I have been very busy these last weeks to set up the new infrastructure for ToolsLib. It’s a long project which covered the last six months in which it was essential to think about every corner cases we could have and to plan the evolution of the platform in terms of traffic, usages, data quantity…
The criteria for the redesign were redundancy, speed & low-latency and security. These points had to be adapted with our increasing traffic and usage. Everyday we have to ensure data integrity and fast access to each visitor, it’s not so easy to do (and not really school-compliant..).
We were close to saturate our network link last summer and it was not enough to follow the evolution and our projects. So we made some choices :
This allow us to offer a capacity of 2x1Gbps with a reduced latency (we gain between 10 and 20ms) and to easily scale, even with peak traffic. It’s really exciting to see these changes being used on production.
More details and figures about the managed infrastructure are explained on ToolsLib’s blog : How do we handle a permanently increasing traffic on ToolsLib.