In my last blog, I touched on the challenges of the modern IT professional and how technologies like Intelligent Infrastructure can help mitigate these challenges and protect us from the exhaustion and frustration that can mark many of our careers.
But what does Intelligent Infrastructure actually mean and to what threshold do we measure intelligence? In this blog, I aim to answer both of these questions.
Intelligence is defined as the ability to acquire and apply knowledge and skills.[1]
At Tintri, we believe that Intelligent Infrastructure needs to acquire knowledge and then apply that knowledge to optimize its own performance, settings or configurations. This “acquiring and applying knowledge” happens automatically, 24-hours a day, 7-days a week, and without human intervention.
So how do we actually measure intelligence? What is being acquired? And how is it being applied? I’m glad you asked!
When it comes to storage administration, the bulk of our time is spent calculating the needs of applications and then configuring a standard storage array to meet those needs. Inevitably those needs change (or *gasp* our calculations were wrong?!) and we start the cycle all over again… and again… and again…
I believe Einstein may have had a quote about doing something over and over again… anyway.
On average, organizations spend 2,100 hours per year or the equivalent of a full-time employee to manage every 500TB in their data center[2]. Growing storage footprints, dwindling storage expertise in the labor market, and the rising cost of software tools to manage these environments make this an untenable situation.
At Tintri, our goal is to drive both operational burden and extraneous tools to zero by building autonomous administration capabilities and tools directly into our products. We address this untenable situation by resolving the single biggest issue in your data center today: your standard storage infrastructure doesn’t understand what a virtual machine is.
Why is this important? It’s important because as much as 100% of the infrastructure in a modern enterprise exists to support virtual machines. It would make sense that virtual machines are the unit of administration across every layer in the infrastructure stack, but in the case of standard storage infrastructure, they aren’t.
Need convincing?
Imagine logging in to vCenter or your AWS console and only seeing hosts. Imagine that managing these environments can only be done by measuring and administering the hosts.
Pretty limited, right? So why then are we content to administer standard storage arrays built around these exact limitations?
Understanding what virtual machines are is step number one in building any relevant intelligence for modern infrastructure. Once the storage infrastructure is engineered to comprehend virtual machines, it can measure their behavior and respond accordingly. Now we have the foundation for the intelligence in Intelligent Infrastructure.
Our Intelligent Infrastructure, VMstore, “responds” by autonomously allocating performance resources to virtual machines so that their individual workloads get 1ms of latency regardless of their behavior. This process happens hundreds of times per day, all day, every day, and without configuration or human intervention. As virtual machine behaviors change, VMstore’s configuration changes with them.
More importantly, these complex analyses centered around learning virtual machine performance and capacity trends can be used to optimize groups of independent intelligent storage appliances.
If this sounds familiar – this is the exact method for how virtualization managers such as VMWare’s VCenter Server, manage clusters of independent compute hosts. The knowledge the storage infrastructure can acquire is not just limited to performance and capacity needs.
Tell VMstore that a certain set of virtual machines need specific data protection, Quality-of-Service, or replication requirements and the system will apply those needs regardless of which physical appliance the virtual machines live on. Should that virtual-to-physical relationship change, VMstore will change in response to maintain those virtual machine-level settings – 100% autonomously. Without these capabilities, as much as 95% of your time is wasted manually configuring (and reconfiguring) these settings every day on standard storage arrays.
At Tintri, we believe that Intelligent Infrastructure should take care of this for you, in turn, freeing up your most valuable asset: your time.
Acquire knowledge + Apply knowledge = Intelligence.
Have I piqued your interest? Are you questioning whether this is too good to be true?
I invite you to our upcoming webcast where three experts including Gartner analyst Phil Dawson, a Tintri customer, and myself will discuss the many benefits Intelligent Infrastructure and provide real-world examples of why it warrants your consideration. I look forward to seeing you there.
Regards,
Erwin
