 Can you hear me? Great. OK, thanks very much for coming, guys. Just to introduce myself, I'm James Norman. I'm the engineering manager at Storage Made Easy. And I'm very pleased to be here today. I've been at Storage Made Easy for coming up to three years now, and I run the engineering team here. And before that, I worked at Sony PlayStation, where I was a developer there on many of their platforms. So in this brief talk today, I'm going to go through how we can optimize our object storage to make it the best that it can be for media entertainment use cases. A lot of the things we'll cover today might be around media entertainment, but the applicability is also across other fields like medical and genomics, and also, in general, just big data as well. So what we're going to do is we're going to go through some of the challenges, what firms are facing these days with their storage, and how we can start to optimize our object storage, and what tools we can use to make best use of our object storage. So the current situation today is that we have a growing number of data feeds coming in to our storage systems, whether we're using object storage, or file storage, or cloud storage. We have more and more multimedia streams coming in. So whether that streams from TV broadcast shows in 4K, 8K, Ultra HD, whether that's feeds, raw feeds from cameras that are filming the content, for example, or whether that's gaming consoles and eSports, the media that that's generating through to surveillance, virtual reality, and of course, mobile phones. We're generating more and more data feeds these days, and we're getting not only more data feeds, but also an increased volume that's coming through that pipe as well. So just to give an example of the sort of scale of data that we're talking about, if we're looking at a situation of a chat show that might be on the TV on the weekend, we might expect them to be filming in Ultra HD, typically 8-bit RGB at about 24 frames per second. Raw UHD footage like that would generate about 33 gigabytes worth of data per minute. Now, if we multiply that up by an hour-long episode, we're starting to talk about nearly one and a half terabytes worth of single-stream footage for that episode. And of course, a single stream is not really what we're recording on a chat show these days. There's five, six, seven, eight cameras recording a particular episode. So we're talking in excess of 10 terabytes worth of data, perhaps for a particular one-hour-long episode of a chat show. That doesn't even take into account the extra data that we're generating on top of that. For example, if we're localizing that into different languages, whether we're doing audio captioning on top of that, we're generating more and more data, and we need storage that's capable of managing that. Now, typically what people have is they might have a mixture of storage, maybe some file storage, and also maybe some object storage, maybe like OpenStack Swift, for example. And people have different storage platforms for different reasons. They might have their file or block storage because it's very fast for them. They can write to it very quickly. They can read from it very quickly as well. But it can't scale. It can't handle the sheer volumes of data that we need to store these days. So they're looking towards object storage, and that's very common. So why do they pick object storage these days? Well, object storage is perfect for handling very large volumes of data. It can handle terabytes, petabytes, exabytes worth of data. And it can scale pretty much endlessly and eliminate hotspots that you might see on typical file storage. It's also perfect for unstructured data. For example, in the media space, there's an awful lot of unstructured data there. And it's very performant. And a lot of the object storage platforms provide tools and techniques for moving this large amount of data up and down between the storage platforms like DLOs, multi-part uploads, range reads, for example. There's also other benefits for object storage, like cost savings, built-in DR, and geo-deployment as well. But these firms that want to use the object storage for all of this data, what sort of requirements do they have? Well, they need storage that needs to scale and always be available. Well, OpenStack Swift can do that, for example. Zef can help you with that. So that's pretty much a good tick box on that one. We need to be able to pull data down and push it up in the quickest possible way. Like we covered on the previous slide, things like DLOs, MPUs, range reads, for example. They're very good for that. We need to be able to secure the data and enforce the security policies because all of these media assets are very important to these organizations that need to protect it. There's some support, for example, encryption at rest, which we can do on the object storage in most cases, but it's not a complete feature set there. There's also a need to access and share this data inside and outside of the organization. Again, the object storage, we can't really share it with people outside of the company without giving them access keys and secret keys and opening up the firewall ports to the storage. And of course, we need access in the crucial workflows through our dams, mams, or if we look in the research and genomic areas through the genome labs and research institutes. And finally, we need ease of use as well. This needs to be used by end users and adopted by end