Last week we deployed the change to serve all texture and mesh data primarily through the CDN, as we've been doing with avatar textures since March. In addition to reviewing feedback from Residents we've been monitoring and measuring the effects of the change, and thought it would be interesting to share some of what we've learned.
First the good news:
- Load on some key systems on the simulator hosts has been reduced considerably. The chart below shows the frequency of high-load conditions in the simulator web services, and you can see the sharp drop as the CDN takes on much of that job. This translates into other things, including region crossings and teleports, being faster and more reliable.
- For most users most of the time there has been a big performance improvement in texture and mesh data loading, resulting in faster rez times in new areas. The improvement has been realized both on the official viewer and on third party viewers.
However, we have also seen that some users have had the opposite experience, and have worked with a number of those users to collect detailed data on the nature of the problems and shared it with our CDN provider. We believe that the problems are the result of a combination of the considerable additional load we added to the CDN, and a coincidental additional large load on the CDN from another source. Exacerbating matters, flaws in both our viewer code and the CDN caused recovery from these load spikes to be much slower than it should have been. We are working with our CDN provider to increase capacity and to configure the CDN so that Second Life data availability will not be as affected by outside load. We are also making changes to our code and in the CDN to make recovery quicker and more robust.
We are confident that using the CDN for this data will make the Second Life experience better. Making any change to a system at the scale of Second Life has some element of unavoidable risk; no matter how carefully we simulate and test in advance, once you deploy at scale in live systems there's always something to be learned. This change has had some problems for a small percentage of users; unfortunately, for those users the problems were quite serious for at least part of the time. We appreciate all the help we've gotten from users in quickly diagnosing those problems. We think that the changes we've begun making will reduce the frequency of failures to below what they were before we adopted the CDN, while keeping the considerable performance gains.