ITFI960066A1 - SYSTEM FOR THE TRANSMISSION VIA SATELLITE OF SYNCHRONOUS AND ASYNCHRONOUS DATA ABLE TO CONTRAST THE ATTENUATION OF THE SIGNAL DUE TO AGENTS - Google Patents

Description

Description of the industrial invention entitled:

"SYSTEM FOR THE TRANSMISSION VIA SATELLITE OF SYNCHRONOUS AND ASYNCHRONOUS DATA ABLE TO CONTRAST

THE SIGNAL ATTENUATION DUE TO ATMOSPHERIC AGENTS "

The present invention relates to a system for the satellite transmission of synchronous and asynchronous data, in particular for video conferencing, capable of counteracting the attenuation of the signal due to atmospheric agents.

Pi? in particular, the present invention relates to a satellite channel access system which satisfies the aforementioned purpose and which for convenience will come? identified in this description as the "FODA / IBEA-TDMA" system.

"FODA / IBEA-TDMA"? an acronym for FIFO Ordered Demand Assignment / information Bit Energy Adapter-Time Division Multiple Access. This is a satellite channel access method for assigning the satellite band upon request. Bandwidth allocation? carried out with technology called TDMA, that is? in time division. Users requesting access to the satellite are allowed to use the entire band at different times, in order not to create collisions.

Purpose of the FODA / IBEA method? to allow the simultaneous transmission of both synchronous and asynchronous data on satellites, maintaining the quality? of the service required by individual applications even in case of strong attenuation of the transmission signal due to bad atmospheric conditions (technique called "fade countermeasure"). The innovative aspect of this system consists in the fade countermeasure technique adopted. "FODA / IBEA-TDMA"? the logical evolution of the FODA-TDMA access method, Italian patent n. 1233264, filed on March 21, 1989

FODA / IBEA allows the simultaneous transmission of both data generated in real time (synchronous traffic or stream) and data generated in non-real time (asynchronous traffic or datagram), guaranteeing them the maintenance of the required service class even in conditions of strong attenuation. signal due to adverse weather conditions. Typically the stream data are generated by video and / or voice applications, while the datagram ones are generated by applications such as file transfer, archive searches, any interactive connection and so on. Street.

The fade countermeasure technique adopted by the system? also made in TDMA. It? particularly useful, if not essential, when transmissions occur in frequency bands above 14 Ghz, such as, for example, in the Ka band (20/30 Ghz), a band on which all satellite transmissions will have to migrate due to saturation Ku band (12/14 Ghz) and C band (6/8 Ghz).

The adopted fade countermeasure technique operates on three fronts: a) uplink power control, if a station has the necessary equipment; b) the redundancy of the data obtained by increasing their coding, c) the redundancy of the data obtained by decreasing their speed? transmission. The latter possibility? it is adopted only for very deep levels of signal degradation, in case points a) and b) are not sufficient. While point a) can? be present or not in a station, depending on the equipment supplied, points b) and c) are made via software by the FODA / IBEA system. Both the data encoding and their speed? transmission rates can be dynamically varied according to the extent of the signal attenuation of the individual ground stations. The speeds usable transmission rates are 1, 2, 4 and B Mbit / s, while the data encodings can be 1/2, 2/3, 4/5 and, of course, uncoded. The weather ? divided into intervals of fixed length equal to 20 ms (time frame or, more simply, frame) in which the single transmissions take place based on the received time allocations. In each frame, a control station (master station) assigns to the other stations (slave stations) their respective transmission times (transmission windows) according to the modality? linked to requests made by stations. The master station itself, in addition to the control function of the common resource "channel", behaves like a normal slave station, forwarding to itself regular requests for its data transmission. Within a transmission window assigned to a slave station, data is generally transmitted to several receiving stations (data bursts). A burst can? be composed of n sub-bursts. A sub-burst gathers the data addressed by a transmitting station to a certain destination station. Within the same burst, the individual sub-bursts are transmitted by choosing the speed? transmission and data encoding according to the signal fade conditions of both the transmitting and receiving stations. There? it means that, within the same transmission from a station A to n destination stations, the fade conditions of the individual stations are taken into account in order to ensure that the transmitted data is received while maintaining the characteristics of BER (Bit Error Rate) requests for individual transmissions.

The purpose of the present invention? to provide a time-division access method that allows simultaneous transmission over satellite of both synchronous and asynchronous data, while maintaining the quality of the data. of the service required by individual applications even in case of strong attenuation of the transmission signal due to bad atmospheric conditions (fade countermeasure) The invention finds its natural field of application as a tool for video-conferencing via satellite, not only point to point but among a number N of stations scattered everywhere in the coverage eye of the satellite used. In parallel to the video-conference, the active stations can use the assigned band to transmit data to several? low priority (datagram data).

Real applications can be tele-medicine and tele-education. In the first case, remote operations can be followed in direct and interactive way by pi? interconnected medical centers, together with the transmission of patient archive data (x-ray images, ultrasound scans, medical history files, etc.). In the second case, lectures held at a distance at a didactic center (be it a university or other) can be followed interactively by students scattered in different centers, together with the simultaneous transmission of didactic material.

In both cases, the quality? of broadcasts? guaranteed regardless of the weather conditions of the moment in which the transmissions take place.

The primary problem faced by the invention concerns the adopted fade-countermeasure technique.

A satellite communication system can? be rendered inefficient due to the attenuation of the signal by the atmosphere. In general, these effects are worse at pi? high frequencies since? signal energy absorbed by rain and water vapor? directly proportional to the increase in the frequency at which transmissions take place.

In the C band {6/8 Ghz) the attenuation of rain is typically 0.2 dB on both the up and down links. On rare occasions it can? get to 2 dB, when a particularly heavy rain can? create greater attenuation. Although the phenomenon of signal attenuation (fading) is present in both the C-band and the Ku-band, how much attenuation is achieved (fade)? small enough to be incorporated as a fixed link margin.

A pi? high frequencies, such as the Ka band (20/30 Ghz), the attenuation of rain varies over a much longer range? wide (even more than 15 dB). A good satellite transmission system, designed for these frequencies, must consider the attenuation problem in such a way as to take advantage of periods of clear sky and be able, at the same time, to cope with short periods of heavy rain in such a way as to not degrade, or degrade too much, system performance. For example, the fading of the Ka band for about 99% of the year? less than 2 dB in most of Europe. There? what then does it need? a technique that increases the reliability? of a link when? necessary, without requiring a link margin that is too high (and therefore expensive) and which takes into account the optimization of an expensive resource which? the band on satellite.

Without prejudice to the need to have a satellite transmission system for the Ka band, given the saturation of the lower bands, a first way to counterbalance the attenuation of the signal consists in the up-link power control, that is? in using pi? transmissive power when? there is an attenuation of the transmission signal. The station, of course, must be equipped with such equipment. Does the use of up-link power control imply that the station is not operating at maximum efficiency and therefore the link? pi? sensitive to fade on the down-link. Assuming that a station is equipped with it, the up-link power control alone is not? sufficient to counter fade levels higher than a few dB; moreover, its disadvantage consists in the fact that, in clear sky conditions, the quality? of the up-link? worse than that in a fixed power system with the same maximum output power, due to the dynamic variation of the power control scheme. For all these considerations, ne? derived the choice of combining a possible up-link power control with another fade countermeasure technique. The choice to implement the fade countermeasure technique in TDMA? was dictated by two fundamental considerations. The first consideration? that the adoption of other fade countermeasure techniques, such as eg. site diversity (with different places) or frequency diversity (with different frequencies), are not feasible, except with enormous and unjustifiable financial expenses, if we take into account that, in our latitudes, the percentage of days with bad weather conditions? minimum in the year. The second consideration, more? technique, was dictated by the challenge that a TDMA realization, which one was in mind, would have entailed, since? did not exist on the market (at the time from the beginning of the project, that is, in 1990, nor does it exist today) a modem capable of managing with the desired dynamism? data transmission. The realization of the FODA / IBEA system has therefore also involved the drafting of the hardware specifications of the speed modem? variable on a sub-burst basis. That modem, together with the satellite channel access controller (TDMA controller),? was made in England

DESCRIPTION OF THE CURRENT STATE OF THE ART

The comparison that pi? evidently leaps to the eye (also by analogy of name)? between the FODA-TDMA system and the FODA / IBEA-TDMA system. Although the second is the logical evolution of the first, the differences between the two systems are substantial. Although both were designed to support the simultaneous transmission of two types of traffic (stream and datagram) with very different characteristics and requirements, the FODA / TDMA system was designed to operate in C or Ku band; therefore it did not adopt any fade countermeasure technique. FODA / TDMA provided for data transmissions at high speed? fixed (2 Mbit / s) and with fixed coding (1/2 coding). In addition, therefore, to a pi? simple architecture of the access method, the satellite channel access controller (TDMA controller), on which the FODA software was installed, provided a modem and a speed codec? and fixed coding, respectively, as it was possible to find on the market.

