DB IP Service

• Protocol
• Warning
• Pre-requisites
• Testing
  • Performance
• Mac OS X
• Acknowledgements

TCP and Websocket service for the dbip MMDB Lite database.

Protocol

Simplest use is to send the IP address for which you want details to the service, and if it is a valid IP address, you will receive a JSON response with all the details.

Note: For the TCP service, finish the query with a single UNIX newline character (\n). The response will be terminated by two UNIX newline characters (\n\n). Other than this, there is no difference between the TCP and Websocket protocols.

The following example shows a request IP address and the response:

184.105.163.155

{
    "city" : 
    {
        "names" : 
        {
            "en" : "Oklahoma City (Central Oklahoma City)"
        }
    },
    "continent" : 
    {
        "code" : "NA",
        "geoname_id" : 6255149,
        "names" : 
        {
            "de" : "Nordamerika",
            "en" : "North America",
            "es" : "Norteamérica",
            "fa" : " امریکای شمالی",
            "fr" : "Amérique Du Nord",
            "ja" : "北アメリカ大陸",
            "ko" : "북아메리카",
            "pt-BR" : "América Do Norte",
            "ru" : "Северная Америка",
            "zh-CN" : "北美洲"
        }
    },
    "country" : 
    {
        "geoname_id" : 6252001,
        "is_in_european_union" : false,
        "iso_code" : "US",
        "names" : 
        {
            "de" : "Vereinigte Staaten von Amerika",
            "en" : "United States",
            "es" : "Estados Unidos de América (los)",
            "fa" : "ایالات متحدهٔ امریکا",
            "fr" : "États-Unis",
            "ja" : "アメリカ合衆国",
            "ko" : "미국",
            "pt-BR" : "Estados Unidos",
            "ru" : "США",
            "zh-CN" : "美国"
        }
    },
    "location" : 
    {
        "latitude" : 35.5248,
        "longitude" : -97.5971
    },
    "subdivisions" : 
    [
        
        {
            "names" : 
            {
                "en" : "Oklahoma"
            }
        }
    ]
}

Fields

If you prefer a simpler listing of field : value pairs, prefix the IP address with f:. The following example shows the same request as above:

f:184.105.163.155

city : Oklahoma City (Central Oklahoma City)
continent : North America
country : United States
country_iso_code : US
latitude : 35.524800
longitude : -97.597100
query_ip_address : 184.105.163.155
query_language : en
subdivision : Oklahoma

Location

You can also retrieve the geo-coordinates (latitude and longitude) for an IP address by prefixing the address with l:. The following examples illustrates:

l:184.105.163.155

35.524800,-97.597100

Warning

There is no authentication/authorisation implemented for this service. The assumption is that users will be running this service internally with no public exposure.

Pre-requisites

Download the MMDB lite database and place under an appropriate directory. The docker/run.sh script assumes the docker directory. Mount this file as /opt/spt/data/dbip.mmdb.gz when running the docker container. The entrypoint script will decompress the file into the running container and use it when running.

Testing

Integration test suite in Lua. Tests basic requests and responses.

TCP test suite for the TCP service using Boost:Asio.

Performance

Performance test suite using hayai and Boost:Beast WebSocket client is also available.

Performance test function sends 4 requests for each IP Address from a pre-defined vector of 50 addresses. Thus each test involves 200 websocket requests internally.

Each test runs the test function 10 times (iterations). Each set of 10 iterations is repeated 2 times to get better average numbers (higher repeats like 5 would be much better). We then run the performance test with different concurrency values. At present the tests are set up to run with 10, 100, 500 and 1000 concurrent threads connecting to the service.

Even with 1000 concurrent requests, the docker container rarely consumes above 25Mb RAM.

Concurrency	Avg Time per run (us)	Avg Timer per iteration (us)	Avg Iterations per second
10	8245983.332	824598.333	1.21271
100	48174280.958	4817428.096	0.20758
500	246399960.709	24639996.071	0.04058
1000	435108315.251	43510831.525	0.02298

All these numbers were against a docker container deployed on laptop. Tests were run from IDE. From previous experience deploying the test as another container could yield better results (depending on how much CPU is allocated to the Docker daemon). This is by no means definitive, since I was doing a few other things (watching a movie, another script running etc.) while the test was in progress.

Mac OS X

I have not been able to build or test using libmaxminddb on Mac OS X. There seems to be some incompatibility (probably some conflicting preprocessor define) with using libmaxminddb with boost. All testing has been via building and running the docker image.

No such issues on Linux obviously, since the docker container works without issues.

Acknowledgements

This software has been developed mainly using work other people have contributed. The following are the components used to build this software:

• GeoLite2++ - C++ wrapper library to read mmdb file via libmaxminddb.
• Boost:Asio - We use Asio for the tcp server implementation.
• Boost:Beast - We use Beast for the websocket server implementation. The implementation is a modified version of the stackless example.
• Clara - Command line options parser.
• NanoLog - Logging framework used for the server. I modified the implementation for daily rolling log files.
• hayai - Performance testing framework.

In future we will look at ftrie. The main downside to it is that we need to package the much (5x) bigger csv file in the image.