Skip to content

Latest commit

 

History

History
91 lines (61 loc) · 4.5 KB

README.md

File metadata and controls

91 lines (61 loc) · 4.5 KB

Twitter-Simulator

Shaishav Shah - 1136-3317

Akib Maredia - 3885-6489

Files overview

  • MessageType.fsx - Contains request response type definitions used for communication between a Client and a Server.
  • Server.fsx - Contains all the logic related to User processing, Tweet processing etc. Runs as a standalone process.
  • Client.fsx - Contains simulation logic and Client actor logic where each Client corresponds to a single user and performs all the functions for a user.

Instructions to run the project

  • Unzip the file AkibMaredia_ShaishavShah_Proj4.1.zip
  • Move to src directory
  • Start the Twitter Engine using the command dotnet fsi --langversion:preview Server.fsx and wait for the message "Server is running!" to allow the Server to boot up
  • In a new terminal instance or in a new terminal tab, start the Simulator process by passing the number of users as an argument - eg: dotnet fsi --langversion:preview Client.fsx 1000
  • The Simulator will run until all the clients finishes sending the tweets to the engine, and would print out the relevant stats to the screen.

Functionalities implemented

  • The user can register himself to Twitter
  • The user can login to or logout from the Twitter (live connection / disconnection)
  • The user can follow / unfollow another user already registered
  • The followers are distributed using the Zipf distribution
  • The user can send tweets with / without hashtags and / or mentions (mention another user)
  • The user can retweet the tweets posted by the users he's following to let his followers know about some interesting tweet
  • The user can see his feed i.e the tweets posted by him, tweets posted by the users he's following, and the tweets where he is mentioned
  • We divided the total number of users in 3 categories: Celebrities, Influencers, and Common Men
    • Celebrities: Very small percent of users having very high number of followers and posting most number of tweets
    • Influencers: Small percent (little higher than the celebrities) of users having high number (lesser than celebrities) of followers and posting second most number of tweets
    • Common Men: All the remaining group of users having very less follower count and posting very less tweets

Performance

For 1000 Users

#Users #Followers Avg. #Followers #Tweets + #Retweets Avg. #Tweets
Total 1000 2577 - 24785 -
Celebrities 1 277 277 244 244
Influencers 5 298 59 653 130
Common Men 994 2002 2 23888 24

Offline users :- 17.0% = 170 (Chosen randomly between 10% - 20% of total users)

Total time taken :- 21 Seconds

Number of tweets handles per second :- 1128

For 5000 Users

#Users #Followers Avg. #Followers #Tweets + #Retweets Avg. #Tweets
Total 5000 67894 - 194899 -
Celebrities 2 2090 1045 571 285
Influencers 25 6458 258 3807 152
Common Men 4973 59346 11 190521 38

Offline users :- 14.0% = 700 (Chosen randomly between 10% - 20% of total users)

Total time taken :- 610 Seconds

Number of tweets handles per second :- 320

For 10000 Users

#Users #Followers Avg. #Followers #Tweets + #Retweets Avg. #Tweets
Total 10000 280825 - 447821 -
Celebrities 5 10312 2062 1539 308
Influencers 50 28692 573 8205 165
Common Men 9945 241821 24 438077 45

Offline users :- 15.0% = 1500 (Chosen randomly between 10% - 20% of total users)

Total time taken :- 2073 Seconds

Number of tweets handles per second :- 217

Maximum number of users tested: 10000

Note: Running more than 10K users is possible, but not tested since it takes more than 30 minutes.