VMstore T7000 Series Icon

VMstore T7000 Series

On-prem VMstore platform.

Tintri Cloud Engine

Container-driven VMstore platform.

Tintri Cloud Platform

Managed infrastructure powered by Tintri.

All Die Hard Tintri News & Events Perspectives Products & Solutions Technical

Artificial Intelligence Storage

Intelligent Infrastructure Blog - Artificial Intelligence Loves Storage

Artificial Intelligence (AI) is everywhere in our day-to-day lives. It’s in your TV making recommendations on content to watch. It’s in your car, removing the headaches of tight parallel parking and preventing nasty high-speed collisions. And it’s probably used in places that you don’t want to think about. But ironically, it’s probably NOT being used where you’d want it to be – like optimizing your business.

In the field of enterprise IT, AI is the new buzzword. AI makes every part of a business better (or at least that’s the promise) – but what does that mean specifically for data storage? What should you make of all of the claims about AI technologies and capabilities from storage vendors? And how do DDN and its Enterprise division, Tintri, deliver on all of the promises of artificial intelligence for mid-range to large-scale enterprise environments?

Let’s find out.

What is Artificial Intelligence?

Without going into a dissertation on the various attributes of an AI and the methods of achieving them, for brevity’s sake, let’s consider AI as any system that can sufficiently process data to augment, mimic or, in some cases, replace human cognitive functions.

AIs perform within a sort of feedback loop that can be categorized by the following key stages:

  • Data analysis – where data is fed into the AI
  • Inference – where the AI infers a condition from that data
  • Prescription/action – in most cases, the intent of the data analysis and inference is to prescribe an action (e.g., movies you might enjoy)

In instances where human beings are “in-the-loop”, a human operator would either accept or reject the prescription and, if applicable, execute the action, affecting the next round of data analysis. In some cases, human beings are not critical to the feedback loop and can either act as overseers or be absent altogether.

Examples of activities where AIs have been most successful include:

  • Process Automation – At Tintri, this is how we’ve redefined expectations for administrative overhead in enterprise storage systems. Baked within our flagship VMstore systems and ecosystem are multiple AIs tasked with ensuring that everything from performance and capacity optimization to data management are handled with little to no human intervention.
  • Cognitive Insight – By essentially making sense of disparate and “non-human readable” data, you can think of cognitive insight as the core AI functionality underpinning complex dynamic processes such as autonomous driving or advanced business intelligence. DDN is the leader in supporting these incredibly taxing capabilities and has long served as the storage platform of choice for data-intensive workloads – even before modern ML and AI applications emerged from research laboratories. The combined performance and scale of DDN platforms, unique in the industry, accelerates applications and integrates larger data sets for deeper, more accurate insight.
  • Cognitive Engagement: This may be the most recognizable and relatable example of AI. We’ve all tussled with cognitive engagement AIs – whether it’s screaming “representative !!” over the phone at your virtual bank teller for 30 seconds non-stop, or asking Alexa to order another 77 rolls of toilet paper, or playing whack-a-mole with a customer service pop-over on your favorite retail website. DDN has delivered solutions that include cognitive engagement, partnering with some of the largest natural language processing firms. Our solutions include machine-learning techniques that remove ambiguity and respond to human input with a degree of nuance and context that has historically been difficult for computers to achieve. On top of this, the unique performance, capacity and on-demand scalability of DDN solutions mentioned earlier meets new needs and delivers tremendous value to these types of AI projects.

AI, Storage Systems, and Their Impact on Each Other

By now you’re starting to get an idea of how AI works with storage systems. And to be clear, there are two aspects to this relationship: it turns out that AI relies heavily on excellent storage systems and excellent storage systems rely heavily on AI.

At DDN and Tintri, our full-service portfolio strengthens both sides of this dance between AI and storage. And there’s more to come as we ambitiously cross-develop technologies to deliver the ultimate AI-focused IT experience for both Enterprise and At-Scale application environments.

For customers leveraging AI as a technology to drive new revenue opportunities, DDN delivers the most powerful storage infrastructure in the industry for critical At-Scale production workloads. The DDN infrastructure is designed for AI applications that help organizations extract and enhance the value of their data. Here’s where AI success is dependent on the capabilities of the underlying storage, and where DDN delivers unmatched simplicity, flexibility, and efficiency with turnkey solutions.  Historically, designing and implementing such an AI infrastructure in-house was an exercise akin to constructing a building using a disparate pile of LEGO® blocks, Lincoln logs and jigsaw puzzle pieces – without assembly instructions. DDN has transformed what is often a months-long endeavor into a turn-key experience – delivering the fastest time-to-production and ease-of-scale to accelerate value to the business.

On the other side of this AI / storage interaction, Tintri has focused on implementing AI as a set of integrated capabilities within our enterprise storage systems that can simplify and, in many cases, eliminate administrative tasks. This reduces management complexity and frees IT admins to do more valuable work. Here’s where storage management and optimization can be greatly enhanced by AI, and where Tintri technologies add incomparable value.

