Hybrid serverless would be a deployment model that links hybrid cloud to the serverless deployment model. It is already started by IT vendors offering hardware rental services in the form of private clouds, putting them to customer's organizations, and charging at the rate of pay-as-you-go.

When the serverless and FaaS computing platforms are deployed on top of that kind of hybrid infrastructure, they become hybrid serverless. This could be the next generation of computing platform that allows you to store sensitive data inside the organization, having some important FaaS functions running on the local system, while leveraging some extra computing resources as pay-per-request. It will be in the scope of the definition of serverless, if the customer's organization does not need to maintain or administer any of the hardware servers. Fortunately, when mixing this model with what we have discussed throughout this book, using Docker as our infrastructure would still be applied to this kind of infrastructure. Docker is still a good choice for balancing between maintaining infrastructure on our own and making the serverless platforms do the rest of the work for us.

In the following diagram, the overall system shows a hybrid architecture. In the case of using a FaaS platform only from inside the organization, requests would be made firstly to the on-premises infrastructure. When loads become large, instances of the function executors would be scaled out and eventually burst to a public cloud infrastructure. However, the data stores are usually placed inside the organization. So, the outside function executors must be able to access them just as if they were running on-premises: