“It’s consumed very rapidly and because of the size, the growth of company at the moment, and the number of projects we’ve taken on and the complexity of the techniques that are used to generate those images - people want 2K, HDR, stereo, 4K (there are a lot of changes in technology driven by the consumer side) - people expect a very high visual quality. We have invested heavily in areas of technology that are cost-effective but give us an increase in efficiency and capacity to be able to deliver more of those iterations through the pipeline,” he said.
The company has also virtualised its core services and updated its core infrastructure and edge switching with VDX switches.
“Our networking gives us the ability to deliver data from one place to another very quickly - the primary areas we want to move data is the render farms and storage and then ultimately back to the users to view the rendered material.”
He said the company also relies heavily on its asset management system, whereby assets can be checked in by a user or by an automated process.
“That allows us, in an automated way, to scan the file system and automatically take things offline once they’ve been protected to tape or once they've been archived. Or if it’s a file not considered an asset, and only needed for a short period of time to make something else or if it was a deprecated version, we have automated processes to scrub that off the system,” he said.
“As you can imagine a human trying to look at two million files just for one show. Lego had 185 million files that were under 2 megabytes on the file system at peak. The current shows is three to four times that size, like Lego Gotham and Lego Batman, so the problems are getting bigger and more complex. We definitely rely on improvements in technology and data analytics and the ability to automate a lot of those processes based on policy to stay on top of it.”
Asked his next steps, he said the virtualisation of infrastructure is on the cards.
“Virtualising the render farm would be a potential for us," he said. "As compute gets more and more dense, it becomes more and more challenging to chop up your environment in a way that makes sense and so virtualising and completely abstracting the hardware layer is something of interest. The same can be said for storage - we are definitely keen on looking at more modular, versatile storage.”
The company is also looking at ways to enable the fastest, most efficient way of deploying infrastructure for short periods, which could mean adopting a cloud strategy.
Sign up for CIO Asia eNewsletters.