As for the FODA / TDMA system, you can? refer to the following documentation:

[1] Italian Patent No. 1233264, filed on March 21, 1989;

"The FODA-TDMA satellite accesa scheme: presentation, study of th? System and results", IEEE Transaction on Communications, Vol. 39, N.12, pp. 1823-1831, December 1991.

As for the fade countermeasure technique adopted in FODA / IBEA, c '? to say that it constitutes an absolute innovation since? just one of the pre-existing techniques? made in TDMA but, as we shall see, it does not appear to have the efficiency reached by the FODA / IBEA system. The choice adopted offers the maximum flexibility? at the lowest cost, being essentially feasible via software. The reasons that led the applicant to choose the TDMA for the realization of the fade countermeasure technique are already? been listed above. Particularly attractive was the definition of the specifications that the modem had to have. The adopted technique is based on dynamic data redundancy (adaptive TDMA fade countermeasure): pi? high? the attenuation of the transmission signal pi? the data is redundant. Redundancy? implemented by increasing the data encoding and / or reducing the speed? transmission, in accordance with tables that link the fade level of the transmission signal with the choice of the data redundancy factor. The fade level of the signal generally changes in the order of seconds; moreover, only the transmissions to / from stations in fade are redundant (otherwise precious channel bandwidth would be uselessly wasted), leaving the transmissions unaffected by adverse weather conditions. Therefore ? it was necessary to give the specifications of a modem capable of dynamically changing the speed? data on a sub-burst basis, and this, to date, has no previous implementation. Also, maximum bandwidth usage on satellite? has always been held in the utmost consideration, never leaving bandwidth unused.

Below we give an overview of the other existing fade countermeasure techniques.

a) Site diversity

The sity diversity (or space diversity) provides for the duplication of each ground station a few km away (about 10 km) from the respective "mother" station. When the attenuation due to rain exceeds a threshold value on a certain link, the transmissions are diverted to the alternative station where, due to the variability? of the intensity? of the rain, the probability? to have more? links in fade simultaneously? lower (see [3] Julius Goldhirsh: "Space diversity performance prediction for earth satellite paths using radar modeling techniques", Radio Science, Vol. No. 17, Number 6, pages 1400-1410, November-December 1982). Such a technique? obviously extremely expensive, since? all ground station equipment must be duplicated. Furthermore, this cost is not? justified in those sites where the days of good weather predominate in the year.

b) Frequency diversity

It is a system where various links, which use frequencies well above 10 Ghz, are designed in such a way as to withstand only moderate rainfall. When the rain becomes excessive on a particular link, that link? varied on a frequency pi? low. Such a technique? very powerful but has two major disadvantages: a)? band request in the frequencies 6/8 Ghz, currently no more? available for problems of total saturation, b) all stations must be equipped with a duplicate of the communication equipment for the frequencies pi? low.

See :

[4] Katsuhiko Kosaka: "Frequency Switching TDMA System for countermeasure against rainfall attenuation", Journal of Radio Research Laboratories, Voi 25, Nos. 117/118, July / Nov. 1978, pp. 117-132.

[5] Carassa F., Tartara G., Matricciani E .: "Frequency diversity and its applications", International Journal of Satellite Communications, Vol. 6, pp. 313-322, Vol. 6, July-September 1988.)

c) Resource sharing: Burst length control

The burst length control technique is based on the adaptive use of the up-link power control combined with an adaptive control of the encoding of the data to be transmitted. In each time frame is a certain amount kept free from assignments? of reserved band, for assignment upon request, to those stations that detect a fade of the transmission signal. This stock of resources is not? used in non-fade periods, resulting in less than optimal bandwidth usage. The stations that detect a signal fade use the extra band to repeat their bursts for a certain number of times, with a proportional increase in the energy of the burst. At the simple increase of energy obtained by repeating H times the message, some coding of the data can? be added, with further gain.

However, this method does not provide for any energy gain obtained by reducing the speed? transmission, can? compensate only partially for signal degradation, as well as wasting bandwidth in the predominant cases of non-attenuation of the signal.

See :

[6] Anthony S. Acampora: "The use of th? Resource sharing and coding to increase th? Capacity of digital satellites", IEEE Journal On Selected Areas In Communications, Vol. JSAC-1, No. 1, January 1983.

[7] "Adaptive Methods to counter rain attenuations effects in th? 20/30 GHz band", Space Comm. And Broadcasting, N.2, pp. 253-269.

d) Fade Spreading

This technique operates by mechanically reducing the speed? of data transmission. The data is combined with a pseudorandom code (PRC), known as a "spreading function". The combinatorial process spreads the data signal over the PRC band. The signal so? "manipulated" is transmitted to the recipient where? recombined in the inverse function with the same PRC and a low-pass filter recomposes the signal. In reality, various effects (such as the interference between symbols, the intermodulation product and the degradation of the implementation of the non-ideal modem and of the filters) concur to reduce the theoretical gain obtainable with the use of such equipment. The efficiency of that system is not? high: ideally c '? loss, but when the effects of modem performance at low S / N ratio values are considered, the method becomes less efficient. Compared to other coding schemes that have a net gain and operate, like this one, in fixed band, this method is less attractive.

See :

[8] Carassa F., Tartara, G., Matricciani E .: "Frequency diversity and its applications", International Journal of satellite Communications, Vol 6, pp. 313-322, Vol. 6, July-September 1988.

[9] "Fade countermeasures for satellite Communications", ESA STM-235, May 1986.

The present invention? defined by the characterizing part of the attached claims and by the following description.

Will the invention come? now described with reference to the attached figures, in which:

Figure 1 schematically shows in an events / time diagram the behavior of a typical transmission channel which operates with the system according to the present invention;

- figures 2a, 2b, 2c show event diagrams relating to the software implemented in the FODA / IBEA system according to the invention;

- figure 3 shows the "scenario" of a communication network system at the level of the earth station;

Figure 4 schematically shows the hardware configuration of the system according to the invention;

Figure 5 shows the allocation mechanism for stream traffic;

- figure 6 shows the "data burst" format of the FODA / IBEA system;

- figure 7 shows the format of the data sub-burst of the system; And

Figure 8 shows a state diagram of the system start-up.

With reference to Figure 1, in this? shown the trend of the situations that can occur in a satellite transmission channel in the presence of weather fading, as a function of time, in a possible sequence of events.

In line I? shown the amount of weather fading MF in the face of an intervention threshold ST, under which not? intervention of the system according to the invention is required.

Once the ST threshold has been exceeded by the MF fading, a fact marked as event E1 on line VI, the system intervenes, as will be seen.

On line II,? Bit Error Rate is shown, normalized to zero before the E1 event. This BER would increase, as shown in correspondence with events E1-E2, E2-E3, etc., without the intervention of the system.

Clearly, between the E4-E5 events, where there is a total loss of the possibility? communication (Outage) the Bit Error Rate goes to infinity anyway.

In row III? showing the intervention of a possible power control, normalized to zero in the absence of significant weather fading.

