Video: How Storage Must Change to Meet Mushrooming Video Growth
Watch Naz Nageer's full presentation from Streaming Media West, Leveraging IBM Cloud Object Storage to Improve Streaming CDN Performance and Cost Efficiency, on the Streaming Media Conference Video Portal.
Read the complete transcript of this clip:
Naz Nageer: Traditionally, we come from an environment where if you're a storage person, you have tier one-, tier two-, tier three-type storage. It's really changed. There are two types of storage, structured data and unstructured data. The big hockey curve is driven by this unstructured data, smart phones, Internet of Things, telemetrics, audio, video, text, blogs. That's all driving the growth. We're seeing a 60-80% year-over-year growth.
Can our existing storage infrastructures support that? They say there will be 30X growth in this by 2020, with 4K videos everywhere. Every one second, one million minutes of video traverse the internet. That's all driven by our smart devices. So I think we're under a challenge from storage infrastructure to have a strategy to be able to manage this growth.
So video resolution has a direct impact on file asset size. You're looking at 4K, 5K, 6K. What's next? All of these devices are being able to support this kind of resolution. Here's a deeper look at how 4K content is driving archive storage. This is just an impact on storage infrastructure. You can't keep this level of growth in traditional storage systems.
Traditionally, how do we manage this kind of growth? File-based storage. Hierarchical storage, directories, directory file storage ... it sort of mimics human behavior. We're used to folders and files and folders and files. That has a breakdown when we start scaling. You can't scale to the mushrooming growth levels that we're seeing. Data is exploding, and yes, you can continue to do it that way, but it doesn't make sense. If I were to analyze your existing storage infrastructure and take a look at the content you're storing there, the majority of it that's on file storage, expensive spinning disk hasn't been accessed in the last 60 to 90 days. Why would you continue to do that?
That kind of storage infrastructure is really designed for a LAN, right? And there's a point where the number of files and directories of a single named space starts to affect performance, and you can spin up more storage. But it becomes a management nightmare to support that kind of growth. And that's where the economics come into play.
So with object storage, you have a flat name space, right? Key value, flat name space, very easy to grow, very scalable, simple to access protocols, it uses our S API, the predominant ones out there are S3 open stack. Cleversafe had their own proprietary one. We are a proponent of S3 going forward.
So for the next generation, it must scale, it must be more efficient, must always be available. We're looking at geographic kind of accessibility with the world becoming larger. We're all one economy these days. We must be able to have data integrity and security, and when we look at our legacy systems, they can't scale to that level, and when you're looking at protecting that data, they're relying on very old technologies that take a very long time to recover. RAID 6, however you want to rebuild a 20 TB drive, whether it's a 74K or a 15K RPM, it takes days to recover that. You need an infrastructure that gets rid of that. So our solution--a lot of the object solutions--are the way you want to be able to do that.
Bruce Ross of IBM Cloud Object Storage discusses the skyrocketing storage requirements brought on by the advent of 4K and HDR video, and the challenges of object storage in the cloud.