SE460750B - TELECOMMUNICATION SYSTEM IN ANY NUMBER AND DATA INFORMATION IN TIME DIVIDED FORM TRANSFER OVER BUSES IN A MATRIX FORMAT - Google Patents

Description

460 750, 2 10 15 25 however data transfer very poorly. If each node is allocated a static bandwidth, the bus's capacity would be utilized sparingly, and the bus' ability to fit different types of systems would be severely limited.

The device according to the invention, which solves said problem, is characterized by the claims and comprises a bus system comprising very fast optical buses for transmitting voice and data information between subscribers without the use of group selectors. A bus of the type specified above can be used to connect a number of distributed nodes connected to the bus in, for example, a distributed telephone system. Several buses can further be connected to each other via so-called BRlDGEmr (bridges) and thereby a whole exchange is formed. The bus replaces group selectors and also RP buses (Regional Processor buses) in known switches and mediates voice and data communication as well as interprocessor communication between the nodes. Control of the information exchange between the nodes takes place with the help of a special communication protocol (DUPER). The protocol provides the ability to exchange speech and data over a high-speed bus, for example 2.4 GHz, with a dynamically variable bandwidth where the single node only needs to handle its own part of the bandwidth.

DESCRIPTION OF THE DRAWINGS The communication system according to the invention is described in more detail below with the aid of an exemplary embodiment with reference to the accompanying drawing in which Fig. 1 shows a bioecdveana of a system according to the invention, Figure 2 shows the division of a bus into static and dynamic time slots. according to the protocol according to the invention, Figure 4 shows in matrix form a network connected in accordance with the invention.

Figure 5 is a block diagram form of an example of the structure of a node in the network.

PREFERRED EMBODIMENT Figure 1 shows a communication system in which a number of communication nodes N1-Nli are connected to a common high-speed bus H81. The bus is in turn connected via a bridge unit B to an additional high-speed bus HB2 which in turn is connected to a number of communication nodes BNI-BNI To 10 15 20 25 30 3 460 750 the nodes Nl-Nli (also applies to the nodes BN1-BN3) are connected subscribers with extremely varying needs for bandwidth, from ordinary telephony to the transmission of TV images.

Through the interconnection of buses via bridges, geographically very separated subscribers can be connected in large networks. The switches of the future will still for the most part cover ordinary telephone traffic, even if the largest communication bandwidth will be occupied by data communication. The bus must therefore be able to handle both static calls with small bandwidth and data that scurfs require greater bandwidth in a simple and cheap way. Speech transmission has long taken place on time-divided buses, where samples from several calls are transmitted under one frame, the next sample under the next frame, and so on. Time-divided buses have also been used for time switches, where several units on the same bus communicate with each other through reading and writing in special time slots.

The fast fiber buses at 2.4 Gbit / s, open up completely new views here. However, the high bit rate would make the hardware unreasonably expensive if traditional protocols were used to control the flow of information on the bus, therefore the protocol according to the invention is adapted for high bit rates and can also be used to connect larger bandwidths than speech in the same way as a voice coupler. .

Traditionally, all devices are allocated the same bandwidth, whether needed or not. According to the invention, the different units on the bus are allocated different bandwidths. The assignment can also be changed during operation with a very large dynamic. The bus's dynamics therefore make it suitable for transmitting data and interprocessor communication as well as speech.

The protocol according to the invention is based on the bus being divided into a number of time slots, each containing a command word and a number of data words, where each data word can correspond to a call. It is these time slots that are to be time-switched on the fast bus, not every call separately. The command word indicates changes (eg up and down) in a leash, which may consist of one or more of the data words. 460 750 4/10 15 20 25 30 'I According to the protocol, each 125 "communication frame is divided into a number of static and a number of dynamic time slots, see Figure 2.

Each unit will have a fixed time slot connected to each other unit on the bus, which is the static part of the frame. The remaining _ part of the frame is the dynamic part. These dynamic time slots are shared between the units at initialization, so that different units can be assigned different band transmitter allocated bandwidth.

The transmitting unit can, with a command word (eg in the static part), give the receiving unit an order to start listening to a new dynamic time slot.

The protocol also includes a distributed resource allocation whereby one device can ask the other devices for more bandwidth when it has used all of its own dynamic time slots.

The application does not explain the principles of 'time multiplexing and PCM' as these principles are known.

The invention relates to a bus protocol which can be characterized as a dynamic time slot protocol. Below, the mechanism itself will be presented.