In row IV? shown queuing of datagram data, normalized to zero in the absence of significant fading, where? to note the increase in stock in correspondence of the outage (between the events E4-E5), and the queue of decrease at the resumption of operations? of the transmission channel, after event E5. With each increase in fading, which can be managed by the system, there is a temporary increase in the stock, due to the delay between the greater demand for band and the allocation of this request.

On line V? shown the energy per bit of the data, normalized to zero before the event El, and increased by the system according to the invention due to the intervention of one or more? countermeasures (power control, coding variation, data speed variation) applied by the system to compensate, as far as possible, the effects of the weather fading.

How can the software of the FODA / IBEA system be deduced from the previous exposition in terms of "events"? organized at events. There? does that normally mean it? waiting for an event to complete in order to take actions related to the event in question. The events are divided into three main categories:

a) events caused by the arrival of a packet from the Ethernet network.

This is usually a stream or datagram data packet that must be sent. Occasionally it can? be a packet containing a request / release of bandwidth for stream data (stream request / release), or a control packet, or a response from the application to a request (made by the system) for momentary data compression ( only for compressible application) or for a return to normal? after a compression (expansion).

b) events caused by the arrival of data from satellite The data received from satellite are divided into three groups: 1) reference burst. His arrival involves a long series of operations since? the Ref. Burst contains a lot of information, such as the list of active stations, the fade level of the individual stations and the stream and datagram allocations of the individual stations; 2) control sub-burst. This ? the part of a burst that precedes the actual data. The CSB of each burst also contains the requests for stream and datagram allocation; 3) receipt of the actual data (subsequent to the CSB).

c) events generated entirely.

Are listed the most? significant events generated internally in the system. 1) Suffering timer: refers to the timeout of a timer that is started every time a stream application works compressed or out of BER specification (ie with a BER higher than its specification). The first case (compression) only happens to those compressible applications, compressed by the system as a last resort in case the fade level is so? high that the maximum applicable redundancy expands the data so much that it does not fit more? in the frame. As a last resort, before automatically cutting certain transmissions, compressible applications in band are compressed. It is clear that in such extreme conditions the BER required by the application is not? pi? guaranteed. Applications will be re-expanded as soon as the fade level decreases just enough to allow it. The second case (out of specification) also includes the first. It refers to all those applications that are unable to compensate the fade level (too high) even with the maximum redundancy, and therefore are transmitted to a BER pi? high. When the timer expires, distressed applications try to re-submit a bandwidth request appropriate to their specification needs.

2) Datagram timer: it expires every 5 frames and serves to update the bandwidth request for datagram applications. Such a request, as already? said, ? calculated on the backlog pi? incoming traffic, the latter averaged over 5 frames. Whenever c '? a dialogue between the system and an application (compression request, expansion request, momentary suspension of the sending of datagram data due to congestion, recovery from this situation, etc.) the system waits for a confirmation response (ack) from the application in question. If at the expiration of this timer (timer ack) the answer is not? yet arrived, the application? deleted.

With "start of transmission time" we mean the opening of the transmission window, that is? the beginning of the time allotted to a certain station for its broadcasts. Similarly, "start time of reception" indicates the opening of the reception window, which does not? always open for the whole frame to avoid the risk of false reception due to noise interpreted as valid data.

Figures 2a, 2b and 2c show, very schematically and macroscopically, the actions taken by the system when one of these three events occurs. As part of an event, eg. the arrival of an Ethernet packet (fig. 2a), the various types of packets are distinguished (datagram, stream, stream request, stream allocation release, control) and, for each of them, the actions performed are indicated. In the figures, the term "controls" refers to the fact that validity is always checked. of the package, its correct format, etc.

Exposed otherwise, one of the main problems for communications satellites operating at frequencies above 10 GHz? the high level of attenuation encountered due to rain.

The attenuation due to rain and the depolarization of a signal occur due to the fact that the individual raindrops absorb energy from the radio waves and the fact that a certain amount? of energy in the waves is diffused out of the propagation path. These interactions depend on the number of raindrops encountered and the distribution of their size and shape.

At frequencies above 10 GHz, rain attenuation has a significant effect on operation. of microwave links. Numerous techniques have been proposed to alleviate this problem and these are known under the generic name "fade countermeasures".

If the speed? (rate) of information of a satellite link is allowed to decrease during an attenuation, can you? use an adaptive FEC (Forv / ard Error Correction) technique as "fade countermeasure". Obviously you can? use adaptive coding only on appropriate links that can support a capacity? of reduced transmission when attenuation occurs.

When the FEC is used, during an attenuation, the speed? of information is reduced and additional coding information is inserted into the channel, while the speed? of the channel remains constant. A net gain is realized with the coding scheme, depending on the coding used. Typical values are between 2 and 6 dB for encodings of 7/8 and 1/2 respectively. "Punctured" codes represent an alternative (and much cheaper) solution than using optimized codes for each encoding.

The technique used by the Applicant is based on a convolution encoder with K = 7 and a Viterbi decoder which, given a frequency of coded bits, performs a maximum probability sequence estimate. to predict the bit sequence of the original information. Other encodings can be derived from 1/2 encoding by erasing (or "puncturing") bits periodically in the encoded sequence and inserting erasures on the decoder to produce a 1/2 encoded sequence. In this way, it is possible to produce other fractional encodings, based on the 1/2 encoding and making use of a 1/2 decoder with the possibility? on the input to decode the various cadence codes. High speeds data are used when not? present atmospheric attenuation, when the signal-to-noise ratio? sufficiently high. Even the speed? of the data is progressively reduced when there is a strong attenuation, so that? the decoder operates with a correct value of the Eb / N0 ratio (bit energy over noise density) and to allow the modem to acquire the data in a reasonable time interval. The fundamental principle used in the system in question to treat different levels of signal attenuation consists in the variation of the energy contained in a bit of information. There? is carried out first by varying the transmission power, when possible, and then the speed? of data encoding and their speed? transmission.

1. THE NETWORK SCENARIO

The applications for the end user are supposed to run on Hosts connected to a LAN (Local Area Network). Several LANs are interconnected via satellite. Access to the satellite? obtained by means of the FODA / IBEA-TDMA software which runs on the hardware developed by MARCONI R.C .. The hardware? consisting of a TDMA satellite controller, with a variable codec codec and a speed modem? variable transmission. (Figure 3)

2. THE TRAFFIC

FODA / IBEA manages two basic types of traffic: synchronous traffic ("stream" data) and non-synchronous traffic ("datagram" data). Applications that use the FODA / IBEA system must declare whether their traffic? of the type "stream" or "datagram".

By stream traffic, or synchronous, we mean data generated by applications such as voice, slow scan television, video conferencing, which are all characterized by a speed? of constant packet arrival. These applications typically require a short and fairly constant delay, cannot tolerate out-of-order packet delivery, but can tolerate occasional bit errors and missing packets. In practice, the stream traffic requires a quantity? fixed bandwidth, assigned on a basis as much as more? regular as possible, and the satellite network should maintain a low and constant delay on the arrival of information. Traffic of the "datagram" type, that is asynchronous,? divided into "bulk" and "interactive" traffic. Bulk traffic? intended for applications such as "file" transfer. Typically, a large number of packets are transmitted and the speed? with which packets are sent and the delay introduced by network crossings are not critical constraints. Generally, such traffic does not have the stringent delay requirement as does voice. Generally, out-of-sequence packet deliveries are allowed. However, particularly on a high delay network which one? the satellite network, the terminal yield of such traffic can? be heavily penalized by bit errors or packet losses. Interactive traffic? intended for terminal accessing computers, database queries, operator message exchange, and interactive traffic. This type of traffic typically requires reliable, error-free delivery and short delays to ensure acceptable response times. Often ? consisting of short messages (a few characters) with arrival times between one packet and another that are not predictable.

3. THE HARDWARE

The hardware provided by Marconi R.C. ? consisting of a TDMA controller, including a variable codec codec and a speed modem? variable transmission.

