Skip to content

Latest commit

 

History

History
294 lines (265 loc) · 7.26 KB

README.md

File metadata and controls

294 lines (265 loc) · 7.26 KB
Raw

embDB

Simple Time Series Database

The database has a simple structure, few dependencies and is therefore intended to be used on embedded Linux devices. embDB scales perfectly and can be used on full-blown linux distros as well.

Facts

Key capabilities

  • Each row can have its own datatype and maximal itemcount.
  • Automated overwriting of old items in ringbuffer manner per row.
  • Each item has its timestamp in milliseconds automatically set by the database.
  • Read out of items can be done on entire row or between timestamps in row.
  • The database is strictly enforcing the usage of datatypes and is typesafe.
  • The access to the database is threadsafe implemented.
  • Serialization / Deserialization is done automatically on start / stop of embDB service.

Datatypes

Supported datatypes are:

  • string
  • unsigned integer
  • signed integer
  • unsinged long
  • signed long
  • float
  • double
  • bool
  • bytes vector

Timestamps

Each entry in the database has its timestamp which has millisecond resolution since epoch and is packed in an signed long

Limitations

The entire database is stored using a protobuf serialization. When the embDB service is running, the entire database is kept in RAM. When the service is stopped, the latest db version from RAM is serialized and stored to the persistent location. Therefore a proper stop of the service is necessary. Proper in this case means by SIGTERM or SIGINT.

Try it out

Docker

Build & Run

From the project folder:

cd integration/docker
docker build -t "embdb:latest" .
docker run -d -p 5678:5678 embdb:latest

Python test scripts

There are test scripts written in Python which can also help as examples how to use the database.

cd integration/python
python test_main.py

This script exits after one test fails or on successful execution of all tests. What it basically does is:

  • Connect to TCP socket of running docker container
  • Add / Delete rows
  • Clear database
  • Write / Read item to rows

To execute the python script, the docker container must be running (see above)

Database structure

The database has the following structure

Hash X Ringbuffer info Type Item A Item B ... Item N
: : : : : : :
Hash Z Ringbuffer info Type Item a Item b ... Item n

Hash: Built from the row-name,

Ringbuffer info: Index, maxItems, overflow, curItem (stuff for ringbuffer).

Type: Datatype of items in row

Item x: One item, containing timestamp & value

Each row has its own itemscount and overflows automatically if more than itemscount items are added.

Interface

Access

The database can be accessed by multible clients over a Unix-Domain-Socket or a TCP-Socket. More infos how to set the path or port are given in the help of the embDB service when calling

root@machine embdb -h
embdb:
 -h | --help         Show this help
 -v | --version      Print the version
 -l | --logging      Logging file (/etc/embdb/logging.ini)
 -d | --daemonize    Run as daemon
 -p | --pidfile      PID file (/var/run/embdb.pid)
 -s | --socket       Path to socket (/var/run/embdb.sock)
 -f | --file         Path to database file (/var/data/sd/database/embdb)
 -t | --tcpport      Tcp socket port (8085)

Protocol

Currently only a JSON protocol is implemented. Other protocols can be added using the implemented interfaces.

JSON

CreateRow

Creates a new row in the database. A row with an unique name must be created to add items to it. The "itemsCount" field defines how many items are stored in this row. If as in the example below 1000 items are stored and the 1001 is added, the first added will be overwritten (ringbuffer like).

Request:

{ 
    "command": "createRow", 
    "name": "yourRowName", 
    "type": "yourRowType", 
    "itemsCount": 1000 
}

Result:

{ 
    "command": "createRow", 
    "name": "yourRowName", 
    "error": 0
}

ClearAll

Clears all rows with their items.

Request:

{ 
    "command": "clearAll" 
}

Result:

{ 
    "command": "clearAll", 
    "error": 0
}

DeleteRow

Deletes the row with the specified name.

Request:

{ 
    "command": "deleteRow", 
    "name": "yourRowName" 
}

Result:

{ 
    "command": "deleteRow", 
    "name": "yourRowName", 
    "error": 0 
}

RowExists

Returns info if the row exists.

Request:
{ 
    "command": "rowExists", 
    "name": "yourRowName" 
}

Result:

{ 
    "command": "rowExists", 
    "name": "yourRowName", 
    "error": 0, 
    "rowExists": true 
    }

RowCount

Number of all rows in database.

Request:

{ 
    "command": "rowCount" 
}

Result:

{ 
    "command": "rowCount", 
    "error": 0, 
    "rowCount": 2 
}

ReadRow

Read all items in a row.

Request:

{ 
    "command": "readRow", 
    "name": "yourRowName", 
    "type": "yourRowType" 
}

Result:

{ 
    "command": "readRow", 
    "error": 0, 
    "type": "yourRowType", 
    "values": [{"value": 89, "timestamp": 1234567891},{}] 
}

ReadRowTimestamped

Returns all items between the given start & end time.

Request:

{ 
    "command": "readRowTimestamped", 
    "name": "yourRowName", 
    "type": "yourRowType", 
    "startTime": 1234567890, 
    "endTime": 1234567899 
}

Result:

{ 
    "command": "readRowTimestamped", 
    "error": 0, 
    "rowCount": 10, 
    "startTime": 1234567890, 
    "endTime": 1234567899, 
    "type": "yourRowType", 
    "values": [{"value": 1, "timestamp":1234567897},{}] 
}

AddItem

Adds an item to the given row in the appropriate type. If you add floating point items you always need to add at least one position after decimal point (e.g. for 0 --> 0.0)

Request:

"data" in appropriate type

{ 
    "command": "writeItem", 
    "name": "yourRowName", 
    "type": "yourRowType", 
    "data": 1234
}

Result:

{ 
    "command": "writeItem", 
    "name": "yourRowName", 
    "error": 0
}

Errorcodes

- 0: SUCCESS

// Database errors
- 10: NOTFOUND              --> Requested row not found
- 11: INTERNAL              --> Internal error
- 12: HASHNAMEMISMATCH      --> Rowname hash mismatch
- 13: ITEMSCOUNTMISMATCH    --> Already existing row with other itemscount
- 14: TYPEMISMATCH          --> Type does not match for this row

// Protocol errors
- 50: NO_COMMAND            --> No command given
- 51: UNKNOWN_COMMAND       --> Command not known
- 52: NO_NAME               --> No rowname given
- 53: NO_TYPE               --> No type given
- 54: UNKNOWN_TYPE          --> Type not known
- 55: NO_STARTTIME          --> No starttime given
- 56: NO_ENDTIME            --> No endtime given
- 57: NO_ITEMSCOUNT         --> No itemscount given
- 58: NO_DATA               --> No data given
- 59: INVALID_DATA          --> Invalid data given

Setup

Environment setup

Please refer to the Dockerfile as an example of how to setup your build environment.

Compile embDB

mkdir build
cd build
cmake ..
make
sudo make install

Without automatic unit-test execution:

mkdir build
cd build
cmake .. -DCOMPILE_GTESTS=OFF
make
sudo make install