According to the example, the number of nodes is limited to 24. A change in the number of nodes only changes the relationship between the static and dynamic part of the bus.

Both a static and a dynamic bus have advantages in certain situations. The device according to the invention is a combination that gathers the good properties of the buses.

The bandwidth is divided into about 1/4 static and 3/4 dynamic proportions.

In order not to make dynamic allocation of bandwidth inefficient, the bus is divided into larger pieces, so-called "time slot assigned packets". As shown in figure 3, each packet consists of 12 8-bit words (two command words, eight data words and two error-correcting words). When the system is initialized, all nodes have a static time slot connected to each other node. This is the so-called static part of the bus that will be stuck and occupy about 1/4 of the bandwidth. If the number of nodes is 24, the number of static gaps will be 23 x 24 = S52. The dynamic part would then be about 1500 d vs 2k gaps per frame. With the current bandwidth, the frame length would be l25, us. See Figure 2.

The static part of the node of a node will be sufficient to transmit 8 calls to S 10 15 _20 25 30 460 750 'every other node, or 0.5 Mbit / s. Connection and disconnection is controlled by the command word of the static door. If more than 8 calls, or if a wider packet-type data communication is needed, the node may ask the receiving node to start juxtaposing a specific dynamic packet, and use this for the transmission.

The connection of this packet is made by sending the command word "Start listening to packets" on the static packet that goes to the current node. In this way, the bandwidth towards other nodes can be increased if necessary with an extra dynamic packet per frame (change the bandwidth by 4 Mbit / s per ms). Increasing the bandwidth by 10 Mbit / s towards another node would take 2.5 ms.

At the initialization of the system, the dynamic packets are divided evenly between the nodes, so in this case each node has about 64 dynamic packets. If a node's dynamic packet starts to run out, a request is sent to the node with the nearest higher number to send an unused slot. The answer with the number on the current slot comes after a couple of frames. In this way, there will be a flow of free gaps that circulate when needed.

It should be pointed out that the distribution of resources only takes place when a shortage of gaps begins to arise in some node. in the bus, therefore, it adjusts itself to a suitable bandwidth distribution between the nodes.

The dynamic time slots have the same appearance as the static ones with two command words, eight data words and two error words. The dynamic command word can make one connection per frame within its own package. A connection can apply to one or up to all eight words in a packet, ie 64 kbit / s to üåMbit / s.

The dynamic 'hatch' will thus control itself. This means that a node's ability to control communication increases at the same rate as its bandwidth.

This also means that the control will take place on two levels - one that adjusts the bandwidth and one that connects the data bits.

The commands that control connection and disconnection in the static and dynamic packages have the following formats: <"l"> <START (3)> <WlDTH (2)> <DEST (l0)> .rg 460 750 s 10 15 20 25 There START indicates which of the eight slots will be the channel's first slot.

WIDTH indicates whether the channel should include 0, 1, 4 or 8 consecutive words and DEST indicates which individual on the node should be the recipient of the packet. DEST is used as the global find address.

The commands in a static packet that change the bandwidth between different nodes can have the following formats: <"0"> <ORDERO)> <SAVE (l)> <NUMBERUl)> r 'Where ORDER encodes some of the commands: "Start listening to package "" Stop listening to packages ". g f "I want a free package" "Here farqdu free package" "I am healthy" \ "Error code" - SAVE can be used to increase some of the fields if necessary. The NUMBER fields are used to enter a number 0-2k. STATUS CODE and ERROR CODE inform other processors of how heavily loaded a node is and its sub-levels, as well as the type of error that the node has suffered. This means that the other nodes can be quickly updated if something catastrophic is about to happen. advantages of the proposed solution include _STATIC STRUCTURE The bus has a static structure, where communication between specific nodes takes place on specific time slots. However, a certain part (dynamic) of these gaps may be allowed to change slowly. This "dynamically static" structure keeps the requirements for the fast part of the hardware (GaAs) reasonable.

CONSISTENT FORMAT Static and dynamic slots have the same format DYNAMIC BANDWIDTH The bandwidth can be changed by increasing or decreasing the number of dynamic time slots between two specific nodes. A connected, _ L '.- ïßfš fi v-faí: 10 15 20 25 Y. 460 750 bandwidth is not disconnected, however, until another node needs_ the resource.

This is an important feature of data communication where data is transmitted in bursts.