3.1 The TDMA controller

The controller in TDMA? divided into two parts (Figure 4): the upward controller (UC), dedicated to transmitting to the satellite, and the downward controller (DC), dedicated to receiving data from the satellite. The transmission control processor hardware? consisting of a Motorola MVME147S-1 processor board and an MVM712M transition module. The processor card employs a 68030 card comprising 4 megabytes of RAM. This runs at 25 MHz and occupies a 4E wide VME BUS slot. The operating system provided by Motorola on the card? called MVME147 BUG. Does the transmission processor board act as a platform for the software that controls and monitors the functions of the unit? transmission control. This includes, through the transition module, the VT100 operator console, SCSI, ETHERNET and RS232 interfaces.

The transmission serial interface card? a double EUROCARD card compatible with the VME BUS. It acts as a buffer between the VME BUS and the three terrestrial interface ports of the transmission controller. The three doors are: a single high speed door RS449 / RS422 (high speed serial interface, at 384 K bit / s) and two CCITT ports type G.703 (low speed serial interfaces, at 64 K bit / s).

The transmission modem interface? controlled by the transmission control processor through a number of memory mapped ports that allow both reading and writing of control data. The circuit also provides the interface to the modem for data transmission and failure monitoring. The interface function? carried out by two cards: the modem transmission interface and the line interface.

The modem interface card for the transmission network? a double EUROCARD card compatible with VME Bus (identity Y-35-9041) 8 W. It contains all the channel coding, framing and speed selection functions. symbols. It analyzes the data coming to it from the VME bus and also provides status information for circuit and modem surveillance purposes.

Is the data routed or to the data FIFO, which can? store up to 512 bytes of data and control information, or to an event memory which? used to synchronize the transmissions of the bursts with the transmission frame. The control information is output to the data FIFO before the data which must have the channel coding. This control information sets the channel coding hardware appropriately. The data is sent coded to the FIFO, via the 32-bit VME data bus. Updating the FIFO of data? controlled by an interrupt that occurs when there? a transition of the flag indicating "half-full data FIFO".

The transmission modem interface processes data which is presented to it by the 32-bit data bus. All data transfers are 32-bit word transfers between the transmission modem interface and the processor. The data FIFO defines the content and timing of a burst relative to the start of the burst. The exact time at which the bursts are sent to the modulator? determined by the transmission event memory (TEM). This ? consisting of two memory matrices organized in "ping pong" mode, each capable of storing up to 512 events. During a frame, the software can? update one side of the TEM, while the other side is employed by the hardware. The events originating from the software are then used by the hardware in the next frame. Events are programmed into the modem interface through a frame event memory which is addressed by the event number within the frame. New events are programmed in offline memory during the current frame and are triggered in the next frame. The update must be done before the end of the current frame. Each event within the event memory contains an 18-bit time code, which defines the time at which it will take place. place a particular action (in 244 ns ticks from the start of the transmission frame), and a 14-bit function code. Transmission events are used, among other things, to open and close the burst gate and to set the output level of the modulator. Can you? employing a special function code to define the end of the transmission frame by providing a reset pulse to the transmission frame counter. When the transmit burst gate is opened, formatted data is sent from the CPU to the transmit modem interface under interrupt control. These data are then coded with the FEC technique, and if necessary they can be "scrambled". The speeds of FECs available are all based on conventional 1/2 convolutional encoding. The puncturing logic allows for additional 2/3 and 4/5 encodings. The hardware hangs the queue of bits needed to download the Viterbi decoder on the receiving side and, if necessary, appender? an 8-bit CRC to each DSB {data sub-burst) that is transmitted. the line interface card? a double sized Eurocard printed circuit board (identity Y 35-9067). It contains the line receivers / drivers to interface the modem and also provides the reference clock at 16.384 MHz, obtained from a 32.768 MHz oven controlled crystal oscillator, placed on the line interface card.

The processor hardware of the receiving controller? consisting of a Motorola MVME147S-1 processor board and a MVME712M transition module. The processor board uses a 68030 board containing 4 Megabyte RAM. It runs at 25 MHz and occupies a 4 E wide VME bus slot. The operating system provided by Motorola on the card? named MVME147 BUG. The Receive Control Processor Board acts as a platform for the Receive Control Processor software that controls and monitors the functions of the unit. reception control. This includes, through the transition module, the SCSI, ETHERNET and RS232 interfaces.

The receiving serial interface card? a double Eurocard card compatible with the VME bus (identity Y-35-9053). It acts as an interface between the VME bus and the three terrestrial interface ports of the receiving controller. The three doors are: a single high speed door RS449 / RS422 (high speed serial interface, at 384 K bit / s) and two CCITTG703 ports (low speed serial interfaces, at 64 K bit / s).

To avoid contention, the reading and writing of memories is organized in such a way that while met? memory is written, the other is read. There? ? known as "ping-pong" operation. Only half? of the memory that does not? in reading it appears in the address space of the VME. The interrupt service routine, what is it like? started, determine which buffer? empty (reading the status word) and then fills the first half? of the memory through the Logic Cell Array (LCA) of the VME interface. During this time, the second half? memory is read through the LCA serial interface.

Data is transferred between the receiving control processor and the receiving serial interface card through the VME data transfer bus (DTB). A data transfer process is initiated by the receive serial interface card when any of the three dual port memories becomes half full.

The receiving modem interface and the channel decoder are two double Eurocard cards compatible with the VME bus (identities Y-35-9051 and Y-35-9052). Together they operate as an interface between the VME bus and the modem control lines and perform all channel decoding functions. In the receiving controller, the 4 "soft" decision bits, received by the demodulator, are passed to the FEC decoder where they are decoded, if necessary. Does the receiving hardware also provide a measure of quality? estimated channel for each DSB by monitoring the soft-decision bits of the correct data, that is? previously affected by error and then corrected. The hardware simply observes the occurrence of particular soft-decision levels and averages these over a sub-burst. Assuming a Gaussian distribution, one can? then calculate an approximation on the short-term BER. This method? much more? quick than to calculate the errors in the UW (Unique Word) and provides an estimate of the channel that can? be used in adaptive mitigation countermeasures algorithms.

The receiving modem interface? a processor controlled through a number of memory mapped gates which allow both reading and writing of status, control and data information between the processor and the hardware. The circuit also provides the interface to the modem for data reception and failure monitoring. The interface function is performed with three cards: the interface to the receiving modem, the channel decoder and a line interface card.

The receiving modem interface and the channel decoder are double Eurocard VME-bus compatible cards, 8 E wide. They contain all the channel decoding, framing and speed identification functions. of symbol. The receiving modem interface processes data from the 32-bit VME data bus and also provides status information for circuit monitoring purposes. The decoded data is routed to the receiving modem interface. This interface uses a receive event memory (REM) to enable real-time reception and decoding of received data bursts. REM? identical in structure to the TEM. The receive function code contains information, among other things, about the initial encoding and speed? of burst symbol within the frame.

Can you? employ a special function code to generate a transmit frame pulse that will reset? the event counter on the broadcast controller. This is used by the slave stations to synchronize their transmission frames to that of the master station. Events are programmed into the modem interface via a frame event memory which is addressed by the event number within the frame. New events are programmed in offline memory during the current frame and are triggered in the next frame. The update must be done before the end of the current frame.

The line interface card? a double Eurocard printed circuit (identity Y-35-9037). It contains the line receivers / drivers to interface the modem and also provides the reference clock at 16.384 MHz.

The codec supports variable encoding: 1/2, 2/3, 4/5 and uncoded PSK. Dotted codes are adopted, obtained from the convolutional encoder 1/2. The decoder operates asynchronously with speed? of information bits at 8 Mbit / s with 3 soft-decision bits on sign / size. The software of the FODA IBEA system guarantees error rates better than 10 <-4>, 10 <-6> and 10 <-8>, depending on the class of service required by the various applications, adapting the coding to the signal / noise conditions .

3.2 Modem

Does the modem supplied by Marconi RC support speed? of variable symbols: 1/2, 1,2 and 4 Msymbols / s, switchable on a sub burst basis, with two different modulation schemes (BPSK and QPSK) to allow operations within the 5 MHz band. transmission rate is variable in the interval 0.5-4 Msymbols / s.

