Planet scale or planet fail?
Systems designed for, and deployed to the cloud are infinitely scalable automatically, right? Well no, not quite. Microsoft's CosmosDb advertises itself as a "planet scale data storage solution" that can facilitate such scalability, but how is such magic achieved? Like most "magic solutions", they need careful consideration and design. With the promise of guaranteed and predictable throughput, CosmosDb sounds great but what are the factors to consider to ensure that this lofty goal is achieved?
This talk looks at some of the considerations we need to take into account to make CosmosDb work as advertised, at planet scale, and not have a "planet fail". We will not only look at the cloud service itself, but also at the client model, specifically at the most popular "SQL like" client with specific code examples and how this actually has some of the greatest impact at getting the most out of CosmosDb to achieve world domination.