Consider an loT application for facial identification. The system uses a camera to capture

an image of a user, and if the user is authenticated, a gate is opened.

The camera captures a $480480$ image using $16-$bit pixel resolution. Before sending the

image to the cloud, the picture is compressed using a compression algorithm that yields

$50$:$1$ compression ratio. Then, the resulting image is sent to the cloud application for

processing using an $Al$ algorithm. After the image is processed, the result is sent back to

the client as a $500$ bytes payload.

Assume that executing the compression algorithm at the edge takes $25ms.$ Also assume

the execution of the Al algorithm on the cloud takes $100ms.$ RTT between the client and

the cloud application is $250ms.$

A$.$ Determine the size of the payload sent in each request $($from the client to the

cloud only$).$

B$.$ Compute the latency experienced by users if non$-$persistent HTTP with

piggybacking is used to transfer the data between the client and the cloud

application. Assume the complete payload is sent in one request.

C$.$ Repeat part B assuming Confirmable CoAP is used instead of HTTP$.$

D$.$ Compute the protocol overhead $($in $\%)$ for each complete transaction $($request

and response$).$ Assume the client is using IPv$4$ and assume a data link layer header

of size $20$ bytes. Also, assume HTTP header is $100B.$

E$.$ Assuming the system is used in a metro station with $25$ cameras, each controlling

a separate gate. Determine the maximum service rate of the system $($passengers

per minute$)$ that will result when using each of the systems in B and C$.$