TrieIndexLogDB is a key/value datastore using log-structured files with trie structured indexes, allowing very fast write access and lowering RAM requirements compared to other log-structured datastores which use Hash Table Indices

1. Clone repo
   git clone https://github.com/yozaam/TrieIndexLogDB.git
2. Go inside directory
   cd TrieIndexLogDB
3. Compile the program
   g++ main.cpp -o main.o
4. Run the program
   ./main.o

"Example of Hash Index: Bitcask is an Erlang application that provides an API for storing and retrieving key/value data using log-structured hash tables that provide very fast access."

• General Architecture for Log Structured DB

  • Every create/update appends new key value pair into the file
  • maintain the pointer to the latest occurance of a key in the hash table / index

• Bitcask architecture / Hash Index

• TrieIndexLogDB architecture / Trie Index

• Bitcask stores every key in memory, along with the byte offset for that key in disk

• This limits the no. of keys that may be stored because they all have to fit in RAM

• My lightwieght storage engine will reduce the space requirements by using a 'Trie' index

(A string Hash (Bitcask) & Trie traversal both have the same asymptotic time complexity but Trie saves a lot of space by avoiding duplication!)

• Divide into separate classfiles
• Implement RestAPI class as well as CLI
• Implement segmentation, by splitting the files (will need to convert offset into (file_segment_no, offset) pair)
• Range Updates (by simultaneous Trie Traversal and sorted Trie Children)