GB2168574A

GB2168574A - Transmission system

Info

Publication number: GB2168574A
Application number: GB08431709A
Authority: GB
Inventors: Keith Caves
Original assignee: STC PLC
Current assignee: STC PLC
Priority date: 1984-12-15
Filing date: 1984-12-15
Publication date: 1986-06-18
Also published as: GB8431709D0; GB2168574B

Abstract

A communication system of the closed-loop local area network type has a number of stations with each of which there is associated a by- pass switch. If a station or its adjacent links are faulty the bypass is operated to short it out. In the present system the bypass switches are all located at a central distribution frame with each station connected to its bypass switch by a two-wire connection. Thus the system is in effect a distributed star ring. To improve system security the stations use braided control whereby each is also connected via a bypass switch with the next but one station. Multi ring and herarchial networks are also described. Note that each station referred to above can serve a number of separate users. <IMAGE>

Description

SPECIFICATION Transmission system This invention relates to a communications system of the serial type.

In any communications system whose users are connected serially via a communications medium, a failure of a single series element can be catastrophic. An example of such serial connection occurs in ring-based communications systems, e.g. the Token ring and the Cambridge ring systems. Here communications between user stations (the point of attachment for a user's terminals) are actively repeated by all intermediate stations on the ring.

Hence the failure of a single user station, or of the transmission medium between stations, results in loss of communications for the whole system.

An object of the present invention is to prevent the failure of a single system element from catastrophically affecting the rest of the system.

According to the present invention, there is provided a communication system, in which a number of stations are interconnected in serial manner by a communication medium, in which each said station is provided with a bypass switch which is normally disabled, so that its station is connected in the medium, in which if a fault is detected in a said station or in the medium adjacent thereto that station's bypass switch is enabled so that the bypass switch switches the station out of the medium, in which all of the bypass switches are centrally located at a common distribution frame, and in which each said station is connected to its said by pass switch by physical transmission paths which extend between that station and its said switch.

Although the invention is described herein as applied to closed-loop systems, it will be appreciated that it is also applicable to linear systems, e.g. of the well-known Ethernet type.

Embodiments of the invention will now be described with reference to the accompanying drawings, in which Figure 1 is a simple representation of a bypass switch, Figure 2 is a closed-loop system embodying the invention, Figure 3 shows some detail of the bypass circuitry for a system such as that of Fig. 2, Figure 4 is a closed-loop system of the same general type as Fig. 2, but using socalled braided control, Figure 5 is shows some details of the bypass circuitry for a system such as that of Fig. 4, Figure 6 shows a bridge station for interconnecting live loops or linear systems, Figure 7 is a simplified diagram of a system embodying the invention with hierarchically arranged loops.

In many ring-based systems, reliability is enhanced by equipping each user station on the ring with a bypass switch. Fig. 1 shows a typical arrangement for a ring station with bypass switch. In normal operation, the station can perform both self-diagnostic checks on its control mechanism, and, via the bypass switch, loop-around tests on its own transmitter and receiver functions. On detection of a fault, a station operates the bypass switch by a signal on the bypass control wire, removing itself from the ring.

However, occasionally a station's self-diagnostic capabilities fail, or fail to react correctly to a fault situation, or the bypass switch itself fails, so that the whole system suffers. Although an individual station's failure rate due to these causes may be typically one failure in 50 years, in large systems with, say, upwards of 100 stations served by a ring, the resulting aggregate averages failure rate of worse than two catastrophic system failures per year is unacceptably high.

In a system embodying the present invention, Fig. 2, each user station on the ring has an active bypass switch such as 1 housed at a distribution frame remote from its station.

The frame thus provides a central location for all of the bypass switches for the ring stations, so that the ring is a distributed starring. The reliability of the system is enhanced by enabling a given station on the ring to operate not only its own bypass switch, but also that of its nearest upstream neighbour. In the event of the loss of acceptable transmission on its ring input, a station may operate its neighbour's bypass switch to receive input from its next nearest neighbour, effectively isolating its nearest neighbour. If desired, the technique can easily be extended to allow a station to operate the bypass switch of its next-but-one neighbour, or its next-but-two neighbour, and so on. This may be useful where more than two contiguous stations fail simultaneously. Additionally, any station's bypass switch may be operated manually at the frame 2.

In this remote bypass arrangement, the transmission medium between the stations and the bypass switches at the frame could be either optical fibre or metallic, depending on the requirements in respect of ring bit rate, freedom from electromagnetic interference, etc. The two bypass control lines per station, which carry only DC conditions ('on' or 'off') are preferably metallic in either case. Since the bypass switches are co-located, the ring medium between bypass switches is short enough also to be metallic (e.g. printed circuit track).

Note that in the system shown, as in the other systems described herein, each station (also called a node) can be a termination via which a number, e.g. 30, user devices have access to the communications medium.

