Question #$1$ Serverless Computing $(5$ Pts$.)$

Answer the following general questions with respect to serverless computing:

$1.$ What is serverless computing? Give three examples of applications that fit into this novel model. For each example

application, give concrete arguments that support the suitability of serverless computing to the application. Also, discuss what

serverless computing cannot do$?$

$2.$ What are the main differences between PaaS $($Platform as a Service$)$ and FaaS $($Function as a Service$)$ delivery models?

$3.$ One key aspect of Cloud Computing is economics. That is$,$ the question of whether Cloud Computing is a costeffective endeavor. FaaS is $$ideal$$ for short and asynchronous computations, such as the creation of a thumbnail for each

image file that is uploaded to an object store $($e$.$g$.,$ Amazon S$3),$ but not for all applications.

Depending on the application, it would pay off to use a classical virtual machine $($VM$).$ To better understand this, it is

enlightening to perform a simple break$-$even analysis of the cost between a FaaS offering $($e$.$g$.,$ AWS Lambda$)$ against a VM

$($e$.$g$.,$ an EC$2$ instance$)$:

As a use case, pretend to be a payment microservice where each payment request takes around $250$ms of CPU time and

consumes $64$MB of RAM on average. Now assume the following hardware specs:

Function VM nano instance

$$ $$0\mathrm{.}20$ per $1$ million invocations

$$ $$0\mathrm{.}0000166667$ for every GB$-$second

$1$ vCPU and $0\mathrm{.}5$GB RAM

$$ $$0\mathrm{.}0058$ per hour

Now compute how many requests per day would be necessary to break even. Under low traffic conditions, which technology is

the cheapest? Why?