In our Journey to the Cloud, we’ve made a few stops along the way to focus on a variety of cloud-based topics. Starting with the storing of content in a public cloud we’ve reviewed how to expand that library from a simple backup to an accessible, browsable content library that can be used for remote proxy editing. In this stage of our journey, we will explore the concept of ‘editing in the cloud’, where editors connect to cloud virtual workstations where they can perform their video production work.
What is a Virtual Workstation?
Conceptually, a virtual workstation is rather simple. Instead of having a high-powered local computer running an editing software suite and asset manager, we have the same software running in a cloud datacenter, on a virtual machine. The benefit of this virtual machine is you are essentially renting it for the duration of usage so your costs are for actual usage time. You also benefit because you don’t have to maintain the server, apply security patches, deal with broken hardware and all of the other headaches associated with that workstation in your office.
Similar to workstations in the physical world, virtual workstations come in all types of configurations suitable for different workloads. One can literally pick from hundreds of combinations of CPUs, RAM, storage, network interfaces and other features. For the world of video editing, workstations configured with GPUs are generally required. This allows the software configuration to run optimally for this video processing heavy software stack.
Virtual workstations are considered headless systems, that is, like all cloud servers they don’t have a monitor, keyboard and mouse. Not a big deal when you are running a database, but that’s not the case for a graphics intensive application like a non-linear editing suite. In order to make this workstation useful it is necessary to be able to attach a monitor. To do this, we rely on a remote desktop client.
If you’re familiar with remote desktop software such as Remote Desktop for Windows or VNC Connect, these applications permit a remote user to log into a workstation and allow the desktop window, keyboard and mouse controls to become remotely accessible much like they are being accessed locally. For video editing however many of these generally purpose remote software desktop packages don’t cut it. They were not designed for the high-resolution and high-refresh rates required to remotely project an editing interface. Fortunately there are solutions to meet those challenges.
Teradici Cloud Access Plus and Nice DCV are two software packages designed specifically to address these video intensive needs. These applications are installed on virtual workstations, relying on that workstation’s GPU to encode and transmit the video monitor ‘output’, keyboard, mouse and USB devices to a remotely connected client. It’s as though you are running a very long HDMI and USB cable from the cloud datacenter to your local computer. Additionally, for many environments, these solutions can support two monitors giving you a larger work area.
Now while these high-performance applications make it feel as though that remote computer is sitting on your desktop there are practical limits – physics, the Internet, and WiFi being the main culprits. Even over the best network, the further the physical distance between the virtual workstation and your local computer, the more perceptible the lag will be. For that reason when deploying a cloud editing solution, it’s important to choose a cloud datacenter physically close to your local office or home. Of course the Internet routing itself will have an impact as well and aside from distance you may want to ensure IP connectivity, typically measured using ping times and throughput are considered.
Lastly, it’s a well known fact that the WiFi network most computers connect to is the biggest culprit. For that reason, hardwiring your local computer to a router can help eliminate many sources of headaches. EditShare recommends a broadband download speed of at least 10-15 Mb/s per workstation monitor for best performance.
Workstation Instances and Software Images
Virtual workstations can be considered a product of the underlying server hardware and the software stack itself, so let’s review this topic in more detail
As stated earlier, all major cloud vendors offer a family of GPU enabled compute instances suitable for video editing. Much like buying a PC or Mac, you can choose how powerful of a system you want, and how much you’re willing to pay. Depending on your typical workload, what codec bit-rates are being edited, how many concurrent streams are being accessed and edited, you’ll want to choose the appropriate compute instance. Adobe has done a great job of publishing a set of guidelines for selecting the appropriate workstations for different personas of video editors. Here’s a table summarizing their choices for AWS, GCP and Azure:
The beauty of the cloud, is your choice is never final. If you choose the right cloud production solution, you can easily migrate from a low-end workstation to a high-end workstation and back again, all while preserving your editing environment and projects.
In a virtualized environment, software images can easily be deployed to a machine, operate for a while and then spin back down when not in use. To make this process efficient, software images are pre-baked into a golden image, so that when they are deployed to a virtual machine they are pre-configured and ready to go.
AWS calls these Amazon Machine Instances, GCP refers to them as an Image and Azure named them Virtual Disk Image. The concept is the same. An image is created with the operating system, applications and optimizations. This image can then be quickly deployed to a system and no further software installations are necessary.
In our example, this software image will include the remote desktop software, editing applications, and other supporting software, providing a fully-function environment to start working right away. Once deployed users can then make changes to their specific environment – customizing their environment, updating software, adding plug-in’s. Then, using persistent storage such as block-storage, each individual workstation can be shutdown while preserving that software image. This frees up the expensive compute instance and saves the image, at a reasonable cost, ready to be used by the user at a later point, on-demand.
A note about virtual workstations, as they are equipped with GPU’s and can the editing applications often require the use of Microsoft Windows, the cost of running a virtual workstation 24×7 can get quite pricey. Saving the software environment to persistent storage allows users to only use workstations when they are active. Since a 40 hour work week is about 24% of the total hours in a week, the savings can be fairly substantial.
Editing in the Cloud
For the most part, the major NLE application vendors are embracing the cloud. Adobe Premiere Pro and DaVinci Resolve can both operate in virtual environments. You can read about DaVinci on AWS here, and Adobe has done a great job documenting their best practices for Virtual Desktop Infrastructures on the big three cloud vendors.
With the other NLE vendors, we expect to see a progression towards more support as well. Final Cut Pro, however, is limited by the fact that MacOS is not available in cloud environments for editing use-cases. (Mac’s have recently been supported in AWS specifically for the development of software applications and at this moment don’t support FCP). Avid Media Composer is also offered in the cloud via an add-on VM license, though there are restrictions Avid applies to the use of their software.
Production in the Cloud
EditShare brings all of the same collaborative storage flexibility that it’s known for in the on-premise world up to the cloud. The ability to support multiple workstations and scale editing bandwidth to meet the challenge of production teams is dramatically more flexible in EFSv as servers and storage can be scaled up and down to accommodate the needs of any budget or project. Multiple editing stations, with any mix of horsepower, NLE applications, and workflow requirements can be instantly provisioned to meet a workgroup’s demands.
Get in Touch
Explore easy ways to get started in the cloud
Schedule time with an EditShare solutions specialist