Context

<p align="justify">
A fundamental problem in distributed real-time systems is the sharing of a communications resource between message streams on different nodes such that real-time requirements are satisfied. TDMA (time division multiple access) communication protocols solve this by assigning messages to time slots such that no two nodes transmit at the same time and queuing delays of messages are bounded. Typically these communication protocols operate on the basis of TDMA cycles, where a node is assigned one or many time slots. Usually each slot has a fixed length and the number of slots per cycle is fixed. Hence, a TDMA cycle has a fix and known time duration, and upper bounds on messages’ queuing delays can be proved.

The majority of research works on TDMA communications address the problem of finding appropriate schedules (TDMA frames/templates) for guaranteeing timeliness to real-time message streams. Unfortunately, the flexibility in assigning time slots to nodes in these approaches comes at a price: an unused slot is wasted and cannot be used for any other hard real-time traffic. A message stream with periodic messages may need a specific time slot in a TDMA cycle during only a few TDMA cycles; in the other TDMA cycles this time slot is not used and hence wasted. One way to overcome this waste is to have a large TDMA cycle serving several instances of a message stream. In the extreme case may need to choose the length of a TDMA cycle to be the least-common multiple of periods to avoid wasted slots. In contrast, however, consider TDMA protocols with slot skipping (TDMA/SS); that is, a slot is skipped when it is not used, hence the next slot can start earlier and this reclaims time for hard real-time traffic.

Within this context, we aim at:
* devising an accurate schedulability analysis, taking slot skipping into account;
* consider the case where a node can be assigned a number of slots;
* study the case of diferent scheduling policies for the output queues in each node;
* develop tools to aid planning distributed real time systems based on TDMA/SS. 
</p>