NAB 2019 Reflections: Cloudian’s Jon Toor on New Telestream Vantage Integration, Power of Object Storage
Combined technologies offer advanced intelligence for search and analysis of media archive
At NAB 2019, Cloudian and Telestream combined their Cloudian HyperStore and Telestream Vantage technologies to provide advanced intelligence for search and analysis of media files by using automated enriched metadata tagging. Under the joint offering, Telestream Vantage creates a media-aware overlay on the Cloudian HyperStore object-storage platform. Vantage processes and analyzes media files, providing such services as auto-transcription and captioning, keyword analysis, automation of ingest workflow and asset QC, and active-library-management functions. It also automatically creates enriched metadata that is stored with the associated content in HyperStore. In addition, Cloudian announced availability of a Cisco Validated Design (CVD) for Cloudian HyperStore on the Cisco Unified Computing System (Cisco UCS) platform, accelerating the transition to hybrid cloud.
SVG sat down with Cloudian CMO Jon Toor to discuss the opportunities created by the combination of HyperStore and Vantage, the increasing use of hybrid cloud/on-premises and multicloud environments, and why he believes the M&E industry is finally embracing object storage.
What is the Telestream-Cloudian integration being demonstrated here, and what functionalities does it open up for users?
We think this partnership is really a step towards realizing the true benefits of object storage. Along with Telestream, we’re leveraging three of the key aspects of object storage: scalability, embedded metadata, and geo-distribution.
First of all, object storage is scalable, which is the main attraction for a lot of people. That opens up totally new possibilities in terms of infinitely large archives of content.
The other piece of object storage is the fact that you have embedded metadata, meaning your metadata is recorded on the storage along with the media itself. Wherever that media goes, the metadata is attached forever. It can go from one Cloudian environment to another or to the cloud. You have the ability to search and find the media at all times regardless of its [location].
What Telestream is enabling is the ability to enrich this metadata using Telestream tools that would provide all kinds of image descriptions, voice transcriptions, ways of describing what’s in your media, such as pulling out keywords that allow you to later search and find that media much faster. Frankly, one of the biggest problems we see in the industry is that [organizations] have so much media to work with but can’t find what they’re looking for [that] it becomes of zero value to them. So the ability to enrich that metadata and use Cloudian’s embedded search tools to find that media increases its value exponentially.
The other thing we’re enabling with Telestream is geo-distribution for remote-production environments that are linked to centralized broadcast and editing facilities. This can help sports producers reduce the cost of sports broadcasting because they can send smaller teams to remote sites. You’re collecting media at that remote site, and you’re using that media and presumably archiving it at a centralized site.
How do you connect those two things together? Cloudian provides an ability to scale a single object-storage environment across multiple locations. The way object storage works, the remote site and that central site act like one device. You have one view of data, one place where you can go get it; you can actually have the media at a remote site and at a central site — both locally accessible, which just accelerates everyone’s workflows.
Are you seeing more deployments of hybrid on-premises/cloud storage workflows these days, rather than companies storing all content onsite or looking to move exclusively to the cloud?
Definitely. The trick to a hybrid storage environment where you’re using cloud and on-prem together is the link. How are those two environments communicating? If you are managing both as totally separate environments, it’s going to be difficult to manage. Cloudian has that integration built in natively because we were built on the [Amazon] S3 API from day one. There’s no translation layer, no access layer. It’s S3 API right down to the core, so it looks and feels exactly like you’re talking to Amazon.
Since we had that A3 API built in from the beginning, it’s very easy for us to treat the public cloud and your on-prem presence as one environment because they both speak the same language. We’ve got integrated tools that can manage the flow of data between those two locations. You’ve got a single view of data. And you’ve still got the embedded metadata, so you can search across your entire hybrid environment with one query. You don’t have to go to the cloud and search once and then go on-prem and search again; You search one time, and you get data from all those environments.
You have the option to either tier data to the cloud and move it entirely off-prem, or you can replicate in both locations. If you’re looking to protect your content for [disaster recovery] or metadata enrichment in the cloud, then you would replicate the data. But, if you’re looking to just do content distribution or archiving, then you would just move everything off your local storage and put it in the cloud on something like [Amazon S3] Glacier. You can still see it as part of your storage environment, but it’s no longer consuming on-prem capacity. Depending on what you’re doing, you can have tiering and replication, and both are being managed from within Cloudian in one seamless environment.
Do you believe the M&E industry — sports organizations, specifically — is finally opening up to the concept of using object-based storage on a larger scale?
Absolutely. They are becoming aware of the value of the immediacy of object storage because this is a limitlessly scalable pool that you can access in milliseconds. Especially in the sports world, where time is everything, that’s huge. We are either having discussion with or deploying [our systems] at virtually all of the major broadcasters right now.
Object storage is limitlessly scalable and searchable, so you can keep the largest possible archive of historical media. You can pull up things from the history and find them quickly through Google-like search capability. Since you can enrich the metadata, you now have a way of not only finding media but also continuously updating how you’re finding media. You can continuously improve your process, enrich that metadata, and you’re not locked into any specific hierarchy or taxonomy.
The hierarchical way of storing and retrieving media dates back to the beginning of storage. But the internet is different; it is random access. The internet is flat, and that’s what object storage is. You can now find things on object storage just like you’d find them on the internet, which I think is very appealing to the [M&E] industry.