MODULAR CONTROL AND DATA RELATIONSHIP The control bandwidth grows at the same rate as the data bandwidth.

HIERARCHIC CONTROL Two levels of control have been introduced. A top level that handles basic communication and adjustment of bandwidth. A local level that takes care of ups and downs within a package.

Q ERROR DETECTION All nodes send at least one packet to each other node under one frame. If there is nothing to 'send, the code "I AM FRESH" is sent, which means that an incorrect node will be detected by all other nodes within a frame.

ERROR CORRECTION Error correction has been introduced, which means that almost all communication errors can be detected and also corrected. oisïRiauz-: ROW REsouRcE MANAGER A simple: kancep: allows to dim the control of free gaps.

SELF-ADJUSTMENT OF BANDWIDTH A bandwidth set by the RESOURCE MANAGER will remain until the needs have changed.

FREE NODE NUMBER A change in the number of nodes means that the proportion of static-dynamic gaps changes. If blocking freedom is desired, only a small number of nodes are connected, but if cost efficiency is important, a larger number of nodes can be connected. the information over the bus consists of a serial bit-phase coded bitstream that contains clock information and data as well as sync pulses. Sync pulses set the bus times so that a time-divided bus is obtained. The T-part or frame is divided into static and dynamic time slots. According to the example, a frame consists of 16 bit control words divided into two 8 bit words then a selectable number of 8 bit PCM samples. The protocol format according to the example includes 8 PCM samples per Pam. 1 'i l i 10 15 _ * Ven:. _ 460 750 - Example of frame design: N 3 2 l 2 l The length of the frame (piece of contents) is dimensioned by node handling 1§ design.

The design of the control word (COM) is conditioned by the 'characteristics of the intended protocol. They all look identical in size. The time slots on the bus are divided, as mentioned, into two groups static and dynamic.

The nodes are now assigned a set of static slots that have a destination defined in the requirement to reach the other nodes' * An example with three nodes gives \ The first NODE a time slot to the second node and one to the third node, the second NODE a time slot to the first one to the third. . third NODE a time slot to first one to second.

Nodi Nodz Noa: Dynamic time slots are assigned to the nodes but have no destination specified.

Through the protocol you can then destination these free slots to nodes according to '°° “°“ ° N0d1 Ex' Nodz 'Nada Q - Dynamic time slots b = Destined static time slots. 10 15 20 9 - 460 750 + A status word, "I am healthy" is continuously transmitted on the bus, which indicates that the sending node is being monitored. If no order information is carried, the communication remains in the previous position.

By using several nodes to connect several buses in a grid pattern, a network with a very large capacity is created. f »BUSES Each node can now, via the static time slot, select which node it wants to communicate with on its two buses. This is internal information that never has to leave the node as it via the static time slot reaches all nodes after both the X, Y, axes. Receiving NOD knows who is transmitting, thanks to the fact that the fixed part of the frame is defined by its time position.

If the connection of a node to a bus applies to 1 PCM channel or the corresponding bandwidth within the frame, the message can be received and the subscriber in the receiving node is designated with 10 bits in one step along his own buses. If it does not apply to the nodes of the own buses, allocation and bridging to the next bus can be arranged through ORDER.

Since not all bits are used in the designation of the node (only 10), a number of the additional 6 bits can be used for the designation of subscribers, which according to the example increases the designation possibilities from 1000 subscribers per node to 32,000 subscribers per node.

When connecting a network, with an ORDER word and a STATUS word, 10 15 20 25 460 750 w each of 16 bits, of which 5 + 10 bits were used for pointing, it is theoretically possible to choose, in a maximum configuration. 32 * 32 nodes * 32000 pcs. 32 million subscribers. 24 * 24 nodes * t 32000 st. 18 million subscribers.

This requires two 125 microsecond frames. If you want to complete the set on a frame, you can choose from 1000 designations per node.

As previously mentioned, the protocol must provide a node whose computing power does not have to correspond to the speed of the fiber buses, since only a small part of the transmitted data is received in the node. The unallocable bandwidth allows for flexible utilization of the bus' transmission capacity. The protocol allows distribution in a bandwidth-efficient way as you can get all nodes to listen on the same time slot or selectable number. Another advantage is the ability to reach any other node in the network via only two buses and to always have two routes there.

The network can be seen as a distributed switch or an intelligent transmission network.

Advantages such as redundancy and modularity are also achieved by the invention.

The system according to the invention is designed as a communication mechanism in a distributed telephone system, but has also proved useful in multi-processing.

