Image your self within the function of CIO at Roblox in 2017.
At that time, the gaming platform and publishing system that launched in 2005 was rising quick, however its underlying know-how was getting older, consisting of a single knowledge heart in Chicago and a bunch of third-party companions, together with AWS, all working naked steel (nonvirtualized) servers. At a time when customers have valuable little persistence for outages, your uptime was simply two nines, or lower than 99% (5 nines is taken into account optimum).
Unbelievably, Roblox was fashionable despite this, however the firm’s management knew it couldn’t proceed with efficiency like that, particularly because it was quickly gaining in reputation. The corporate wanted to name within the know-how cavalry, which is actually what it did when it employed Dan Williams in 2017.
Williams has a historical past of fixing these sorts of intractable infrastructure points, with a background that features a gig at Fb between 2007 and 2011, the place he labored on the know-how to assist the younger social community scale to hundreds of thousands of customers. Later, he worked at Dropbox, the place he helped construct a brand new inner community, main the corporate’s transfer away from AWS, a significant enterprise involving transferring greater than 500 petabytes of knowledge.
When Roblox approached him in mid-2017, he jumped on the likelihood to tackle one other main infrastructure problem. Whereas they’re nonetheless within the midst of the transition to a brand new trendy tech stack as we speak, we sat down with Williams to learn the way he put the corporate on the highway to a cloud-native, microservices-focused system with its personal community of worldwide edge knowledge facilities.
Scoping the issue
#Roblox #fully #reworked #tech #stack #PJDM