When I was working at EA, both as CTO at the EALA studio and as a Tech Director at Redwood Shores / Maxis studio, there was a pretty large group of engineers devoted just to infrastructure, and another just to auto-test. It was all bespoke technology but those teams created some pretty impressive pipelines and tools that made a huge difference on large projects like the Medal of Honor and the Sims. It has been quite a few years since then and I have continued to look for those kinds of solutions that a smaller game studio could buy off the shelf without having to make a huge bespoke technology investment. There is just nothing like it available. Larger studios have invested in custom solutions that are often questionably cobbled together and costly to maintain, while smaller studios are left out in the cold.
This was the genesis for why we set out to build Wevr Virtual Studio (WVS).
A lot of the thinking behind WVS is a synthesis of the ideas born from having a glimpse of “what good could look like” at EA, and ideas from the absolute revolution called “DevOps” that has happened in the general software development space in the last few years. Especially the notion that effective devops should be self-service: there is no devops department, but rather individual team members should be able to leverage devops services when they need them in their personal workflows.
The analogy between what we are building and GitHub is a strong one, but for it to make sense, one really needs to think about it in terms of what GitHub has become in the last few years (a full featured dev-ops platform for software development) rather than just where it started out (a hosted repository).
Modern devops platforms are having a monumental impact on how software is developed, distributed and maintained. Devops pipelines have become an integral part of the project, rather than optional external processes.
We are at a point, where an individual can sit down on a platform like Github, start a new software project from scratch, and in one evening have automatic builds on dozens of platforms happening, automated unit tests running on every check in, notification of failures being distributed, packaged builds automatically created and a whole slew of other automation triggered. All of that can be pretty easily put in place by a single software engineer without any support from specialists.
The interactive space (gaming, real-time 3D), however, has not seen a direct benefit from these platforms for a number of reasons. First of all the teams have far more artists, technical artists and designers than they do engineers. In fact, there are an increasing number of projects that get by without pure engineers and just have gameplay scripters. The automation that is necessary is less around code builds and far more about data and design pipelines, build distribution, team coordination and data organization. The builds are more specialized and often involve tools that have not been containerized so that they could be usable as building blocks on a dev ops platform. A large part of the data flow comes from artists and designers rather than engineers and in many studios there is a bespoke hodgepodge of tools for managing different types of data rather than a single source of truth in the form of a unified repo. And finally, solutions to many of the pain points are best dealt with pre-checkin processes rather than the continuous-build post-checkin approaches that the software world commonly uses..
What we are building is a devops platform that is specifically tailored for interactive 3D production. It does not target a “devops department” or only the engineers, but rather the entire team, starting with artists. Artists should be able to control and take advantage of automation and test capabilities on the platform just as easily as designers or engineers. Producers should be able to control and configure the build creation and distribution features. etc.
We are starting with the repository because interactive projects have strong code/data and data/data dependencies and having a single source of truth helps enable automation across the project. As we talk to a lot of developers, a shocking number of them do not use any form of revision control for large sections of the team. We also see that the biggest initial efficiency impact WVS has on teams is across their non-engineering roles, simply because those have been historically underserved in technology based workflow improvements.
We want to increase the time game teams spend on making their project awesome, less on managing support infrastructure tasks, by relying on WVS - a purpose built cloud platform for game-engine-based development. WVS offers the whole team of technical and non-technical members a cloud hosted platform to collaborate and share progress on their game engine project. WVS securely provides all the necessary software features and cloud compute for teams using Unreal Engine or Unity to quickly work together and share tasks, iterate, generate renders and builds and easily set up playtests and reviews, in one place. WVS improves creative velocity for small to large teams and reduces a lot of headache and costs.
We’ve been battle-hardening WVS all through last year, and now have several productions up and running in private beta. If you’re starting to develop a game or real-time 3D, sign up for early access and set up your very own virtual studio with WVS.