With their ability to provide video storage cheaper and more flexibly than on-premises systems, cloud-based video archiving services have established themselves in the broadcast TV and streaming video markets. Their competitive edge is now being sharpened by the integration of artificial intelligence/machine learning (AI/ML) to cloud-based archiving systems—making these systems more accessible to users than ever before.
The AI/ML Advantage
The simplest way to understand AI/ML’s positive impact on cloud-based video archiving is this: AI/ML helps these platforms do everything they do better.
“Machine learning in the cloud has been a huge step forward in expanding its range of capabilities,” said Sam Peterson, COO of Bitcentral. “Speech-to-text and language translation have been two of the big advances lately, but these are only the tip of the iceberg.”
“Metadata AI and Generative AI are all the rage since ChatGPT opened eyes to what can be achieved,” agreed Jonathan Morgan; senior vice president of product and technology at Perifery, a division of DataCore Software. “Like a glove fits a hand, AI is built on data, and when it comes to media archives there is both a wealth of information and monetizable assets.”
Newsbridge’s cloud-based platform manages both archived and live video content in highly sophisticated ways, according to James Fraser, vice president of U.S. Sales for the company. Newsbridge’s MXT-1 is “our own in-house AI that we’ve built over a number of years,” he said. “The idea behind MXT-1 is to produce a human-like description of what AI is seeing in your content, to make it highly searchable. So rather than search for a recorded conference using a file name, you can ask MXT-1 to find a specific moment.”
Meanwhile, customers moving to AWS are exploring the company’s full portfolio of AI/ML products, said Josh Mello, AWS’ Senior Worldwide Storage Specialist. “One helpful offering is the Amazon Media2Cloud solution, which provides a framework for enhancing metadata collection media archives via AI/ML services, as these assets are placed in Amazon S3 Glacier,” he explained.
“Additionally, many customers are pulling out insights on their consumers, as well as data on their content, using Amazon Rekognition. Amazon Transcribe is another AI service our customers are turning to in order to easily transcribe information pulled out of their media archive.”
Pros and Cons
There is no doubt that cloud-based archiving has caught on with TV broadcasters and streamers alike in recent years, even without the advantages being offered by AI/ML.
Cost is a major reason why: It is far cheaper to store content on third-party web-connected servers on a fee-for-service basis — which is what the cloud actually is — rather than invest in on-premise server farms and the IT staff to maintain them.
Then there’s data security: “In addition to offering 99.999999999% (11 nines) of data durability in every AWS Region, Amazon S3 Glacier storage classes simplify archive management,” said Mello. “Once a customer migrates to AWS, that’s it. They no longer need to worry about end-of-life support on existing infrastructure or get bogged down in the endless cycle of generational updates.”
Better yet, cloud storage providers like Newsbridge offer tiered levels of service for their customers, with instant retrieval cloud storage offering faster and more expensive access to customer content than “cold” or “deep archive.” This choice of options is made available for “cost optimization,” according to Fraser, and is proving popular with Newsbridge’s broadcast and streaming clients.
“Another significant feature is the capability to store content off-site, providing a safeguard against on-site hardware failures or unexpected incidents,” Bitcentral’s Peterson said. “In the cloud, users can leverage multiple layers of storage and implement replication across different physical regions for added resilience.”
Best yet for M&E customers, AWS’ cloud-based archiving opens up the art of the possible for more interactive viewer engagement, new revenue streams through opportunities like FAST channels, plus “additional monetization efforts that leverage the vast amount of content available in customer archives,” said Mello.
Hybrid Approach
Still, the cloud isn’t a “one-size-fits-all” solution, Peterson readily admitted. “This is why some of our customers are fully cloud-native. But for other customers, it makes more business sense to employ a hybrid approach that simply adds to the capabilities of their existing physical storage.”
As well, some of the apparent advantages of cloud-based archiving can turn out to be disadvantages in certain circumstances. For example, when a broadcaster/streamer decides to use a large number of AI/ML apps to manage their cloud-stored media, there can be “large associated compute and access costs related to running those apps.” said Morgan. “Alternatively, the AI/ML apps required might be available on-premises and are likely to be significantly less expensive and quicker to run there.”
In the same vein, on-premises storage can prove to be a more economical option than cloud-based storage in some cases. “For instance, if a media archive can be stored in a ‘cold’ public tier and is rarely accessed, then public cloud might be a cost-effective way to go,” Morgan said. “But if the media archive is actively accessed, then public cloud costs can quickly mount.”
This being said, the many advantages offered by cloud-based archiving is ensuring its viability in the marketplace. This is why “the broadcasting and streaming industries are now far beyond the point where the cloud is treated merely as a byword for storage,” concluded Peterson.
