The AI Platformization of Storage – The Data Intelligence Platform

The IT industry uses the word “platform” all the time. Often, I find myself shifting between the many jargons circling “platform”, loosely. I am pretty sure many others are doing so as well.

I finally found the word “platformization” giving right vibes in a meaningful way in February last year, when Palo Alto Networks pivoted to platformization. Their stock tumbled that day. Despite the ambiguous definition “platformization” when Palo Alto Networks (PANW) mentioned it, I understood their strategy.

Defence-in-Depth in cybersecurity wasn’t exactly working for many organizations. Cybersecurity point solutions peppered the landscape. There were so many leaks and gaps. Platformization, from the PANW‘s point-of-view, is the reverse C&C (command & control), if you know the cybersecurity speak. PANW wants to take charge all the way for all things cybersecurity, and it made sense to me from a data perspective.

Paradigm shift for Data. 

For the longest time, networked storage technology has been about data sharing, be it blocks, files or objects. The data from these protocols is delivered over the network, mostly over Fibre Channel and/or Ethernet (although I remembered implementing NFS over Asynchronous Transfer Mode at Sarawak Shell in East Malaysia), in a client-server fashion.

By late 2000s onwards, unified storage or multi-protocol storage (where the storage array is able to served all 3 SAN, NAS and S3 services) was all the rage. All the prominent enterprise storage vendors had a solution or two in their solutions portfolio. I started viewing networked storage as a Data Services Platform which I started explaining it in 2017. Within the data services platform, various features revolve around my A.P.P.A.R.M.S.C. framework (I crafted the initial framework in 2000, thanks to Jon Toigo‘s book – The Holy Grail of Data Management). This framework and the approach I used for my consulting and analyst work worked well and is still relevant, even after 25 years.

But AI is changing the data landscape. AI is changing the way data is consumed and processed through the networks between the compute layer and the storage layer. It is indeed, for me, a paradigm shift of data, and the storage layer, better known as AI Data Infrastructure now, is shifting as well. And this shift will accelerate the exponential growth in innovations, with AI and super-charged data leading the way.

DDN Infinia Data Intelligence Platform (screencapture from DDN Beyond Artificial webinar)

From Data to Knowledged Data

Artificial Intelligence is driving the paradigm shift. Storage for AI or AI Data Infrastructure plays a vital role in the innovations of AI. While the storage services protocols deliver data to the AI compute nodes, the data is raw from an AI perspective.

However, the rise of object storage has given for the data to be labelled, tagged, imbued with vector embeddings information, and giving this raw data residing in the shared storage meaning, reasoning, context and more. We are on the cusp of having Knowledged Data in the palm of our hands. AI applications are riding that rocketship now.

I once wrote “The Future is Intelligent Objects” way back in 2011. I am beginning to see this happening right now, thanks to the realization of this vision by DDN‘s Infinia Data Intelligence Platform.

Welcoming DDN Data Intelligence Platform

The data services platform that I mentioned earlier still rely on a layer of data delivery services. This is underpinned by the multi-protocols of Fibre Channel, iSCSI, NFS, SMB, AFP(?), S/FTP, WebDAV and the darling right now, S3.

DDN Infinia 2.0 was launched last week in the Beyond Artificial worldwide webinar (full video here and below). DDN Infinia (from the picture above) still has the usual data protocols, that include S3, block-level integration through Cinder and CSI drivers, file-level protocols and SQL (yes, Infinia has a SQL Query Engine) for data management and data access. Every data infrastructure vendor is trying to complete the checkboxes of unified network storage protocols.

The thing that got me super excited about DDN Infinia is the rectangle box on the top right of the photo above that says “AI Data Acceleration Libraries“. This is the part where DDN is creating a future for putting knowledged data from the Infinia Data Intelligence right into the AI applications, the AI framworks and the AI models. That is an incredibly powerful reality that is happening right now.

Listen to Jensen Huang, the CEO of NVIDIA®, explain exactly this reality in his conversation with DDN’s CEO, Alex Bouzari in the video below. As quoted:

“During use, the AI has to access information, and AI would like to access information not in raw data form, but in information form and this is the reason why the reframing of storage, of objects and raw data into Data Intelligence, … and providing data intelligence for all of the world’s enterprise as AI runs on top of this new Fabric of Information …

Jensen Huang (CEO, NVIDIA®) in conversation with Alex Bouzari (CEO, DDN)

Watch that part of the video here between minute 46.30 to 47.25.

DDN Infinia is software-defined and Infinia 2.0 acts as a unified, intelligent data layer that dynamically optimizes AI workflows.

AI Inference efficiency and accelerating innovations in AI

The world is beginning to shift from AI data training to AI inferencing. Many of the trained data are being put into production, often through inferencing. And the knowledged data of information I mentioned will bring forth optimization of data access and data management, supercharging the AI data pipelines and accelerate AI applications data processing at an unprecedented speed and scale, whilst maintaining security and efficiency. DDN Infinia is built from ground up to deliver these tenets for AI execution engines, AI inferencing and AI applications.

DDN AI Data Intelligence AI Framework

The diagram above explains how the DDN Infinia is redefining the AI data architecture landscape. The AI data acceleration libraries brings a new level of security, efficiency, performance (both in terms of very low latency, and extreme high throughput) for AI everywhere, at scale.

From DevOps to Full Stack

As AI transforms the applications landscape, we are also seeing the change happening from the developer end as well. We once were trying to integrate between the software developers and the infrastructure backend. DevOps tried to happen, in a rather fortuitous way, and was fractious at best. There is still a noticable divide between developers and infrastructure people.

The traditional software development lifecycle and data pipelines are increasingly being transformed by AI. Data intelligence is now automating development, from data movement, and data management to data security and data governance. These tasks once have to be coded into the software, but AI with knowledged data is augmenting these tedious and infrastructure tasks with greater automation, dynamic metadata tagging, smarter search and overall data efficiency. Smart developer tools and intelligent AI software agents are transforming DevOps into Full Stack.

This already pervasive change liberates the software developers, the AI application developers to better data movements, design better data engineering and much more, having AI supercharging AI. AI is leading the way.

Time for AI-Ready Data. For Real.

In the same way about the platformization strategy taken by PANW, DDN is taking storage platformization to take on the AI phenomenon, with NVIDIA® leading all the way. DDN has taken a leaf from NVIDIA®’s playbook to deliver the full stack of enterprise AI applications, for data. It is bringing a whole world data, knowledged data, with a whole new level of data management, data movement efficiencies, just for AI.

Thus the platformization journey I am talking about is where we are seeing and learning in real time, that storage or data infrastructure is no longer a reactive server to clients’ request for raw data but an active participant in the AI data management, data acceleration and data processing. This is what I meant when the DDN platformization first step is the Data Intelligence Platform.

AI is driving data towards data intelligence, and data intelligence is driving toward better AI. It is time for AI-Ready Data, and this new Data Intelligence Platform from DDN is leading the way to changing the storage technology landscape for the present and the future.

Tagged , , , , , , , , , , , , , , , , , . Bookmark the permalink.

About cfheoh

I am a technology blogger with 30 years of IT experience. I write heavily on technologies related to storage networking and data management because those are my areas of interest and expertise. I introduce technologies with the objectives to get readers to know the facts and use that knowledge to cut through the marketing hypes, FUD (fear, uncertainty and doubt) and other fancy stuff. Only then, there will be progress. I am involved in SNIA (Storage Networking Industry Association) and between 2013-2015, I was SNIA South Asia & SNIA Malaysia non-voting representation to SNIA Technical Council. I currently employed at iXsystems as their General Manager for Asia Pacific Japan.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.