SandDB

"I'm gonna put some sand in Cassandra's eye."

Custom Implementation of Apache Cassandra DB for SUTD ISTD 2022 50.041: Distributed Systems and Computing Group Project.

Project Structure

sanddb/
├─ anti_entropy/
├─ client/
├─ config/
├─ data/
├─ db/
├─ messages/
├─ read_write/
├─ ring-visualiser/
├─ utils/
├─ main.go

API Endpoints 🔚

Note: both partition keys and clustering keys are required since they form the primary key.

Create Table

HTTP Method

POST

URL

http://localhost:<port>/create/

Request Body:

{
  "table_name": "hospitals",
  "partition_key_names": ["HOSPITAL_ID", "DEPARTMENT"],
  "clustering_key_names": ["ROOM_ID"]
}

• table_name: name of the table to be inserted/updated
• partition_key_names: headers of the partition keys
• clustering_key_names: headers of clustering keys

Insert/Update

HTTP Method

POST

URL

http://localhost:<port>/insert/

Request Body:

{
  "table_name": "hospitals",
  "partition_keys": ["1", "GENERAL"],
  "clustering_keys": ["AA-1"],
  "cell_names": ["Bed", "Oxygen Tank"],
  "cell_values": ["3", "10"]
}

• table_name: name of the table to be inserted/updated
• partition_keys: values of the partition keys
• clustering_keys: values of the clustering keys
• cell_names: column headers to be added into the row
• cell_values: values of the columns to be added into the row

Read

HTTP Method

POST

URL

http://localhost:<port>/read/

Request Body

{
  "table_name": "hospitals",
  "partition_keys": ["1", "GENERAL"],
  "clustering_keys": ["AA-1"]
}

Params:

• table_name: name of the table to be queried from
• partition_keys: values of the partition keys
• clustering_keys: values of the clustering keys (optional)

Delete

HTTP Method

POST

URL

http://localhost:<port>/delete/

Request Body

{
  "table_name": "hospitals",
  "partition_keys": ["1", "GENERAL"],
  "clustering_keys": ["AA-1"]
}

Params:

• table_name: name of the table to remove an entry from
• partition_keys: values of the partition keys of the row to be deleted
• clustering_keys: values of the clustering keys of the row to be deleted

Database Structs 🏛️

table_name.json:

type Table []Partition

type Partition struct {
	Metadata PartitionMetadata `json:"partition_metadata"`
	Rows     []Row             `json:"rows"`
}

type PartitionMetadata struct {
	TableName          string   `json:"table_name"`
	PartitionKey       int64    `json:"partition_key"`
	PartitionKeyNames  []string `json:"partition_key_names"`
	PartitionKeyValues []string `json:"partition_key_values"`
	ClusteringKeyNames []string `json:"clustering_key_names"`
}

type Row struct {
	CreatedAt           time.Time `json:"created_at"`
	UpdatedAt           time.Time `json:"updated_at"`
	DeletedAt           time.Time `json:"deleted_at"`
	ClusteringKeyHash   int64     `json:"clustering_key_hash"`
	ClusteringKeyValues []string  `json:"clustering_key_values"`
	Cells               []Cell    `json:"cells"`
}

type Cell struct {
	Name  string `json:"name"`
	Value string `json:"value"`
}

Anti-Entropy

For future work in implementing the entire full Merkle Tree, as well as its comparisons, these repositories might be useful:

• https://github.com/cbergoon/merkletree
• https://github.com/aprimadi/merkle-tree

Acknowledgements

Credits and thanks to:

• Group 5 Team Members:
  • Filbert Cia
  • James Raphael Tiovalen
  • Ong Zhi Yi
  • Yu Nicole Frances Cabansay
• 50.041 Course Instructor: Professor Sudipta Chattopadhyay