Specialized services in public clouds – the value for active media workflows
The Media & Entertainment vertical has been heavily influenced by public cloud. From the early days of using cloud storage as a long-term, inexpensive data archive, media companies are now seeing the primary value of cloud services to enable active media data workflows. Cloud services are evolving more toward specialized services – those that can provide value through processing of the data, versus just simple storage.
Many people have considered the potential value of on-demand compute services such as AWS EC2 to perform compute-intensive jobs. For media, this would naturally lend itself to workflows such as file transcoding, and potentially for rendering. In the last years, we have now seen the cloud vendors introducing yet more specialized services for media data, such as the Azure Video Indexer and video encoding in Google Cloud Platform. There are now Content Distribution Network (CDN) services in all the major clouds through AWS CloudFront, Azure CDN and Google CDN. These services now make it possible to consider media workflows being partly or mostly in the cloud, as opposed to the captive solutions that have been employed to date.
Cloud Media Workflows
Media production and post-production is a complex business, comprised of multiple workflows that have all become digital. Video payloads are also becoming immense – with multi-Gigabyte and even Terabyte files becoming commonplace. A big cost consideration, therefore, is related to minimizing data egress charges for massive media payloads – especially the impact of data transfer charges from multiple public cloud services.
In today’s world where content is delivered through multiple channels, on different devices (cable boxes, desktop, mobile), in different resolutions, and in dozens of languages – the transcoding effort can now entail over a dozen ultimate target file formats. The end stage of the workflow is the delivery of the content to the end-user devices. Services such as Akamai and Limelight are well known as popular Content Delivery Networks (CDNs) to provide both the video storage repository (in common terms, the “origin server” for the video content) and perhaps more importantly the “edge servers” which typically provide local edge caching and HTTP-based delivery of the content to the consumer. In recent years, customers have adopted on-premises object storage solutions such as our own Scality RING, as the best TCO storage solution for their own video origin servers. Now we see adoption of multiple cloud services for extending media workflow to transcoding and CDN.
Why do I need more than one cloud to do this?
If all that is needed is a cloud-based CDN service, isn’t a single cloud sufficient to enable a cloud media workflow? As mentioned before, specialized media services in the cloud are rapidly emerging, but in a non-uniform manner. Today’s compelling cloud service that offers distinct value for certain applications will almost certainly be challenged by others in the future.
AWS is incredibly popular, and one reason for this is that they now offer over 90 cloud services, including a specific set of Media Services through AWS Elemental. In addition, there are offerings from cloud vendors such as Google, Microsoft Azure, Alibaba, IBM and others that can further help optimize and refine TCO. Differences in the various cloud offerings by region, by service and even cost can make a multi-cloud deployment the best approach for both the media workflow itself, as well as for the best TCO. Retaining flexibility to choose the best services today versus tomorrow is the insurance policy that customers in media will need to ensure they reap the best TCO.
Optimizing for cost-effective cloud media workflows through Zenko
Media customers understand that being selective in their usage of cloud is the greatest lever they have for reducing their TCO, and by intelligently placing data to minimize data transfer charges, especially for unnecessary egress from clouds.
Zenko can immediately reduce TCO by simultaneously transfering files to multiple clouds, thereby taking advantage of free cloud data ingress. This can eliminate data egress charges from one cloud to copy the data to another cloud service. Zenko also makes it possible to be very selective about exactly the files that are needed in the cloud, either by specifying policies that filter those that are needed, or by providing automated lifecycle management policies for expiring files in the cloud, once they are no longer relevant. These optimizations alone can help save large media customers significant amounts of money on their monthly cloud services charges, while also streamlining their media workflows.