The modem? consisting of the following units:

- Y-35-9030 A 19''X6U rack including a real PK110 power supply.

-? -42-1504 A 21HP amplitude IF converter module.

-? -35-9031 Analog interface.

-? -35-9032 FIR modulator.

-? -35-9038 Carrier evaluation device.

-? -35-9033 Timing evaluation device.

-? -35-9037 Clock generator / line interface.

A special feature of the modem is the capacity? to dynamically adjust its speed? transmission within a data burst. There? allows the individual DSBs of a DB to have different speeds? transmission (and consequently, different bit energy), according to the needs. The speeds available transmission rates are 512, 1024, 2048 and 4096 KBaud, with BPSK or QPSK modulations. Is therefore a range of speeds available for the system? bit transmission rate of 512-8192 Kbit / s. In order to meet the speed requirements? dynamic, the modem? realized using digital signal processing (DSP) techniques.

3.3 Hardware Limitations

This hardware allows the realization of a wide range of TOMA systems made via software, but imposes some limitations. These were necessary because, in some cases, the software would not have been able to react quickly enough? to external events.

Each burst must begin with a sequence of timing and carrier recovery bits (CBTRS). This ? in effect a limitation of the modem.

The frame must include a reference burst (RB) containing the "unique word" of the master (MUW). The MUW maintains the synchronization of the receiving hardware within the frame. Aside from the MUW, the remaining content of the RB is not? important for hardware. Only one MUW must appear for each frame. The maximum frame length (allowed by the hardware)? 64 milliseconds.

Each data burst must consist of a control sub-burst (CSB) followed by a number, possibly zero, of data sub-bursts (DSBs). The CSB is decoded by the hardware and provides the information necessary for the reception of subsequent DSBs.

4. THE FODA / IBEA SATELLITE ACCESS SCHEME

The design of the FODA / IBEA-TDMA satellite channel access scheme? based on a centralized control performed by a control station (master) that can? also operate as a "slave" station.

The other active stations in the satellite network operate as slaves. AND? responsibility of the master station to govern the organic functioning of the network, controlling the resources and maintaining the synchronization of the "slave" stations. Although there is only one master station at a time in the network operation time, all controllers are capable of operating as "master".

The master:

to. allocates time within the TDMA frame for the stream and datagram transmissions of the slaves;

b. maintains synchronization of all stations on the network;

c. maintains the optimal level of transmission power which is used by all slaves as a reference level;

d. it uses the frame space allocated to itself to send data over the network in the same way as a slave station.

The slave:

to. adjust the transmission power level taking the master as a reference ,?

b. sends allocation requests to the master for stream and / or datagram transmissions;

c. it uses the frame space allocated to it by the master to send data over the satellite network;

d. choose the encoding and speed? transmission of all data to be sent, according to the attenuation level of each link, asking the master for the correct allocation;

And. spreads its fade level.

The master operates on a signaling channel within the TDMA frame. The slots for data transmissions to the satellite are assigned upon request, following the criteria discussed in paragraph 4.3 below.

The master station can? be replaced, in case of failure, by any traffic station among those declared available to take on this task. At any time, a traffic station can? decide to give up becoming the new master. The master spreads the physical address of the station from which it will come? replaced in case of failure. The successor ? the first active station (which has declared its availability) in the station list maintained by the master. A station can? refuse, at any time, the possibility? to become a new master. The current master communicates this condition in the state associated with that traffic station and, if necessary, the next station available to become the master is selected as successor.

4.1 The FODA / IBEA frame

Will come the term "stream sub-frame" is used to indicate the part of the frame dedicated to containing all stream allocations. Similarly, "datagram sub-frame" means the portion of the frame dedicated to holding all datagram allocations.

The sub-frame stream extends to a Normal Stream Upper Boundary (NSUB) upper limit. If not ? present traffic stream, the datagram sub-frame can? practically occupy the entire fraine. When is there? attenuation of the transmission signal, the stream transmission is privileged and space is sought in the frame to allow the broadening of the stream data by increasing, if necessary, the upper limit of the stream sub frame up to an ESUB (Extended Stream Upper Boundary) value. The sub-frame datagram is compressed as a result. The enlargement of the data? due to the choice of an encoding pi? redundant (among those supported) to counteract the attenuation in progress. If the encoding? insufficient, the speed? data transmission is reduced in octave steps.

The FODA / IBEA frame has a length of 20 milliseconds. Each frame contains:

-) a reference burst, sent by the master station in order to spread attenuation levels and time allocations for stream and datagram transmissions. The reference burst is always sent with coding and speed? fixed. The values of 2 Mbit / s and 2/3 coding were chosen.

-) numerous transmission windows to send stream and / or datagram data. Generally, the trend? that of setting only one transmission window to send both stream data and datagram data (see section 4.4). The part of each transmission window dedicated to stream data remains unchanged (once assigned), while the part dedicated to datagram data can? change in size in each frame. Consequently, the transmission window (s) must be recalculated by the master station for each frame.

In each frame a small number of control slots are assigned (transmission windows of fixed size, dedicated to the transmission of short data, such as request updates or other control messages) by the master station on a "round robin" pattern. among all active stations that have not received any assignments in the current assignment. The number of control slots? been fixed to 1 control slot when the system? calibrated to operate up to 8 ground stations, 2 for 16 ground stations, 3 for 24 ground stations and so on? Street. The control slot? assigned at 1 Mbit / s. The coding? 2/3 like all other control sub-bursts. In the current implementation,? assigned only one control slot, whatever the number of stations. The position in the frame of the control slot is not? fixed. If all stations have an allocation in one frame, the space dedicated to the control slots is added to the remaining space in the sub-frame datagram.

-) a first access slot (FAS) every 32 frames. This slot? dedicated to allowing a new station to join the network. It is used in contention between all the stations that want to become active. In the frame where? FAS slot is present, control slots are not allocated. The FAS? sized as a control slot pi? an uncertainty (currently fixed at? 150 microseconds) due to the current position of the satellite relative to its nominal position. The FAS has a fixed position within the frame, just before the end of the frame. Is access to the FAS with speed? 1 Mbit / s and with 2/3 coding.

4.2 Classes of service and redundancy factors

The data to be transferred between two applications running on two different satellite-connected LANs must usually meet certain requirements associated with quality. of the requested service, such as:

the typical end-to-end delay of the data (ie the delay between their generation and their final reception);

- maximum packet arrival time jitter; - the BER (error rate) required.

In the system in question, the first two parameters are not negotiable. In the case of stream data, the delay depends on the round trip time (RTT) pi? a fraction of the frame (approximately 300 milliseconds in total). The maximum "jitter" of the arrival times of packets leaving FODA / IBEA can? be assumed to be equal to twice the length of a frame pi? the maximum temporal jitter of packets entering the system.

In the case of datagrams, both delay and jitter depend heavily on the traffic conditions of the overall system and the saturation control mechanism. These parameters cannot be specified by the application.

The BER, on the other hand, must be specified by the application to choose the quality? of the communication channel. The required BER interval is specified by means of a value called "class of service (COS)". There are four classes of service, as shown in table 2:

Table 2

When signal attenuation occurs due to unfavorable atmospheric conditions, the BER specified by the transmitting application must be maintained using the coding and speed. of transmission pi? convenient. The operating value of Eb / N0 must be sufficiently high to allow a conveniently high probability? acquisition of the modem. There? translates into the probability? of UW revelation.

The information redundancy factor. Ri,? the ratio between the energy per bit of information of the data sent with a certain encoding and speed? and the energy per bit of information of the same data sent uncoded at 8 Mbit / s.

Ri pu? be expressed as the product between the coding redundancy RC and the redundancy Rs. Re expresses the expansion of data due to the increase in coding and the possible reduction in speed? transmission. When should you compensate for increased signal attenuation (in order to maintain the same BER value for that type of data),? it is preferable to first increase the encoding (taking advantage of the encoding gain) rather than reduce the speed? transmission. The latter must be halved only when Eb / N0 falls below the value necessary for the acquisition of the bursts.