How AI Helps Solve Our Customers’ Problems

For DDN’s At-Scale customers, a top priority is getting to the value buried deep within their mountains of data. Unfortunately, many organizations have embarked on this journey toward implementing AI apps without fully understanding the potentials implications – complexity, exploding data set sizes and emerging regulatory requirements, just to name a few. According to a 2020 Deloitte survey on the state of AI, 90% of businesses say that AI is critical to their success, but only 40% feel prepared to fully address their AI initiatives. The major risk to most companies is that they will get left behind, or even driven out of business, when faced with savvy AI-driven competitors.  A recent ESG survey found that organizations with AI/ML in production are more than twice as likely to identify as being in a very strong business position (55% vs. 26%) than those still in early stages.

What Options Do Organizations Have When Facing These Challenges?

To meet these challenges head-on, DDN has pioneered turn-key AI infrastructure solutions that are easy to deploy and scale, while simultaneously maximizing agility and accelerating time-to-value. In fact, DDN is the only storage supplier to have reference architectures in place with NVIDIA for AI deployments ranging from a single processing system all the way to AI supercomputing clusters featuring 100s of processing nodes. DDN has already delivered solutions to a wide range of industries around the world, including financial services, healthcare, manufacturing, autonomous vehicle, and energy suppliers who are already changing the way they do business – and changing the world – using AI.

For Tintri’s Enterprise customers, IT agility has been a long-standing challenge. Over the last 20 years, IT systems have grown in scale and complexity and, according to a recent ESG study, over 75% of IT leaders concede that their environments have become exponentially more complex in just the last two years.

This growing complexity – driven by more apps…using more data…on a greater variety of systems – has compounding effects throughout the IT organization; chief among them is a widening time and skills gap. In short, enterprises are running out of time and skills to keep existing stuff running, let alone to get new stuff going.

So where does AI from Tintri fit into this mix? Tintri’s flagship product, VMstore, is an AI-powered enterprise storage system that we’ve designed to run itself.

“Hmmm…it runs itself” you say?  Yes – there’s a reason why VMstore is commonly referred to as “Intelligent Infrastructure”. We’ve designed the AI in the VMstore OS to integrate with the infrastructure and application ecosystem to learn the context of the files that live on its file system. Some of these files can make up virtual machines, some of them can be database files, others can be persistent volumes for containers, but they’re all understood the same way by VMstore: as “managed objects”. The AI within the VMstore OS is able to manage resources with much finer granularity because, unlike standard storage, there are no LUNs to provision and administer.

VMstore’s AI can measure the IO behaviors of each object individually and can automatically configure its resources so that each one achieves sub-millisecond latency for reads and writes. As the IO behaviors change, VMstore’s AI responds in kind – measuring up to 56 datapoints per managed object and self-adjusting to optimize resources every 10 minutes. 24-hours a day. 365-days a year.

VMstore systems can even federate together to share their analyses using Tintri Global Center (TGC), the unified control center for VMstore systems. TGC can analyze the aggregated behaviors of a group of systems and make recommendations to further optimize data placement across them all. Over the last decade, Tintri customers have realized tens of millions of hours in operational savings – savings they’ve redirected to more valuable initiatives.

Real World Benefits of AI and Storage

For DDN customers, the benefits of turnkey AI infrastructure are immediately realized. With the ability to scale seamlessly from POC to production and grow on-demand, customers can significantly accelerate their AI projects. During a recent interview with ESG, a DDN A3I customer stated, “The biggest improvement is just the ability to look at more data. We have big fast storage so we can look at more workloads, which means we have more days, months, years of data we can sift through – or it means I can get much more granular with the stuff I am looking at which means my results are going to be much more precise.” This combination of performance and granular visibility and insight is creating differentiation for companies in all industries.

But all this power can’t be harnessed by most companies if it’s not also easily integrated and managed. Another DDN customer told ESG, ““The management has definitely met my goals. I have a much larger group of engineers that feels comfortable handling the system. It saves time and it allows me to take on projects I wouldn’t have been able to.” By removing the need for specialized management skills to deploy systems of this class, DDN has democratized AI infrastructure across a far greater number of enterprises.

For Tintri customers, the impact of our AI on storage administration results in as much as 95% OpEx savings. This AI also manages VMstore’s Auto-QoS feature, which ensures that individual managed objects always have the best latency, in response to their measured behaviors and needs. Through TGC, this AI keeps tabs on things like snapshot and replication policies– ensuring that regardless of where your managed objects are stored the policies are consistent and there are no gaps in your data protection SLAs – even when objects are migrated to different systems.

These capabilities save thousands of hours per year in administrative overhead – delivering relief to our customers and freeing them to be more agile and accommodating to their respective businesses.

Hopefully, you now have a better idea about the relationship between AI and storage systems and how DDN and Tintri deliver intelligent infrastructure that strengthens this relationship and delivers impressive results to organizations across all industries. This is both true for businesses that are developing AI applications to extract and enhance the value of their data and for those that want to leverage integrated AI to simplify storage management and eliminate administrative complexity and cost. In both cases, it’s Intelligent Infrastructure delivering exceptional results and an exceptional experience.

 

Get more info about AI and Intelligent Infrastructure and have fun while getting up to speed at the Tintri Laugh & Learn Event on March 25th, featuring comedian Marty Simpson. The event will be recorded so you can watch this upbeat, virtual discussion anytime.

This site is registered on wpml.org as a development site. Switch to a production site key to remove this banner.