MF-TDMA Applied to DVB-RCS2

Slicing time and spectrum

Posted by Yoann on October 28, 2015

Categories: notes,satellite

Traditionally in video broadcasting services, there is one entity broadcasting its content over different frequencies. That is a classical bandwidth sharing technique called Frequency Division Multiplexing (FDM). Each flow is sent over a specific frequency dedicated to it so that it doesn’t interfere with other users.
In digital telephony (e.g. ISDN), a different approach is used: users transmit only during short time slots, periodically and on the same frequency. This is call Time Division Multiplexing (TDM). The timeslots are assigned statically to users, and hence unused when the user doesn’t use it.

But when the potential user number overpasses the capacity, it is necessary to use an access method to dynamically assign resources. FDM and TDM have their corresponding access methods with Frequency Division Multiple Access (FDMA) and Time Division Multiple Access (TDMA) techniques. Instead of assigning statically the frequency/time resources, they just define an Frequency/Time division plan that will be use by the MAC layer as allocable resource. FDMA was used in mobile telephony and some satellite systems, TDMA is used in GSM and some optical networks.
MF-TDMA is a combination of FDMA and TDMA: multiple frequency channels are available, and each is divided in timeslots. Users are provided dynamically one or more slots of time/frequency for their traffic.

The DVB-RCS2 standard uses MF-TDMA. In the following we’ll try to explain how this standard organises all (frequency, time) couples that are available.
In this article, we’ll focus on a specific case of the standard, with linear modulation and burst transmission.

System view of a DVB-S2/RCS2 satellite. The Forward link uses DVB-S2, the return link the DVB-RCS2
Figure 1: System view of a DVB-S2/RCS2 satellite. The Forward link uses DVB-S2, the return link the DVB-RCS2 [PNG]

DVB-RCS2 Waveforms

DVB-RCS2 defines several waveforms to be used. Some (up to 128) are predefined by the standard, and others can be defined by the system user (up to 128). Each waveform defined for a system shall be provided in advance to each RCST. A waveform is defined by a few parameters such as the modulation scheme, coding rate, burst length in symbols, pilot scheme etc.

Below are a few examples of waveforms. Many waveforms use the same symbol length, so that ACM can be done easily by interchanging the waveforms.

Some waveform examples with 232, 536 or 1616 symbol bursts
Figure 2: Some waveform examples with 232, 536 or 1616 symbol bursts [PNG]

MF-TDMA entities

Bandwidth-Time Unit (BTU)

The smallest unit in MF-TDMA is the BTU. A BTU is a slot, with defined duration, carrier bandwidth and symbol rate.
Hence, only a limited number of waveforms can fit to a BTU type, given its duration and symbol rate.

Timeslot

A timeslot is a sequence of BTU of the same type over a single carrier frequency. Hence, the duration of a timeslot is a multiple of the BTU duration, and the bandwidth is the same as the bandwidth of a BTU.

There are different types of timeslots, depending on how many BTU they cover, or their function. As far as I know, there are 5 types of timeslots:

  • Type G (Generic): they are 1 BTU large and can be aggregated together just in time (right before being transmitted, according to the need) to form other types of timeslots.
  • Type CB (Control Burst): like the type G BTUs, they are one BTU large, but they are used for signalling only.
  • Type LB (Logon Burst): this timeslot type is specific and uses 3 BTU, used for logon.
  • Type TRF1 (Traffic 1): they are 2 BTU large and used for user traffic.
  • Type TRF2 (Traffic 2): they are 6 BTU large and used for user traffic.

Frame

A frame is an combination of timeslots. It is possible to combine different types of timeslots in the same frame, as long as all the timeslot types use the same type of BTU.
It is possible to define several frame types, according to the need of the implementer.

Below are 3 examples of frame templates, without the repartition of timeslots.

Three examples of different-sized BTU frames, without timeslot scheme applied
Figure 3: Three examples of different-sized BTU frames, without timeslot scheme applied [PNG]

A frame is the combination of one of the above “templates” and a timeslot scheme, as shown on the following figure extracted from the standard.

Example of timeslot repartition, grey grid in the background is the timeslot base for the frame
Figure 4: Example of timeslot repartition, grey grid in the background is the timeslot base for the frame [PNG]

Superframe

A super frame is an aggregation of one or more frames. Frames do not have to be contiguous (time and frequency wise). A super frame is identified through its superframe count. A superframe total duration is generally between 25ms and 750ms (even though it can be higher in some implementations.).
Frames inside a superframe can be of various types, as long as their common carrier frequency bandwidths are the same. An example of superframe is shown on the simple example below.

A simple superframe example, with 3 types of frames, timeslots are ommited for clarity
Figure 5: A simple superframe example, with 3 types of frames, timeslots are ommited for clarity [PNG]

A group of super frames of the same type is called a SuperFrame Sequence (SFS).

Conclusion

In this article we presented the MF-TDMA access method used in the case of linear modulation with bursted transmission for the Return Link of DVB-RCS2 standard.
We overviewed the decomposition of the bandwidth, broadcasted to every user connected to the terminal periodically, along with the dynamic allocation scheme.
This is done to allow thousands of users to access the channel without interfering with each other, and it works !


Sources:

  1. Jean-Baptiste Dupé Thesis
  2. Second Generation DVB Interactive Satellite System (DVB-RCS2); Part 2: Lower Layers for Satellite standard


2 Comments

Jens -- 2018-04-01 21:05

Source of a superfram
Very good overview! But I‘m not sure about the source of a superframe. Is the originator of all superframe parts an only RCST or contributes multiple RCST to the superframe?. The latter seems to be more probable for me.

Yoann -- 2018-06-21 10:53

Re: Source of a superfram
Hi Jens, Thanks ! You are absolutely right, the actual transmitters of bursts (timeslots) are the RCSTs, but it is the gateway that organizes the superframe and assign the resources (timeslots) to RCSTs. This is done through the DAMA mechanism, which is described in the RCS2 standard, and can be summarized as follows:

  1. The RCST requests a resource allocation (via already allocated resources OR via a random access channel).
  2. The gateway assigns resources the each RCST and forms the corresponding superframe.
  3. The gateway broadcast the allocation table to every RCST on the forward link (through TBTP2 or FCT2 tables of RCS2)
  4. The RCST transmits on the assigned timeslots
I'll try to do an article on this when I have more time...
Cheers! Yoann


Post a comment

All comments are held for moderation; basic HTML formatting accepted.