users, and that's not something that's particularly easy with object storage when we're dealing with users who perhaps aren't tech savvy. So storage made easy. We're a UK company and we produce a product called the Enterprise Farfabric. Now the Farfabric at its core is a virtual application that can be deployed into any virtualization platform like OpenStack, for example. And at its core, it provides connectors to 60 different storage backends. So it can talk to things like OpenStack, it can talk to Ceph, it can talk to Amazon S3, it can talk to Google Dropbox, whatever you have, it can pretty much talk to it. We were able to provide this vast amount of connectors to lots of different storage types, and that gives you a single pane of glass for you into the storage. So you can see your Amazon storage next to your local file storage, next to your Swift storage, for example. And on top of that, we're able to offer a lot of features to the users there. So for example, there's lots of benefits for the organization, such as being able to secure the data, track it, audit it, trace where users are and what they're doing on the system, for example, apply policies like encryption at rest on the storage. And for the users of the organization, it gives them primarily access, ease of access to the storage. We provide access through web, desktops, mobiles, range of different platforms they can access the storage through. It also gives them a very nice set of collaboration tools so they can work together. So if you have different editors, they can share files with one another, collaborate on different files with one another. And there's also productivity tools in there to help them be more productive in their roles as well. There's a lot that this particular diagram explains, but there's also things like protocol gateways. So we provide web-dav interfaces, FTP, SFTP interfaces, desktop applications, mobile applications, and of course at its very core, we have a RESTful API that they can connect through. So why is the file fabric great for these particular use cases? Well, firstly, it's integrated into every point of access. So it can go into the web, there's no plug-ins required, there's a simple web file manager, desktops, it can integrate directly into your Windows file shares. You can see it as a network share on your PC. There's dedicated applications as we talked about on iOS and Android, legacy protocol gateways, supporting FTP, SFTP, and web-dav. So if you have in your genomics lab a microscope, it can talk to the FTP server. And this all means it integrates into your tool set. Just an example here, we have, if you can see it, this is a Windows desktop connected to a range of different storage platforms. In our case, we're interested in the Swift storage, for example. This is object storage, but we can drill into it as files and folders, open up particular folders, double-click on files, and instantly see videos and download content from the object storage. It's very end-user-friendly, and that's the key here. There's also no more command-line access required as well. People who are using object storage are probably very familiar with command-line, but if you wanna grow the object storage out to the rest of the organization, they can't use the command-line. So one of the really neat things about the file fabric itself is it provides the, touched on the end-user view into the storage, gives them very easy ways to access the storage. You can do away with the command-line, and you can also get a lot more benefits by moving to the file fabric. Data movement is also really important for people who are dealing with large data sets in particular. Let's say, for example, you're working with your colleague who's in a different country, and you have a very large, maybe a movie file or a genome piece of data, and they need it there on their systems by the morning. With the file fabric, you can simply do a drag-and-drop movement of that data from one system to the other, and it can be there, you can walk away, leave it, and you don't have to monitor it, and when the user wakes up the next morning, they can access the files. So that uses our Mstream technology that we've been developing on for quite a while now, but it doesn't use anything proprietary. It uses a lot of the underlying technology provided by the storage platforms like DLOs, MPUs, Range Reads, to paralyze the uploads and the downloads and the transfers of these files between the different locations. So there's little or no vendor lock in there. And what it helps accelerate is things like cross-cloud transfers, transfers from the end-user, from your desktop up to the storage, and also likewise from the storage back down to your desktop as well. So just an example, here we have a very quick example of the drag-and-drop which you can perform from your desktop up to your object storage. So here I'm dragging from my local Mac into my Swift storage, no browser plug-ins, nothing proprietary there, just works in any web browser, and this is now doing multi-streamed upload from the web browser, and also multi-part uploads to the storage on the backend. So it goes pretty fast and it pretty much minimizes the effect of latency across the network there. And when this is going on between two different servers, it's pretty much using the full data pipe that is available between the two storage locations. There's also the cross-cloud transfer as well. So for example, if you have some