The features that have been sought are: A network with a simple addressing algorithm, ie routing table should not be needed.

Short distance between all nodes, ie a minimum of bridges (BRIDGEs) must be passed as a channel is connected between two arbitrary nodes.

There must be several alternative routes for safety reasons.

Quick and easy way to connect a channel between two nodes. This is to also handle interprocessor communication over the network.

(IC '10 15 20 25 30 m f 46Û 7' a 'Large dynamics in bandwidth between nodes, without using a central control.

Each node is equipped with at least two gates to different buses, so that a single bus fault cannot knock out several nodes. Very little hardware is required to allow these buses to communicate with each other via a direct channel.

To enable this, a so-called node handler (NODE HANDLER) is required, the function of which is described in more detail in connection with Figure 5.

There is only one type of node handler in the system. The node handler has two ports for the said DUPER buses, as well as one port for a node processor. This means that the node manager can serve as a bridge (BRIDGE) between two buses, at the same time as there is a free port for a node processor. As can be seen from Figure 4, the network is built up as a matrix network M max 24x 24 with a node handler at each intersection point. This provides a system of a maximum of 576 nodes, where the connection between two arbitrary nodes can be connected along two different roads that only pass one bridge on the road. The two paths are obtained by first going to the correct X-coordinate and then to the correct Y-coordinate, or vice versa. Of course, a Sadant network is not dependent on routing tables or the like.

The system is built around 24 horizontal Y-buses and 24 vertical X-buses of the DUPER type, designated DB. These are numbered zero from the left and from below and are denoted, for example, "X-bus 1" for the other horizontal bus from below.

The names of the nodes are formed by their X and Y bus numbers (X and Y coordinates) and are denoted (1, 14) for node 14 on X-bus 1 and also node l on Y-bus 14. In the network the node name is given in 10-bit binary form, where the first 5 bits indicate X and the last 5 bits indicate Y coordinate. Node (1, 14) is indicated in binary form as 0000101110.

Each intersection between the buses is controlled by a node handler, designated NH. NH transmits data words from one bus DB to the other bus DB, or to the node's local node processor, designated NP.

The communication on DB takes place in packets, in 12 words. Two words are Command (COM), 8 Data words (DATA) and two are CRC check (CRC). An entire PACKAGE is addressed to the same node handler NH. However, the different data words may be addressed to different destinations. Approximately 2000 PACKAGES are sent within a RAM of 125 us on each * “= '» fï' ^ * f '* * mms 460 750 12 10 15 2D 30 bus, i.e. corresponding to 16,000 calls.

Node processor NP leads and distributes the work at the nearest lower level, which is assumed to be a DB-type bus, but slower. Each NP serves about 500 subscribers.

Messages that are not addressed to node processor NP are forwarded by the node handler NH on the other bus DB. This forwarding is called bridge function. It can be noted that a bridge coupling always means a perpendicular course change.

Figure 5 shows the structure of a node handler NH. As can be seen from the figure, it is possible to, in order to increase the degree of freedom, connect more than one node handler.

The example shows, with reference to the matrix network, two horizontal node handlers and two vertical node handlers. The input logic IL from the fiber bus DB to the node handler NH is built in Gallium Arsenide (Ga As) and contains, among other things, an opto / electrical converter for converting the fiber bus' optical signals to electrical signals and a series / parallel converter for converting the bus's serial information flow to information in parallel. form. A clock sensor CL provides the system with the necessary clock signals. The receiver R receives incoming signals and further forwards them to a demultiplexer D which in turn outputs the information to a first buffer memory BM1 which serves as a buffer between the incoming fast information flow and a slower internal bus IB. A second buffer memory BP adjusts the information flow between the internal bus IB and the node processor NP. A third buffer memory BL adapts the information flow between the internal bus IB and a local bus LB to which the subscribers belonging to the node are connected.

In the other direction, when the information goes from the subscriber through the node N1 to the bus DB for connection to other nodes on the common bus or for brewing over to another bus, the information passes the said 'third buffer memory BL and according to the example a further buffer memory BM2. Via a multiplexer M, a transmitter T, parallel / serial conversion and “in electricity / opto conversion in the unit IL which thus also contains output logic, the information is supplied to the bus DB. A by-pass unit FK allows bypassing of the node manager's transmitter logic as the node itself has nothing to transmit. The node processor NP, which may 10 15 20 25 ß j - 460 750 be a microprocessor made by MOTOROLA type M 68000, controls the information exchange through the node in a known manner. All current data about the information flow is stored and read out via control memories CM. A time slot counter C serves as a pointer for writing and reading in the memories. As can be seen from Figure 5, it is possible to connect more node handlers to the node where, for example, node handlers 1 and 2 refer to horizontals in the previously mentioned matrix network M and node handlers 3 and 4 refer to verticals.

