DEVBLOG: THE FUTURE OF DU - Part 2: Under the Hood

[NQ-Naerais]

 

PART TWO: UNDER THE HOOD 

In our previous blog, we spoke about improvements we’re making to our processes in order to get more efficient use of our resources and deliver the best game possible. Today, we’re taking a look at operation costs and how we’re working to make improvements in that area, too. 

THE COST OF DOING BUSINESS 
The server and database architecture behind Dual Universe is new and quite complex. As we had about the game itself,  we had made assumptions about how much running a game like DU was going to cost in various areas. What we found was that our estimations proved to be far too conservative. We had to confront the actual cost of scaling up operating costs for a larger, global player base. And that’s okay. We have a plan. 

 

After reevaluation, we determined that some of our design decisions have had a major impact on our operational costs (i.e. what it costs to run the servers). In particular, databases are a major part of these costs. For example, because the world can be entirely edited and every edit needs to be stored in a database then communicated between the server and the clients, it eats a lot of our I/O allocations, which in turn increases our database costs. 

 

THE IMPACT ON UPDATES

The consequence is that, in the upcoming months, we will be rolling out significant backend improvements in order to optimize these operational costs. We made the conscious decision to keep pricing low in order to make the game more accessible for people who want to play, but that means that we need to keep operational costs in check. Developing these optimizations is time-consuming, but it is fundamental if we want to have a viable game as the playerbase continues to grow. 

 

It’s important to make players aware of this because some of our upcoming releases may not seem like much to you. While they won’t appear to include much in the way of new features, they will be updates to deploy these fundamental under-the-hood changes. For instance, the upcoming 0.25 release will be primarily focused on introducing a “game-changing” incremental storage mechanism for edits to the game world that will have a major positive impact on our database costs. The good news is that we think that some of these modifications will improve performance for the players.

 

Speaking of performance: that’s another area where we’ve been doing a lot of work, using the in-game telemetry to optimize areas where performance was dropping. There are several projects in the works to address performance. Although these may not seem particularly sexy when you read about them in the patch notes, the difference in how much better the game runs and feels while you’re playing should convince you it was worth the time and effort we spent on them. 

 

One example is the in-game screens. They ended up being quite popular among players, but the use of HTML for screen customization has proven to be quite a performance bottleneck in areas where there are a lot of screens. We recently started testing a new system using simple Lua draw commands instead of HTML to achieve the goal of screen customization. This change should seriously alleviate performance issues in areas with a lot of screens while providing the added benefit of unifying customization languages in a more user-friendly way.

 

There are other optimizations like this in our pipeline, and they will be presented in due time. 

MORE IN STORE

Tomorrow, in the third and final part of this series, we’ll talk about DU gameplay, and the challenges of maintaining a delicate balance between staying true to the original vision while taking player feedback into consideration. 
 

Discuss today's blog here!