.
Like Formula 1 cars, conventional storage systems and first-generation scale-out systems are really, really complicated.
The problem is that you know it, I know it, and most importantly, the customer feels it – every single day (said slowly and with emphasis) …
Equally complicated and technically advanced are self-driving cars, but the application of technology is completely different.
The explicit goal in autonomous driving is to deliver radical simplicity.
An autonomous vehicle is focused on getting you somewhere. You don’t need to understand how it works to benefit from it. Give it your destination, and it figures it out from there.
Like the self-driving car, Datera is an application-driven autonomous data services platform for modern data centers.
Datera incorporates advanced technologies, delivers incredible performance and provides continuous operation. Most importantly you get these benefits without having to know how it works – it is autonomic by design. Give it your intent, i.e. your destination, and it takes you there. Simple.
Datera had a recent win at a customer that was attributed to our radical simplicity.
Their previous supplier, a three-letter storage company, requires 27 steps using three different tool sets to provision each and every single volume.
- Step 1: Specify the size.
- Step 2: Specify the protection.
- Steps 3 – 27: Specify the media, ports, hosts, replication, snapshot, compression, etc.
Keep in mind that before starting on their 27-step program, someone had to figure out the answer to all those questions and how they mapped to the different tool sets and layers of storage systems.
I think their tag line was “One Volume, Hours of Fun.” OK, maybe not. But you get my point – it’s complicated. Imagine mapping 27-steps-per-volume to a DevOps environment with hundreds to thousands of volumes, and in single day, dozens of ephemeral volumes that may be created and destroyed. Remember F1 – 14 people just to change the tires? What happens when it is only you?
In today’s data centers, radical simplicity is an absolute necessity.
So – how does Datera deliver radical simplicity?
Datera came at simplification from the viewpoint of managing by intent. One API call, one second – the system does the rest, provisioning all the storage for an application autonomically. In today’s world where an application may be instantiated ten times, or perhaps even a hundred times a day, why would you create or manage the application’s storage one instance at a time, or worse, one volume at a time?
It’s not just about accelerating the provisioning of storage – that’s just the first lap of the race. It’s about adapting during the race in response to flat tires, changes in weather and even fixing the track. Autonomy is not about the first lap, it’s about winning the race – delivering data to an application, for many years, through good times and bad, for better or for worse.
With Datera, there are five related mechanisms that enable autonomic operation at scale.
The customer only interacts with the first and simplest mechanism:
- A single API call creates all the storage, connectivity and data management for an entire application. The customer specifies the application and who can access the data. That’s it! The subsequent mechanisms are all autonomic, requiring no interaction with the customer – i.e., they’re self-driving.
- The control plane inside the system maps intents to policies used to create and holistically manage all the storage for an instance of an application over the lifecycle of the data. These policies are the conversion of the simplified customer intent to the internal activities necessary to deliver on the intent. Policies can be changed anytime.
A hierarchical policy structure allows the policies for application instances to be adapted based on best practices.
Multi-tenancy enables fine-grained tenant control by creating leaves or branches from the hierarchy.
The hierarchy enables a change to a base policy to autonomically propagate to tens or hundreds of application instances with a single operation.
Dual maps of the current and future state enable the control plane to reshape data storage, distribution and protection. Reshaping occurs in response to:
- Internal changes – adding storage, failure recovery, optimizing resource utilization, etc.
- External changes – changes in policy, adding more tenants, etc. The system operates with these dual maps all the time, not as an exception.
- The maps point to nodes in the cluster. The extent store software in each node collects QoS telemetry and performs policy-specified data management functions, e.g. tiering, snapshot, de-duplication, encryption and replication, all optimized for the type of media present in the node.
If we put voice control on Datera, provisioning and managing storage would go something like this:
Customer: Datera, create MySQL storage for Bob and Jane.
Datera: OK, I created MySQL storage for Bob and Jane.
Like an autonomous car, Datera is full of complicated technology. Hundreds of sensors, thousands of components, millions of lines of software. Also, like an autonomous vehicle, you don’t have to understand any of it to benefit.
Let me close with a radically simple thought:
At the end of the day, winning in business and in Formula 1 is all about spending more time on the track and less time in the pits. And isn’t life in the winner’s circle a whole lot more rewarding than doing time on pit row?