ServerlessConf: Managing your Serverless Servers
The backbone of serverless architecture is using Functions as a Service to handle your logic needs. This is done by spinning up containers to run these functions on demand. For many use cases, this would be enough. Other times, you would need to manage small configurations such as time outs and memory. More often, a developer needs to use third-party libraries or a company built and maintained library. For small changes, the timeout and memory can be adjusted through configuration settings with the cloud provider such as AWS Lambda or Google Cloud Functions. To import third-party or other pre-built libraries, cloud providers will accept deployment packages that can be uploaded with a zip file or other deployment procedure after building the package locally. AWS has also introduced Layers as a way of having a standard way of deploying libraries to already existing runtimes. In more complicated scenarios, such as an IT Department requirement that any runtime be ‘blessed’ or maintaining a golden archive of runtimes, a combination of OpenWhisk and Docker Blackbox can allow you to choose your own runtime. This talk will go over all of these options, how to choose which is the best for your particular use case, and how to implement them.