The attenuation level represents the number of dB of C / No available (carrier power with respect to the noise density) under the theoretical value of 81 dB taken as an attenuation level equal to 0. When operating with 0 attenuation at speed? of 8 Mbit / s, the Eb / N0 ratio? of 12 dB. This value of Eb / N0 allows the reception of uncoded data with a BER lower than the best combinations of encoding and speed. transmission were chosen considering the BER requirements of the four classes of services. These combinations are shown in table 3, considering 4 dB (5 dB actually measured on the channel, minus IdB of margin of implementation of the modem) as the minimum value of Eb / No necessary for the acquisition of the modem.

Table 2

Minimum value of Eb / N0 for modem acquisition = 4 d B.

where:

F [dB] attenuation level expressed in dB Q [dB] quality? of the channel (Eb / N0) expressed in dB M bps speed? of data transmission expressed in Mbit / s

Rs speed redundancy factor? (8 / current speed)

R-ck coding redundancy factor for service class #k (l / current coding used)

Rik redundancy information factor for service class #k (Rs * Rc)

If the minimum value of Eb / No necessary for the acquisition of the modem? set at 6 dB (7 dB actually measured on the channel, -1 dB of modem implementation margin), table 3 becomes table 4:

Table 4

Minimum value of Eb / No for modem acquisition = 6 d B.

The fixing of the Eb / N0 ratio necessary for the acquisition of the modem at 6 dB makes coding 1/2 not used, how? evident from table 4.

4.3 Requests and Assignments

Once active, a station can? make a request for a stream transmission window (stream channels) and for a datagram transmission window. Normally the requests are "piggy-backed" with the data in the transmission windows. If a broadcast window is not assigned to the station, it can? use the control slot, when assigned to it by the master.

4.3.1 Stream

The stream request is forwarded by a slave station only once, and, if accepted, is considered valid until an update request or an explicit abandon request is sent. The request is made as a multiple of the 32-bit word (channel word).

Each stream application sends to the FODA / IBEA system (to which it is connected) a request indicating the number of stream channels requested Cr, the minimum acceptable number of stream channels assigned Cm (which can be set equal to Cr) and the class service request, expressed by means of a number ranging from 1 to 4. The request to the FODA / IBEA system is made by issuing a command of the GAFO protocol.

The station software assembles all the incoming Cr values together and sends a global request to the master, the sum of all the Cr received. The request also includes the calculated transmission "overhead". The master station fulfills the stream requests received from active stations until the NSUB value is reached in the sub-frame stream. The NSUB limit is never exceeded in the absence of attenuation.

If the master assigns exactly the required number of stream channels (the assignment is spread in the reference burst), the system distributes the assigned channels among the applications requesting stream channels according to their Cr requirements. The control station ignores both the Cm value and the COS value declared by the requesting stations.

When attenuation occurs, each FODA / IBEA station tries to keep stream transmissions already active. in progress, with the specified BER. To compensate for the attenuation of the signal, stream transmissions require more? space in the frame to send the same amount? of information. Consequently, the station in attenuation sends to the master station an update of the original stream request in order to obtain a greater number of stream channels. This request ? marked with a flag as EXTRA to distinguish it from a normal update (not in FADE condition). The control station tries to fulfill the new request within the ESUB limit. Increasing the upper limit of the sub-frame stream? granted only in case of mitigation, and not to satisfy requests from new applications. When the sub-fraine stream crosses the NSUB boundary, stream requests from new applications are rejected.

If the increased size of the sub-frame stream does not? sufficient to satisfy the requests for increase, the system notifies (using the GAFO protocol) to the stream applications the need? to reduce the required band up to the declared Cm value. There? ? possible only if Cm? less than the required value Cr (compressible applications). After receiving compression approval from the indicated applications, another attempt is made by the attenuating station, sending a new request to the master. The new request? the sum of all possible Cm values and the incompressible Cr values. If even the new request cannot? be fully satisfied, all (or part) of the stream applications will transmit under degraded BER conditions. The decision to continue or not the session with degraded performance, or to give up, concerns the application itself, freeing the stream channels occupied by other users.

Figure 3 illustrates the stream assignment mechanism adopted by the master station (see Figure 5).

4.3.2 Datagram

Does the station make a request (r) for a datagram broadcast window at any opportunity? transmission, usually at every frame. This request ? expressed in multiples of 16 32-bit words. The request takes into account both the instant traffic entering the station (s) from various applications and the quantity? of data (q) lying in the station and waiting to be sent to the satellite (backlog). We have:

where H? a time constant of proportionality. Simulation results, obtained by loading the channel with datagram traffic generators according to Poisson for 10 stations, led us to choose a value of 0.4 s for the parameter H. For a station with no or small backlog, a request? therefore equal to the traffic entering the station in H seconds. For H equal to 0.4 s, the request? then equal to the traffic entering the station in 20 frames, if the backlog? negligible.

The datagram requests are forwarded as often as possible to give the master station the best possible situation. updated as possible of the traffic entering the station. The requests are organized by the master in a circular queue, which is scanned to calculate the individual time assignments. New datagram requests are analyzed first. There? it allows to reduce the delay time between the first request and the assignment, after a period of absence of transmissions. Any further datagram requests received by the same station (apart from the first one)? considered an update and replaces the residual of previous requests.

The length of the assigned transmission window (a)? proportional to the request in a range of values between a minimum and a maximum threshold

We have: where f? the coefficient of proportionality? in the assignment. In the current realization f? been chosen equal to the number of active stations N divided by 100, with 5% as a minimum and 50% as a maximum. ? was introduced for efficiency purposes. Avoid too small allocations when transmission overload - due to preambles and headers -? too large in comparison to the data to be transmitted.

avoids that a station too loaded takes away too much capacity? from other stations.

After each assignment, the datagram request is decremented from the assignment and the next request is parsed, if? space still available in the phratry. The first assignment that fits entirely in the current frame is analyzed again as the first assignment in the next frame where the rest of the quantity? calculated will come? finally assigned. All the space up to the end of the frame (if insufficient for a minimum allocation) is given as an over allocation in the last treated station.

Self ? space still available in the frame after an entire assignment cycle (ie the time interval between two consecutive assignments to the same station), the available space? shared among all active stations, including even those stations that have no assignment in that frame. This feature provides unused time sharing between stations. The unused space can not always? be divided between the active stations, since? the quotient may be less than the minimum allocation. If all active stations have datagram assignment, the allocation algorithm first tries to divide the remaining unused space between the stations that made a non-null request and were not fully satisfied. If there are no such stations, an attempt is made to divide the remaining unused space among all stations. If these stations are too numerous and the quotient? below the minimum threshold, an attempt is made to divide this space between the stations that have not received datagram allocation {since? these did not ask for any allocation). If the quotient? still too small, the unused space is divided between the stations that already? they have some allocation. This last division? always successful, since? there? no minimum threshold given that those stations we already have? assigned their overhead.

Generally speaking, the system has a behavior of the kind of a Fixed-TDM (F-TDMA), - called "preassigned mode" - all the more? accentuated at least the system? loaded. The system migrates to the pure FODA / IBEA assignment scheme when the channel load increases beyond a certain limit (pre-assignment limit). In summary, when the system? lightly loaded and the assignment cycle falls within the frame, each station has a capacity? pre-assigned and within the conditions of linearity? each loaded station generally has an over-assignment. There? allows the station to absorb a certain quantity? of sudden change in traffic.

The datagram request does not include the upstream transmission overhead due to preambles, headers, and guard times. The master station estimates the overhead required for each slave and adds it to the datagram assignment.

For datagram transmissions, the attenuation conditions cause an increase in actual backlog and instant traffic. There? automatically increases the station request in an attempt to obtain more bandwidth. Due to the increase in the request itself and the possible compression of the sub-frame datagram (caused by an expansion of the sub-frame stream), the capacity? overall of the datagram can? be significantly reduced, particularly under heavy load conditions. Some efficient action of the channel saturation control system is required to avoid congestion.

