Cloud services have been around for a fair while now, in various forms and from a multitude of providers. As a platform, it has now matured to be a very stable, established environment in which to store, process and serve data in a massively more efficient way than previous methods. Much of the web is supported by a backbone of this largely transparent service, including Google searches and applications, YouTube video, Facebook photos and many more. Here at MultichannelTV, we leverage a particularly large Cloud provider to support our vastly growing archives of videos, in addition to much of our less dynamic (but otherwise essential) content. But why do we use it? And, more importantly, how does it work?
We'll tackle that last question first, and explain what the "Cloud" actually is. The answer to that is surprisingly simple, and frustratingly vague;
The Cloud comprises of thousands of computers which are all controlled by various software packages all tailored to manipulate the computers in very specific ways.
As that’s about as clear as mud, let’s look at one usage for the Cloud and show an example. We’ll look at using it for storage, and look at how and why MultichannelTV leverages it.
Web files (photos, videos, scripts, etc) are traditionally stored on a server. Let us say that company “Foo Ltd” saves all of its web files to a server based in London. For safety, this server probably has either a RAID configuration or an automated backup solution. This is expensive, and is still often limited to one data-centre. What if there is a fire at the data-centre and all the servers were destroyed? This would negate the RAID set-up, and assuming automated backups were housed in the same facility, these would be redundant too. At MultichannelTV we store our media content on Amazon Web Service’s (AWS) cloud platform, a platform called “Simple Storage Service”, or “S3”. This platform ensures our data is deployed across multiple machines in multiple facilities, in a high-reliability, cost-efficient way – this is apart from having our own physical backups located locally on-site and off-site. Using this platform not only ensures our data is physically secure, but it also opens up the use of other Cloud services, explained next.
Once again, let’s take a look at Foo Ltd. Foo saves its data to a web server based in London. If someone in France wants to view the content, they have to download it via unseen hops from their rooftop terraced apartment in Paris all the way to the data-centre in London. This could take a few seconds longer than someone who is trying to view the content from just a mile away. Take this even further, for users in Tokyo and New York. The data travels thousands of miles through countless cables – all of which increase the time. Secondly, what if 500 people try to access a large video file on the same server simultaneously? Assuming the server hasn’t succumbed to time-outs, overloading and other errors resulting in complete failure, any more users will have to wait a long time for their content. At MultichannelTV, we make use of AWS’s “CloudFront” platform to rid ourselves of this problem.
CloudFront is the core of a Content Delivery Network, or “CDN”. You may have seen this in web addresses already, eg: http://cdn.example.com - this has become the de facto format in URI's for CDN's. A content delivery network is comprised of clusters of servers all working together to serve data to any given user in the most efficient way possible. This takes into account (among other things) network conditions, geography and file-specific information. AWS CloudFront tightly linked to the S3 storage platform and consists of multiple machine clusters in strategically deployed locations (eg, London, Paris, Dublin, etc). If a user in London requests a video from our S3 stored location in London, then the video is duplicated across machines in the London cluster and served to the user. If a user in Paris requests the video, then the video is served from local clusters in Paris (who at some point in the past have retrieved the original file from the London based S3 facility). This setup ensures that users across the globe can access the videos from servers which are geographically closer to them, thereby reducing latency and resulting in a better user-experience. Also, as files are located on potentially thousands of servers, there is no problem if a traffic spike occurs. For example, one of our clients received over 50,000 views on a video within a week with absolutely no down-time or lag. Replicating that with traditional servers (implementing load balancing, mirroring, etc) would be comparatively hugely inefficient and incredibly expensive!
Here at MultichannelTV - to better serve our clients - we now incorporate many Cloud based services into our offerings. I have mentioned storage and CDN’s, but we also use it for email campaigns and notifications, processing data, encoding video into multiple formats, providing analytics and database-related tasks.
In essence, it provides us with a cost-effective, stable, scalable and exciting platform from which to serve our clients in a way which suits their needs best.
vaporwave
origamiaround
pixelskylines
tangentwithtanner
prguitarman