gRPC-programming

A subset of Assignment 4 of the Distributed Systems course in IIIT-Hyderabad (Monsoon '24)

K Nearest Neighbours

More details available in the source directory.

The above was run with k=100.

Implemented with two separate clients, one to send data to be stored in the server, one to query each server for the $k$ nearest neighbours for their local data points.

Sequential Time Complexity consists of a simple heap implementation, such that for n data-points, to find the k nearest neighbours, we would need a time complexity of:

$$ O(n \cdot log(k)) $$

Parallel Time Complexity consists of querying for the local nearest neighbours of each server, and merging them together on the client side efficiently. Overall, we require a time complexity of:

$$ O(\frac{n}{s}\cdot log(k) + k\cdot log(k) \cdot log(s)) $$

To see run-times and qualitative and quantitative analysis, justified with plots for varying n, varying s and so on, see the report.

My Uber

More details available in the source directory.

An over-engineered rider-driver system, consisting of:

• load balancing - random picking, pick-first, and round-robin policies available
• mutual TLS - clients (riders and drivers) must provide a valid certificate that is issues by a trusted CA. All communication between client and server is encrypted using SSL / TLS.
• interceptors - for authentication, logging and other metadata.
• timeouts - the driver client must accept or reject a ride within a specified timeout period.

The servers function completely independently. For a more detailed analysis of the implementation and functionality, see the report.