camera footage, perhaps from a chat show episode, you wanna take it into the editing suite, maybe they have the file systems in the editing suite. Again, it's a very simple drag-and-drop between the two platforms. Here we have our Swift storage. What we're gonna do is we're gonna drag a 10-gigabyte file from one to the other. We're gonna drag it over to our file system. And now that transfer will just kick off in the background. The user can close the browser, walk away, and be assured that when they come back, they'll see the file in the directory. So you can see that in this case, it's using 100 megabytes per second data transfer there, which is pretty much the upper limit of the pipe between the two storage systems. Object previews as well are pretty important for customers. When dealing with large files, they don't wanna have to download the whole file or the whole object in order to see whether it's the one that they're looking for. So the file fabric itself provides previews for lots of different file types. Some of those are particularly handy, for example, media files. Media files can be very large, terabytes in size. So with the file fabric, they can simply initiate previews of those files so they can start to stream those files without needing to download a one terabyte-sized file to see what's actually in it. Again, this is applicable for other use cases as well, for example, around Dicoms and genome files as well. And of course, there's some generic viewers in there as well for things like Word, Excel, PowerPoint, and many of the end user-friendly tools that people are using. So just an example, again, here we're dealing with a 150 megabyte MP4. We can just open it up, have a look at the preview, click the preview, no need to download the full file, starting to stream straight from the object storage, and then we can start to scrub through the video to see different points and see whether this is the actual video file that we're looking for. Content distribution is also really important for a lot of organizations. When you're sharing data with people outside of the company, quite often there's some shadow IT going on. They want to, maybe they're using some insecure FTP servers to transfer files to other parties, or maybe they're using unsecured Dropbox accounts, so whatever they're using, there's often some insecurity there in terms of how they're transferring the data from one to the other. It's also a case of replication from one storage system to the other. However, with the file fabric, what they can do is they can replace those maybe the existing FTP servers and replace those with the file fabric's secure sharing capabilities where everything is logged. As soon as someone accesses that shared link, it's all logged. They can do both inbound, so they can receive files and outbound sharing, and they can also individually protect files, put time-based expires on files, and get real-time reporting of access as well. So just a quick example here. Again, we have our Big Buck Bunny movie. What we can do is we can go and start to share this file by URL. We can set expirations on it, limits on downloads, and also passwords. What we can then do is we can then send this across to our colleagues by email if we want, or we can send them the link, or we can send it via SMS if we want to. And so in this case, we can generate the shared link. And then on the next slide, what we'll do is we'll present that we are the user receiving the link, and now accessing the link, they need to enter the password. Again, if they access this after three days, then the file would no longer be available, and then they can now start to download the file. So now they've achieved the secure distribution of files, and there's also no need to replicate that file into lots of different storage platforms as well. It's also storage agnostic as well. So because we support a very broad range of clouds, it doesn't really matter what we're talking to, we can provide all of this functionality for all the different clouds. And also the file fabric doesn't write in any proprietary formats. So if you're using Swift or Cep, for example, or whatever your storage system is, you can still continue to access the storage directly, as well as using the storage through the file fabric as well. We don't write in any proprietary formats, we use all of the underlying technology like I said before, around that. There's also some built-in intelligence into the platform as well, where the product can integrate into platforms like Google Vision, Amazon Media Services, and Elemental, to give you things like automatic tagging and classifications, automatic transcriptions, and all of this is searchable at your fingertips as well. So just to recap again on where we were, what do the big firms need? Some of the key points are sharing the data. We can now do that. We don't have to duplicate the data off. We can now set policies. We can use storage encryption and also enforce policies through the file fabric. We can also integrate this into our workflows, through our Windows tools, for example, if we wanted to do it through those platforms. And of course, the key one is of use for end users as well. We can see how easy that is to interact with the storage for all the different platforms that we support. So thank you, thank you for your time. If you have any questions, feel free to ask. But we're also over on stand C17, which is just on the back corner of this room.