A relatively simple method of avoiding saturation? block the backlog growth for a certain time, exerting a "counter pressure" on remote users, when a dangerous situation for congestion is revealed. Since datagram data is collected from high-speed networks, the only effect of this procedure? to slow down datagram traffic for a convenient amount of time. To detect saturation, the user must be able to calculate his queuing time in transmission at frequent time intervals. The mechanism to block the backlog is activated as soon as the estimated delay exceeds a certain upper threshold and is deactivated when it falls below a lower threshold.

The following procedure is used by each station to make a rough estimate of the queuing delay.

The capacity channel reserved for the datagram (Cd) is calculated by decreasing the capacity? overall of the channel of the capacity? total assigned for stream broadcasts:

Let Dc be the part of the Cd assigned to a certain user U. It? the relationship between the quantity? of time of its own allotment and the quantity? total time allotted to all stations to send datagram traffic (assignment cycle). The delay D? roughly estimated as:

where all the variables are considered to be averaged over an appropriate time interval.

This mechanism currently does not? implemented. The current implementation blocks datagram traffic and re-enables it by employing two markups on the length of each of the station's internal input datagram queues.

4.4 Setting the transmission windows Each station finds the allocation of its own transmission window in the reference burst. Generally, the allocation assigned? equal to the request stream pi? a percentage of the datagram request. Is the speed specified in both stream and datagram requests? transmission of the preamble of the data type relevant to the request. Since a station can? send data to different destinations, this calculates the speed? of the preamble in order to adapt to the needs? of the destination station having the worst attenuation conditions. The reference burst is broadcast, so that each station is informed about the time assignments of all stations that send data. This data is received after an RTT (round trip time from satellite) (13 or 14 frames). The transmission time plan is used to preset the acquisition of the modem at the speed? specified of each expected data burst and to set the relative acquisition windows of the UW.

Generally, the trend? to set only one transmission window to send both stream data and datagram data. There? ? possible only if the speed? transmission of the preamble of the stream assignment? lower or equal to the speed? transmission of the preamble of the datagram assignment. In this case, in effect, the stream data must be transmitted with worse transmission characteristics than the datagram data and the worst speed? transmission of the preamble (that of the data stream) pu? be used for both types of data without removing anything? to stream assignment.

On the other hand, if the speed? transmission of the preamble specified in the stream assignment? greater than the speed? transmission of the preamble of the datagram assignment, then two different consecutive but separate windows must be set, one for the stream data and the other for the datagram data. There? happens since? the adoption of the worst speed? transmission of the preamble (the one specified in the datagram data would deprive the stream allocation of part of the allocation itself, due to an undue enlargement of the preamble.

4.5 The FODA / IBEA transmissions

AND? It is convenient to provide some definitions at this point. Data burst (DB)? the quantity? of information that can? be transmitted to the satellite from a certain station during its transmission window. The transmission window? the time allotted to a station by the master station to send stream and datagram data. A burst can? contain data addressed to different destination stations. A data burst? consisting of a control sub-burst (CSB) which contains all the control information with respect to the subsequent n data sub-bursts (DSB) each with individual destination stations and with individual transmission characteristics (coding and transmission speed ). Another zone, called the channel control area (CCA), is included, when needed, in the control subburst to send assignment requests and other control information.

A sub-burst stream can? be made up of pi? of a packet or packet fragment. A sub-burst datagram? consisting of only a data packet or packet fragment.

The reference burst (RB)? a particular type of burst, broadcast at the beginning of each by the control station. This is sent for synchronization purposes, to provide the schedule of transmission times to the stations at which? allowed to transmit in the frame following the one in which? received the reference burst. The reference burst does not contain any sub-burst. (Figure 6)

The "preamble" sequence? consisting of an initial segment of an unmodulated carrier (carrier) followed by inversions, that is? an alternating sequence of 0 and 1 in order to reconstruct the time synchronization to the bit. The carrier is sent as successive binary "00" and is used in the receiver modem to retrieve the carrier. The inversions are sent as successive "1100" and are used by the receiver to recover the symbol clock. There? corresponds to alternating "1" and "O" for each of the I and Q channels of the QPSC system.

The configuration of the preamble? followed by the single word (UW).

The UW? a sequence of 1 and 0 (48 bits long) both on phase I and on phase Q of the carrier, chosen to have good properties? of correlation. This is to mark the beginning of the data transmitted on the channel. The instant in which the correlation peak of the UW occurs marks the reference for decoding the information in the data part of a burst. The subsequent recognition of the UW? necessary to mark the arrival time of the reference burst and to establish the reference time to decode the received bursts. A sequence of 2 UWs is used to create the single word of the master (MUW), used for the reference burst, while a single sequence of UW and its complement are used to create the single word of traffic (TUW), at the start of each CSB.

Both the preamble and the (UW) are always sent unencrypted.

The control sub-burst (CSB) contains information about each of the successive sub-bursts of data in the burst. At choice,? also contained a number of extra words that can? be used to contain status information about the transmission conditions of the burst (the CCA). For each sub-burst of data in a burst, the control sub-burst specifies the number of words, encoding, and rate. used. Since an incorrect decoding of the control sub-burst causes the rejection of all the data sub-bursts that refer to it, the CSB must always be transmitted with the highest degree of security. Satellite headers are included in the control sub-burst.

4.6 Quality estimate of connection and dissemination.

Each transmitting station must know the signal / noise ratio available in the station where its data is addressed, in order to choose the transmission parameters (encoding and transmission speed) that allow the data to be received with the required BER. The parameter chosen to express the conditions of the link? the ratio Indicating with r the speed? transmission expressed in bit / s, the ratio

(bit energy versus power density of

noise) ? given by

The estimate of the ratio is carried out by each station, at each arrival of sub-burst using the measure of quality? provided by the hardware. The "soft" quantizer that provides the input to the Viterbi decoder also feeds a finite state machine that calculates the quality? of the signal, that is, a weighted sum of the number of bits in a burst that falls into each of the quantities obtained by the "soft" quantizer. The quality signal? divided by the number of bits in the sub-burst and is then used to enter a lookup table that provides the for that sub-burst.

System parameters

4.8 Structure of the data sub-burst.

The sub-burst structure of data? shown in figure 7:

The DSB Preamble (SBPRM) (Sub Burst PReaMble) and the DSB UW (SBUW) are required for all DSBs. The capacity of clock tracking and of the modem carrier requires different lengths of the SBPRM, depending on the speed? transmission of the previous and current DSB. The DSBs within the same DB must be organized in increasing order of speed? transmission.

The SBUW (Sub f? Urst f? Nique Word)? formed by combinations of UW using the 24-bit binary sequence of UW = OxSACCFO on each of the P and Q channels.

The queue overhead (queue zero CRC) is passed as a single null word to the modem interface card. The modem places the sub-burst CRC in the first 8 bits, and uses the remaining bits to download the encoder.

5. OVERALL SYNCHRONIZATION OF THE SYSTEM

The synchronism to the frame? given by the master station. The frame counter of the master (FC) goes from 0 to the end of frame (FE) value. The RB is sent at the beginning of each frame, at a fixed time. The master UW acknowledgment event (MUW) clears the FCs of all Rx receiving controllers (including that of the master). The master does not require any synchronization, since? it itself provides the reference time. The offset between the reception and transmission times is set automatically in the case of the master. Each slave station must continuously measure its RTT time and consequently adjust the position of its transmission frame time axis relative to the receive time axis, continuously adjusting the position of the transmission frame pulse (TFP) in memory. of Rx event. The master does not measure its RTT since? does not set the TFP.

A "default" value of the RTT, supplied by an operator, is used to initialize the system. Once a slave has synchronized its reception frame with the MUW, the initial TFP is placed, using the default RTT, whose error must not be greater than 150 microseconds (600 symbols at maximum speed). The first burst sent by the new station entering the network? centered in the first access slot (FAS). The FAS, sized to double the normal control slot (CS)? 150 microseconds, has a fixed position at the end of each group of 32 frames, and is accessed in contention with other possible stations that arise.

