https://github.com/astynax/assessment-kotlin

Solution wor the "IPv4 address counter" task
https://github.com/astynax/assessment-kotlin

Last synced: 19 days ago
JSON representation

Solution wor the "IPv4 address counter" task

• Host: GitHub
• URL: https://github.com/astynax/assessment-kotlin
• Owner: astynax
• Created: 2024-08-26T17:38:25.000Z (5 months ago)
• Default Branch: master
• Last Pushed: 2024-09-16T10:05:15.000Z (4 months ago)
• Last Synced: 2024-09-16T11:43:23.339Z (4 months ago)
• Language: Kotlin
• Size: 65.4 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
### IPv4 Address Counter

Here is my solution for [this problem](https://github.com/Ecwid/new-job/blob/c3189567685c3f63604f5c34d8f8270656b832fc/IP-Addr-Counter.md).

#### Rationale

Since the task didn't specify that I can only *roughly* count unique addresses, I decided to count each one precisely instead of using something like Bloom filter or HyperLogLog.

I thought that if I'll

1. use bit sets to store the last two bytes of the address

2. keep these sets in a `Map` using the first pair of bytes as a key

then I'll get a pretty decent overall storage density (~1 bit per unique address) and possibly will save some space because of the `Map`'s sparseness if some address ranges won't appear.

#### Results

I have my code tested with the mentioned inf the task sample file (120 GB). The program counted exactly 1 billion unique addresses using ~1.2 GB of heap on average.