Asia's Source for Enterprise Network Knowledge

Sunday, June 25th, 2017

Storage

AI and machine learning to play a big part in Pure Storage's future

AI and machine learning to play a big part in Pure Storage’s future

Artificial Intelligence (AI) and machine learning will be one of the cornerstones for the continued growth of Pure Storage learned attendees at Pure Accelerate 2017.

Pure Storage CEO Scott Dietzen said during his keynote that AI is one of the top priorities for today's IT professionals.

"In IT priorities today AI ranks only behind cloud and mobile in terms of the impact people perceive it's going to have. Data-driven prediction, including deep learning, demands much more of the underlying data platform. It's got to be simple and cost effective to scale really big data and have high enough bandwidth to capture video and sensor feeds and the performance necessary to run next-generation applications," Dietzen said, "Mobility and the Internet of Things (IoT) are driving the explosion of data, and data-driven predicting, including deep learning, is giving us new tools to mine the value therein. It's an immense opportunity to improve our current businesses and build wholly new ones."

Matt Kixmoeller, Vice President of Product at Pure Storage said that thanks to advances in machine learning algorithms, IT organizations are about to witness a new era of self-driving storage. If the value of all that data is ever to be derived, the first thing that needs to be addressed is how to efficiently store it all.

According to the company, self-driving storage requires the elimination of manual operations and must also provide a level of protection and safety to accommodate all the interacting elements. Finally, self-driving storage must also extend beyond single array optimization and deliver global intelligence.

“Big Data is becoming big intelligence for AI and machine learning but we need to process data at a larger scale and velocity. This is causing problems at the storage tier because we need to be able to process large pools of data rapidly and efficiently,” Kixmoeller said.

Kixmoeller added that rapid adoption of AI applications across multiple vertical industries in the past year has increased demand for high-capacity storage. “The more data AI applications have access to, the better the outcome,” said Kixmoeller.

Pure1 META is delivers global predictive intelligence by collecting and analyzing over 1 trillion array telemetry data points per day and enables effortless management, analytics and support.

The key driver behind the global predictive intelligence of Pure1 META, is the META AI Engine, which analyzes a data lake of more than 7 petabytes of data to generate both Issue Fingerprints and Workload DNA. META scans all incoming array telemetry against a library of issue fingerprints to predict and resolve incidents in real-time, before they impact customer environments, and captures hundreds of variables related to performance that are used to forecast performance load.

Customers can leverage META's Workload DNA through the new Workload Planner in Pure1. This new capability will allow customers to answer questions about new workload deployment, interaction, performance and capacity growth, and workload optimization, helping reduce risk, increase consolidation, and provide better visibility to plan for upgrades or expansions.

"For the first time, customers can leverage intelligent storage that constantly analyzes historical data across our global network of devices to improve their overall environment," said Kixmoeller, "Pure1 META helps customers more accurately forecast performance through growth, answer questions about capacity and performance additions, and how their environment will change over time, which allows organizations to manage future needs."

What this could also affect is analytics at the edge as the sheer volume of data generated and needing analysis is too large to moev it back to the core for analysis. Kixmoeller said that devices on the edge generate data so fast that there was a need to analyse it in place to make effective decisions very quickly.

"I believe edge will be larger than multi-cloud and core combined in terms of data because of the proliferation of IoT and sensor data," Dietzen noted.