Message Transport Layer (MTL)

A TCP-based message exchanging system for local area networks targeting low power embedded systems.

Developed by Dimitrios Karageorgiou,
during the course Embedded and Realtime Systems,
Aristotle University Of Thessaloniki, Greece,
2017-2018.

Message Transport Layer (MTL) is a message exchanging system on local area networks (LANs). It is based on TCP protocol and provides reliable message transfer. It consists of a server module and a client service. Server module -the core service of MTL- is expected to continuously run on low-power embedded devices such as routers or other low-power computers. Client service can be used by any client application that wants to communicate with MTL server for exchanging messages.

Server acts as a forwarder for the messages. It receives messages from connected clients and forwards them to their destinations. If a destination is offline, message is NACKed through a simple Negative Acknowledging mechanism. MTL-level error correction capabalities are solely based on this Negative Acknowledging mechanism, thus saving bandwidth compared to a possible Positive Acknowledging mechanism. In any case, error handling on MTL-level is expected to be something rare, if at all needed by the applications that make use of it.

Original target of the server was ZSun Wi-Fi card reader, running an OpenWRT port as its OS.

How to compile:

make : Builds both a server and a demo client for testing.

make server : Builds only server.

make client : Builds only demo client.

Executables are located inside bin folder under project's root.

In order to successfully compile, a compiler that supports GNU-11 C standard is required.

How to run server:

./bin/server <port> [<log_file> [<min_rate> <step> <max_rate> <period>]]

where:

• port : Port number to be used by server.
• log_file [optional] : Path to a file that will be used for storing log data.
• min_rate [optional] : Minimum sending rate of MTL in messages/sec.
• step [optional] : Step of reduction for sending rate of MTL in messages/sec.
• max_rate [optional] : Max sending rate of MTL in messages/sec.
• period [optional]: Period of rate limiter in milliseconds (ms) to reduce rate by given step.

min_rate, step, max_rate and period provides a way to setup a rate limiter that periodically reduces sending rate of MTL server. It starts from max_rate and at each period reduces sending rate by step. When rate drops below min_rate it starts again from max_rate. Normally, rate limiter is expected to be turned-off (i.e. none of the last four args provided). Though, it's useful for conducting various tests.

Demo client:

In order to experiment with MTL and conduct a series of tests, a demo client is implemented. Demo client can operate either in interactive mode or in testing mode.

Interactive mode:

At this mode, demo client reads messages from stdin in the form of destIP:destPort message and sends them. Client in interactive mode can be invoked as following:

./bin/demo_client <server_hostname> <server_port> -mode=i <port>

where:

• server_hostname : IPv4 address in dot format or hostname of server.
• server_port : Port number on server where MTL service is running.
• port : Port which will be used by demo client.

Testing mode:

In testing mode, demo client is able to start multiple clients at once. It is also able to generate and exchange a given amount of messages between started clients, while verifying their correct receipt. Testing mode can be invoked with two different destination selecting options. It can either be invoked with random argument, which selects as destination for each started client a random client from the rest, or with all argument, where each client sends the specified amount of messages to all other clients. The executable can be invoked as following:

./bin/demo_client <server_hostname> <server_port> -mode=t <clients_num> <send_mode> <messages_num> <if_ip>

where:

• server_hostname : IPv4 address in dot format or hostname of server.
• server_port : Port number on server where MTL service is running.
• clients_num : Number of clients to be started.
• send_mode : 'all' for send to all, 'random' for send to random
• messages_num : Number of messages to be send by a client to each of its targets.
• if_ip : IP assigned to the interface which will be used for communicating with the server. It should be the IP visible to the server. If device is behind a NAT, the public IP of the NAT should be provided.

Licensing:

This project is licensed under GNU GPL v3.0 license. A copy of this license is contained in current project.