[ad_1]
Flash storage provider NetApp and real-time data science platform developer Iguazio said Monday that they have teamed up to offer channel partners and business customers a joint offering aimed at improving the efficiency of intelligence applications. artificial.
The joint offering aims to build a high-speed pipeline between data storage and analytics to improve AI performance and improve governance around applications, said Asaf Somekh, CEO of the New York company e Israel.
“At the company, there is a lot of hype around AI, but a lot of inefficiencies,” Somekh told CRN. “You see data scientists building models and showing efficiencies in the lab getting the green light to launch products. But 85 percent of successful lab projects never see the light of day, according to Gartner. We’ve partnered with NetApp to change the way AI enters the business. ”
[Related: NetApp Introduces Keystone: Simplifying Storage For Hybrid, Multi-Cloud World]
Joint AI platform combines NetApp storage and its Ontap AI and NetApp Cloud Volumes software for hybrid cloud capabilities, Iguazio’s data science platform for end-to-end machine learning pipeline automation, and Nvidia DGX GPUs that can be used on a service basis.
It is closely related to a new trend, machine learning operations, or MLops, aimed at putting machine learning into practice, Somekh said.
“Many ‘unicorns’ have built technology in the first step of AI, which is the laboratory,” he said. “Our Ignazio data science platform is about taking lab work and spreading it across the business. It runs in real time to collect the different pieces of live, real-time and historical data data in storage to make the process “.
Hoseb Dermanlian, senior AI manager who entered the market at NetApp, Sunnyvale, California, told CRN that the joint offering is largely a channel game for NetApp.
NetApp has been building an Nvidia-NetApp channel for the past year, but has not been selling it directly to companies, Dermanlian said. The company already has five channel partners in the United States working with NetApp’s AI-centric offerings, he said.
“We will grow our partner base,” he said. “We will not boil the ocean, but we will build the base.”
Creating a high-performance data channel for machine intelligence and machine learning has been a challenge for data scientists and architects, said Juan Orlandini, chief architect at Insight, a Tempe, Arizona-based solution provider and partner at NetApp. and Iguazio. .
“It has been done for the cloud, but it is more difficult for installations,” Orlandini told CRN. “Everyone thinks data scientists are doing difficult things, and data science is difficult. They need doctorates. But the other difficult part is getting the data from disparate services.”
Orlandini said this is due to the need to normalize data from multiple sources and make everything look the same before an experiment can be run on the data.
“Then you will want to run another experiment on a different portion of the data, or use a different model, and then you will need to put it into production and take it to the field,” he said. “Having a pipeline so the data architect can do his job and the data scientist can do his job is key. Iguazio manages the pipeline.”
The storage part is also really difficult due to the need for performance, repeatability and efficiency, Orlandini said.
“That’s the really good thing about the two companies working together,” he said. “On the one hand, you have the data consumer, and on the other hand, the infrastructure. Data architects and data scientists don’t want to be storage experts. But in the past, they were forced to be storage and computing experts, and we were really working on two jobs, working with the data and making sure the data is available fast enough. ”
The joint offer brings together the best of both sides, Orlandini said.
“With Ignazio and NetApp, if I’m an infrastructure person, I can provide the APIs next to the software and let them do their job,” he said. “And at Insight we can do this for our clients.”
The joint offering will be sold as a meeting solution on the channel, Dermanlian said.
