Wasabi Technologies' latest object storage product adds automatic tagging of video data using AI and machine learning capabilities for the media and entertainment industry.
Wasabi's AiR service is based on technology created by Curio AI, which Wasabi acquired from GrayMeta Inc. in January. AiR allows customers to store video data in S3-compatible object storage with automatic data tagging and metadata indexing. Tags are deployed in seconds and provide facial recognition, speech-to-text, translation, and logo recognition capabilities.
AiR, like Wasabi's other storage products, has no inbound or outbound fees, and there are no charges for using the service. Instead, it will be offered at an initial flat rate of $12.99 per terabyte per month, according to a Wasabi spokesperson. According to the vendor, AiR is expected to be released in the coming months, and pricing is likely to change after launch.
According to Ray Lucchesi, founder and president of Silverton Consulting, sifting and tagging relevant video data is a painstaking and underserved task, even when automated. It's business.
“Video is ubiquitous and there is a huge need at this point. [a cataloging service]”It's the same with audio,” Lucchesi said.
Frame-by-frame data playback
Wasabi AiR will initially target media and entertainment customers, according to Aaron Edel, Wasabi's senior vice president of AI. Edel previously served as President and CEO of GrayMeta.
He says these customers are starting to use cloud object storage as a new form of cloud storage without the physical concerns and limitations associated with tape, but they aren't always taking advantage of the computing capabilities the cloud brings. .
AiR aims to change this by eliminating the need for customers to connect object storage to external cloud services for tagging capabilities. Customers may typically use physical tape media to store infrequently accessed large files such as video because it's cost-effective, Edel said, but object storage can offer similar savings. It is possible to connect to more services.
“We don't want customers to think of this as a cold archive that will never be accessed,” Eder said.
When you upload a video file to the service, the technology performs an initial frame-by-frame search and tagging operation. Users can later review this data and adjust the tagged information, or add additional search terms or objects they recognize for later upload. Customers can also train AI solely on data for features such as content moderation.
Wasabi already has sponsorship deals with media and entertainment companies that can take advantage of this feature, including sports teams such as the Boston Red Sox, Edel said.
AiR old footage
Steve McDowell, founder and principal analyst at NAND Research, said the hype surrounding generative AI is creating legal and technical challenges for companies. Worse, he said, GenAI has undermined non-generative AI and machine learning (ML) capabilities that are beneficial to enterprises, such as automating the tagging of video data.
steve mcdowellFounder and Chief Analyst, NAND Research
Tagging and metadata generation is not something many employees want to tackle, regardless of what business utility they derive from their data. Most metadata tagging services can handle static documents and images, McDowell added. Some products, such as Komprise and its platform, focus on tagging from a data management and classification perspective rather than examining the content itself.
“This is a great use of AI and ML,” McDowell said. “Data Metadata and Tagging” [are] This is a serious problem. The fact that it's on a storage platform is a big differentiator for Wasabi. ”
McDowell and Lucchesi said customers using hyperscalers such as AWS can combine similar capabilities. However, these features may have associated upstream and downstream charges, which can increase costs.
Lucchesi said demand will determine whether the service can maintain performance for media customers, who typically handle large video files.
“The question is what kind of metadata is being created. Are they good enough for what the customer wants to use?” he said. “If a customer starts streaming his 8K, will they be able to sustain it? [resolution] video? ”
Tim McCarthy is a TechTarget editorial news writer covering cloud and data storage.