Track 2 | 11:25
In this talk, Michael will cover how along with his team he pushed RavenDB to process hundreds of thousands of requests per second by using zero-copy data storage and autodidactive software, by reducing external dependencies and getting chummy with the operating system. Michael will discuss topics ranging from complete architecture overhaul to the using SIMD instructions to increase memory bandwidth and instruction rate on the CPU. The speaker will take a look at the guts of the CoreCLR and how it is possible to push it ever further to get the last push in service for amazing performance.
Michael is an experienced full-stack developer with background in development of applications and distributed systems. By day, as member of RavenDB Core Team he works mostly on the server side of things, currently participates in development of a new cross-platform version of RavenDB. By night he becomes a tinkerer with passion for new technologies, new ideas and coding in general, creating lots of different pet projects along the way.