With all that’s going on in the world today, the remote working scenario has hit the headlines. There are numerous approaches to achieve a solution – each with its own advantages and disadvantages – but they can be broadly categorised into moving the data or moving the pixels.
The old school approach to working remotely is to use FTP or file transfer sites such as WeTransfer to send data to whoever needs it. This comes with some obvious restrictions on file size and how much you can send offsite.
File sharing solutions like Google Drive or Dropbox can also be used to ferry data to and fro, providing a slightly more advanced approach to transferring data, but these systems come with a fair use policy.
Both of these options require the remote user to have a machine capable of completing the workload and access to the actual software to do that work (which includes licencing, installation, and version control). While the user is only loading data, this data is also limited to the size of the internet connection they have and also comes with numerous security issues with data being sent offsite.
Chief among the methods of moving data to remote locations deploying a VPN (Virtual Private Network). But first and foremost: is your dataset small enough that we can work successfully over the internet?
A VPN is essentially a secure connection from a remote machine to your company network. This enables users to see all your mounted drives, workstations, and even printers from wherever they are. The downside is that the size of files you are trying to load over your home internet connection can be limited and downloads can take some time to complete.
The other issue is at home you might have a laptop or consumer level machine not necessarily capable of dealing with the files you need to load or have access to the software to actually work from home.
VPNs are normally fine for people who are working with small files (Word documents, PDFs, spreadsheets, small Photoshop files). They don’t function well in large pipeline scenarios, when dealing with large amounts of data, or access to things like central storage and render farms.
The most secure method of enabling remote working is not to touch the data at all. In this scenario, we leave all the data at the office – safe and secure – and give access to remote workers by distributing their work desktop back to another machine.
Among the advantages to this solution is the fact that no data actually leaves the office. You have access to the same machine you use day in, day out and you have access to all the same tools and pipelines that you have in the office.
In this workflow we only move the smallest amount of data: the pixels of the screen, streamed to your home much like Netflix or BBC iPlayer, but with a mouse and a keyboard. Think of it as using the internet as a really long monitor, mouse, and keyboard cable.
This means the remote device is only decoding an encrypted stream of pixels and, therefore, you could do this on any cheap computer. It also means you have access to the software and power of your workstation at the studio.
There are a number of approaches to this workflow, but the primary two are VDi (a virtualisation approach), or a one-to-one remote workstation.
All HP Z workstations running Quadro graphics ship with one of the world’s best kept secrets: HP RGS (Remote Graphics Software). This is a software agent that runs on the workstation. It consumes some of your compute resources (CPU and GPU) to send your desktop to end devices such as a laptop, tablet, zero client, or thin client.
Zero Client = machine with no OS (operating system), only firmware with minimal RAM - this offers good performance and is relatively low cost but has limitations on the number of monitors and resolution it can deliver.
Thin Client = machine with minimal OS, memory and GPU. It has a better processor but comes at a price but it has the benefit of no limitations and runs with RGS.
RGS does not support Wacom tablets and ideally needs to run behind a VPN. It also requires a good machine on the other end. It’s a heavily, well-compressed stream giving excellent bandwidth, but does require a lot of machine power to decode it especially at higher screen resolutions and FPS rates.
You can find the software here:
You will need to have a VPN and a remote machine capable of decoding the streaming pixels.
The disadvantage of RGS is that when the machine is completely maxed out then – if no resources are left or reserved for the RGS software – you will find your remote desktop performance is severely affected and in some cases can lead to dropout and disconnection. This can happen when the system is rendering/computing heavily on the CPU, or if a video stream compressor (such Adobe Premiere or Media Encoder) is utilising the GPU heavily again the remote session can be disconnected.
Teradici are a company who make a streaming pixel format called PCoIP. This is one of the most efficient and well-established streaming formats on the market today. It’s been ratified for use in Amazon, Azure, and Google for delivery of workstation.
The PCoIP pixel stream can be achieved in a hardware or software version.
Remote workstation card
This is a card that can be plugged into any PC and has a processor that compresses your GPU output into a streaming pixel format on the fly. The GPU output plugs directly into the card, which in turn consumes a network port, and this pushes pixels out over the network to any end device.
The advantage of this workflow is the machine has no overhead: it’s completely secured from everything the PC is doing and has no issues. This stream can be served out to remote workers via VPN (similar to RGS) but also has the benefit of using a Cloud Access Gateway Broker that allows you to bypass the VPN. It’s still secure as no data leaves the studio or is actually accessible. The only thing to leave the building is a stream of pixels.
The software version is similar to how RGS works and allows the use of the host machine’s CPU and GPU. It’s a much cheaper solution, but has the same limitations as RGS.
All of these solutions rely heavily on the bandwidth of your studio internet connection and the performance of your firewall as to how many remote workers could function at any given time.
The overall experience will also be defined by the remote users’ internet connection, bandwidth, and contention.
The contention problem is the elephant in the room. As anyone with children who has worked from home will know, come 3.30 pm when all the kids have returned from school the internet slows right down. With several people streaming YouTube or playing internet games the contention of everyone using a single home network makes working via the internet problematic to the point of not functioning.
And when everyone is working from home then the contention could come to a point where the internet crawls to a halt depending on who’s using what provider, where.