Below are some methods for route selection, so-called routing through the system.

Thanks to the network's simple structure and numbering, routing tables will not be needed. In the first place, the two-stage X-Y routing paths will be tried. Should these give errors or barriers, a three-step routing can also be used, which gives an additional 64 possible paths.

CONNECTION OF A TWO-STEP ROUTING.

When a node processor NP wants to set up a connection to a node, the receiver's logical name is translated into the physical X-Y name. NP sends the channel request containing the physical name, as well as the current bandwidth and an identifier unique to the node, to the node manager NH. After a certain time, NH will respond with the same identifier and a green light will start transmitting. The bandwidth of a channel can be 64 kbit / s - SOUkbit / s. If a larger bandwidth is desired, node processor NP must request multiple channels, which it administers itself.

The emergency manager NH examines whether it has sufficient bandwidth against the indicated bridge (BRIOGE). If this is not the case, it connects one of its free packages to the bridge under the next frame.

THREE-STEP ROUTING Three-step routing can take place via any of the other nodes located along the two DB buses to which the node manager NH is connected. The node manager can see from old status reports which node seems to fit best. The "geographical" location of the nodes will not matter. If necessary, a new slot is set up against the selected node, and a channel is established in the usual way by sending connection orders together with a 10-bit destination code. The receiving node will detect that the channel is not to be set up to its own node processor NP, and will connect the channel perpendicularly further.

The same destination code will be used, which means that a normal two-step routing is started here.

A network according to the invention has a very good safety against bus faults. Occasional errors will be corrected by the CRC word. If a bus completely falls out, it does not stop a single node, because all nodes on the bus due to missing "I am healthy" messages understand that the bus is broken, and use the alternative route. Should two X-buses fall out, this will not knock out any node either. If an X-bus and a Y-bus should fall out, only the node at the intersection point is knocked out.

As can be seen from the description, the telecommunication system according to the invention presents a unique solution to the problem of transmitting information of extremely varying bandwidth in addition without using group selectors as in today's telephone system. In addition, through appropriately selected parameters, hardware costs are kept at a very low level. f "N * .lqtßlvavf

Claims (4)

10 15 10 1s 460 750 PATENT CLAIMS Telecommunication system in which voice and data information in time-divided form is transmitted over buses to which a number of nodes are connected, each in turn connected to a number of subscribers with varying bandwidth requirements and to bridges. connects the buses to each other so that a matrix-shaped network is obtained in which all subscribers have the possibility of connecting to each other and in which each node has connection possibilities to at least two different buses in the network, characterized in that to enable the transmission of information with little bandwidth requirements as well as information with very large bandwidth requirements, the information transmission on the buses (DB) is controlled by a protocol with a format adapted to the transmission and that each node (N) is assigned time intervals comprising static and dynamic time slots which include at least one command word and a number of data words which data words can each correspond to a call or other information on a connection and that for each node at least one static time slot is designated so that each node has a fixed connection to each other node on the bus common to the nodes, and that said nodes are also assigned dynamic time slots in relation to their special requirements for bandwidth and that each node is allocated unused dynamic time slots from other nodes if necessary. Telecommunication system according to claim 1, in which voice and data information in time-divided form is transmitted over buses to which a number of nodes are connected, each in turn connected to a number of subscribers with varying bandwidth requirements and that bridges connect the buses to each other so that a matrix format network is obtained in which all subscribers have the possibility of connection with each other and in which each node has connection possibilities to at least two different buses in the network, characterized in that each node contains a node handler (NH) which, under the control of a node processor (NP), converts and exchanging information between subscribers-nodes-kbuses and bus-nodes-subscribers, respectively, and that a node handler (NH) is connected as a bridge in each of the crossing points in said matrix network (M) for transmitting information between buses. Telecommunication system according to claim 1, characterized in that said buses (DB) are high-speed buses designed as optical fiber ono 460 1750 16 buses. Telecommunication system according to claim 2, characterized in that forwarding of information via a bridge always means a perpendicular change of direction.