In the case of a collision (missed reception of its own burst) the station tries again after an n-frame block time, where n has a constant distribution of the interval (1; S), where S? the maximum number of stations allowed in the system. Following a correct reception, the station compares the actual reception time with the expected one (time from recognizing the UW of the RB), and uses this information to correct the RTT.

After entering the system, each station must continuously adjust its RTT, to compensate for satellite shifts and clock drift. The same regulation mechanism used during the subscription is not? easily made for subsequent bursts, since? their position is not? fixed within frames, so each station should keep a history of the last bursts transmitted along with their transmission times. these are compared with the reception times. Since the errors predicted in the RTT are of the order of 1-2 ticks for each frame,? was used a method pi? simple.

Each burst contains a 7-bit field containing the least significant bits of the transmission time. The difference between the expected and actual reception time is then calculated using this 7-bit field, assuming that the error will not be. never greater than? 63 frame ticks. This difference is used to adjust the RTT.

The regulation algorithm? constructed in such a way as not to give erroneous results even in the presence of measurement errors. The stability of the feedback mechanism? ensured taking into account some conditions largely satisfied by geostationary satellites. The feedback mechanism? been verified using the unit? Marconi delay that allows manual variations of the RTT, and yes? proved stable by entity? of displacement of the satellite far beyond that of real satellites.

6. ADJUSTMENT OF THE POWER OF THE TRANSMISSION SIGNAL

If the output power control is disabled, the system transmits at maximum power (OdBW at the IF output). The transmission power can? be varied by using the "+" and keys on the controller terminal in downward direction.

Self ? enabled the output power control, the system tries to adjust the transmission power by itself in order to maintain the same power at the input of the satellite. cos? like the other stations. The master and slave stations have different behaviors.

The attenuation levels on the up and down legs are 0 under clear skies. The transmission reference level (at the modem output) is the power level that allows an Eb / N0 of 12 dB at the modem input in clear sky conditions. The transmission reference level depends on the antenna and the transmit amplifier.

It is assumed that a variation of 1 dB of attenuation in the uplink leads to a variation of 2 dB (in the same sign) of attenuation on the downlink on a satellite channel at 30-20 GHz. This assumption? a simplified and extrapolated consequence of some considerations.

All ? pi? simple for the slave stations, which follow the reference of the master. That is, each slave station employs a feedback loop to keep its own received power level equal to that of the master. In this way, if the station has sufficient power, the power levels on the satellite are the same as the power levels on all the stations' modem inputs. After a slave station? entering the system, pressing the "p" key freezes the reference level, so this? available for the station even if it becomes a master.

Claims (11)

CLAIMS 1. Data transmission system on microwave carriers, in particular in the Ka band (20/30 GHz) between a multiplicity? of receiving-transmitting ground stations, connected to each other via geostationary satellite in which each ground station? equipped with the means that guarantee both the data transmission / reception system on satellite (as regards the earth / satellite / earth connection), and the processing means for the management of the incoming data from the terrestrial network and for the acquisition of the band transmissiva on satellite, said data transmission system being characterized by the fact that one of the N stations belonging to the satellite transmission network? enabled to operate as manager of the common resource (ie the band on satellite); said station being operating as a "master"; the other stations operating as "slave" stations forwarding requests to the master station to obtain band allocations in order to transmit data, and in which, as regards the transmission of its data, the master station behaves like any slave station. 2. Data transmission system according to claim 1, characterized in that when the current master station suddenly stops operating, the role of master is automatically filled by the first station in the list of stations that have declared, upon their entry into the network via satellite, your availability? to operate as a master. 3. Data transmission system according to claims 1 and / or 2, characterized in that the data transmitted simultaneously on the satellite are both of the asynchronous type (such as those generated by applications such as file transfer, remote access, etc.) and synchronous type (such as those generated by applications such as voice transmission and moving images), particularly suitable for video conferences. 4. Data transmission system according to claim 3, characterized in that the signal transmissions operate according to a time interval, called "time frame", in turn divided as follows: a sub-frame for the reference and synchronism signal, in which the master station transmits a reference burst, - as many transmission windows as there are stations that have access to the channel in that given frame, in each transmission window, except the FAS, both synchronous and asynchronous data are transmitted from the applications pertaining to the ground station via a LAN or a WAN (Wide Area Network). - a transmission window, every 32 frames, dedicated to the entry into the network of a new station, if the maximum number of operating stations has not yet been reached (First Access Slot); this window being accessed in contention between the stations that in that certain instant decide to become active. 5. Data transmission system according to claims 1,2,3 and 4, characterized in that data transmissions take place with a technique capable of counteracting the attenuation of the transmission signal due to bad atmospheric conditions. 6. Data transmission system according to claim 5, characterized in that the countermeasure technique of the attenuation of the transmission signal is based on three points: a) control of the output signal power (uplink power control); b) data redundancy obtained by increasing the transmission encoding, being the possible encodings: 1/2, 2/3, 4/5 and not encoded; c) data redundancy obtained by decreasing the speed? of transmission, being the speeds? possible: 1, 2, 4 and 8 Mbit / s; points b) and c) being made via software, and both the coding of the data and their speed? being dynamically varied according to the extent of the attenuation of the signal of the individual ground stations; within the same data burst, the single sub-bursts being transmitted by choosing the speed? transmission and data encoding based on the fade conditions of both the transmitting and receiving stations, or the worst of the receiving stations, in the case of broadcast transmission (to all active stations on the network) or multicast (to a subset of active stations). 7. Data transmission system according to the preceding claims, characterized in that in clear sky conditions the overall space in the frame reserved for allocations for synchronous traffic does not exceed a certain number of bytes, or, in other words, a certain limit indicated as NSUB (Normal Stream Upper Boundary), while the rest of the frame? dedicated to allocations for asynchronous traffic; in this situation, new allocation requests for synchronous traffic being accepted only if they can be satisfied within the NSUB; in fade conditions, the transmission of both synchronous and asynchronous data being redundant according to tables that associate the fade level found with the appropriate coding and speed? transmission in order to keep the BER of the data unaltered, according to the specifications of the generating application, since? the maintenance of synchronous data transmission sessions is privileged, the NSUB limit being extended to the ESUB value (Extended Stream Upper Boundary), which can? also coincide with the end of the frame, with consequent cancellation of asynchronous transmissions for the entire duration of the attenuation; in this situation, requests for allocations for synchronous traffic by new applications are not accepted, and when is the situation of fade? exceeded, the limit? reported to the NSUB value. 8. Data transmission system according to the preceding claims, characterized in that the allocations for synchronous traffic, the requests of which are accepted according to the modalities? claimed in claim 7, are the same as the request made; requests for asynchronous traffic following the rule: request = stock H (incoming traffic), where H? a constant of proportionality? equal to 0.4 seconds; their allocations following the rule: being f the proportionality coefficient? of the request (r) used in the assignment, chosen equal to the number of stations N divided by 100, with 5% as minimum and 50% as maximum, min being the minimum value that an allocation can? have and max being the maximum value ,? after each assignment, the relative request being decreased by the assigned value and the subsequent request being analyzed, if space? still available in the frame for asynchronous allocations; the space possibly still available after an allocation cycle (time between two successive allocations to the same station) being divided among all active stations (pre-assignment mode), including those stations that have not made any request, in order to ensure to all stations a lung of advantage in the event of a transient traffic. 9. Data transmission system according to the preceding claims, characterized in that within a transmission window, the multiplexing between synchronous and asynchronous data? fully charged to the transmitting station, which can? also use any spaces not used by your synchronous transmissions (see moments of silence in voice transmissions) to transmit your asynchronous data. 10. Data transmission system according to the preceding claims, characterized in that each data burst? preceded by a control sub-burst (CSB) containing, among other things, the length of each sub-burst, its encoding and relative speed? transmission. 11. Data transmission system in the Ka band (20/30 GHz) between a multiplicity? of ground transceiver stations, connected to each other through a geostationary satellite, as described above and illustrated with reference to the attached drawings,