San Mateo CA-based object storage vendor Cloudian has announced a partnership with Skymind, an artificial intelligence [AI] vendor which developed the open source deep learning library Deeplearning4j. The deal will see the two companies collaborate on integrated AI and machine learning solutions. Cloudian expects that this will make these kind of tools much more widely used than they are today – and provide a new kind of opportunity for their channel partners.
“Skymind is not a shrinkwrapped product,” said John Toor, Cloudian’s CMO. “Skymind is more of a toolkit to create spoke solutions for different industries. AI and machine learning are a tool set to be used. Where Cloudian fits in is providing the infrastructure to support solutions for those problems. What Skymind does generates a lot of data. When you train an algorithm, you need a large amount of data to do that. We have a storage solution that is well suited for this because of its scalability, metadata, and ease of use.”
Toor compared what Skymind is doing for AI as comparable to what Red Hat did for open source software, and believes it will have a similar impact.
“Skymind’s model is all about democratization of these tools, providing education and documentation and making them available to anyone,” he said. “This has huge potential. It’s a great moment for partners to learn about something they can take to their customers. It’s similar to when Red Hat made it easy to adopt open source software. I think Skymind is doing the same thing with some of these tools and I think it will have a similar impact. Over the next six months, we will see more use cases of these tools, and further democratization, as it is proved out in actual customer environments.”
“We are excited about this because when we sell our storage and data management, people ask about additional things we can do to get value from our data,” said Mike Tso, Cloudian’s CEO and co-founder.
“Skymind often runs into data management issues, and they would wind up being referred back to us by one of our partners,” Tso said. “A key part of AI is taking a subset of data and retaining the system. That gets into our model well, because our metadata makes it easy to take a subset and only train on that. In addition, Skymind was born in the cloud and do all their R&D in the AWS cloud. But they sell to hospitals and banks where training data can’t leave the premises or are too big to leave the premises, because their data take too long to copy to the cloud. Because we are 100 per cent compatible with AWS, they can move their toolset from AWS to on-prem and we can migrate everybody’s data.”
The first solution from the partnership is a security monitoring tool that identifies likely network attacks, using pattern recognition to identify suspicious network traffic, and making both the source and the destination of the attack visible. Cloudian provides the data management that supports the training environment, the algorithm development and its ongoing refinement.
“We have collaborated around AI solutions before,” Tso said. “We developed a deep learning-based system with Intel and Dentsu [Japan’s largest advertising agency]. It allowed the identification of the year, make and model of cars as they approached a billboard, so the advertisements could target specific demographics.”
Toor said that the security monitoring tool will be the first of many joint offerings.
“There will be a whole range of solutions addressing different marketplaces,” he said. “There are 10,000 different tools out there for analyzing genomes. This will be similar, in terms of the broadness of the potential market. Cloudian’s role here is to sell the shovels and tools that enable the gold rush.”
The opportunity for partners to cash in is also significant.
“AI is rapidly evolving, but today it is a very consultative sale – not like our own platform which is turnkey,” Tso said. “Over time I think Skymind wants us to help them better leverage our sales channel, to make it more repeatable and easy to sell. From our channel partners’ standpoint, they want tested solutions that are easy for partners to offer. This AI story is another angle to begin conversations for them.
“Our partners know how to handle the data migration, and help the end user bring out value from the logs,” Tso continued. “I see this as a solution that brings additional benefit to the users, and thus to the partners who can be very involved in enabling it.”
Cloudian is likely to develop similar relationships with other AI vendors in the future.
“We are a standards-based solution that is easily put into environments,” Toor said. “I would see us working with a variety of players over time.”