Fig. 3 is a schematic of a remote active bypass switch with dual control suitable for use with a distribution frame. It operates as follows: (a) Under normal conditions, the bypass control (and the manual bypass) inputs to the OR gate G3 from the associated station and the downstream station are inactive. The A/B select output from the OR gate G3 is thus set to select the 'A' inputs to the two gates G1 and G2. Input from the ring received via G1 is sent by the transmitter TX 1 over the link to the station where it is received by the receiver RX1.

Output from the station is sent by the transmitter TX2 over the link to the remote bypass switch, received by receiver RX2, and sent to the next downstream bypass switch via gate G2.

(b) With any of G3's bypass inputs active, the A/B select output is set to select the 'B' inputs of gates G1 and G2. Input from the ring is thus routed via gate G2 to the ring output, bypassing the station. Output from the station is looped around to the station's input via gate G1, permitting the station to perform a self-test function on its transmit and receive components.

In such a system, in which a single bypass switch may be operated by its parent station or by the nearest downstream station, a failure in the bypass switch could affect the whole system. Thus gates G1 and G2 and the OR gate G3 are the critical bypass switch components since they are effectively in series with the ring. Clearly, the details of the mode of any failure in these components determines whether or not the results are likely to be catastrophic for the ring. Pessimistically assuming that all failures in these components are catastrophic, the ring system suffers approximately one failure every 7600 years per station (based on figures supplied by 'Handbook of Reliability Data for Electronic Components Used in Telecommunications System s'-Post Office Telecommunications May 1980).For systems of 100 stations, a mean time between catastrophic failures of approximately 76 years is thus obtained. It should be noted that other failures-whether in the station itself or in other bypass switch components-do not prevent the downstream station from operating the bypass switch and so are not catastrophic.

For systems very much larger than 100 stations, or where the MTBF figure obtained by means of Dual Bypass Control is inadequate, an arrangement using so-called braided bypass control can be used. As before, each user station on the ring has an active bypass switch housed remotely from its parent station at a distribution frame. Also as before, a station can perform self-diagnostic tests and operate its bypass switch on detection of an internal failure. However, the present bypass switches are more complex than those described above, as they provide a bypass braiding capability.

Using braided bypass control, each bypass switch is connected via the ring medium to bypass switches other than those belonging to its nearest neighbours, see Fig. 4. This shows the simplest form of bypass braiding, with bypass switches connecting to nearest and next-nearest neighbours, but more complex arrangements are possible. The output transmission from each station is relayed by the station's bypass switch to its nearest neighbour bypass switch via the ring primary, and to its next-nearest neighbour bypass switch via the ring secondary. Each station can thus receive transmissions from either its nearest or its next-nearest neighbour station.

With the system functioning normally, a station receives transmissions only on its ring primary. In the event of malfunction causing loss of acceptable transmission on the primary input, a station signals its bypass switch via the primary/secondary select line to switch inputs to receive transmissions on the ring secondary. Thus a faulty station, bypass switch or ring segment may be isolated to restore service to the rest of the system.

Fig. 5 shows a schematic of an actice bypass switch suitable for use with braided bypass control. The switch operates as follows: (a) Under normal conditions, the primary/secondary control from the station is set such that gate G1 accepts input from the ring primary via receiver RX 1. Then, with the station bypass control and the manual bypass control both inactive, the A/B select output from the OR gate G4 is set to select the 'A' inputs to the two gates G2 and G3. Input from the ring primary received via G1 is sent via G2 and the transmitter TX 1 to the station's receiver RX3.

Output from the station is sent by the transmitter TX2 over the link to the remote bypass switch, received by receiver RX4 and sent via G3 and transmitters TX3 and TX4 to the nearest and next-nearest downstream bypass switches, respectively.

(b) Under fault conditions, with loss of acceptable transmission on the ring primary input, the station changes the condition signalled by the primary/secondary select line to cause gate G1 at the bypass switch to select the ring secondary input via receiver RX2. In this manner, the nearest neighbour station, bypass switch and ring segment are effectively isolated to remove the fault condition. Apart from this change, the remainder of the system continues to operate normally.

(c) With either of G4's bypass inputs active, the A/B select output is set to select the 'B' inputs of gates G2 and G3. Input from either the ring primary or the ring secondary, as selected by the primary/secondary select line, is thus accepted by G1 and routed via gate G3 and transmitters TX3 and TX4 to the two ring outputs, bypassing the station.

Output from the station's transmitter TX2 is received at the bypass switch by RX4 and looped around to the station's receiver RX3 via gate G2 and transmitter TX 1, permitting the station to perform a self-test function on its transmit and receive components.

The advantage of such a system is that a catastrophic system failure requires simultaneous failures to occur in two or more adjacent station assemblies (a station assembly consists of a ring station, a bypass switch and a ring segment). The probability of occurrence of this event is so low that, even for systems with may hundreds of ring stations, the theoretical mean time between catastrophic system failure can be demonstrated to be many hundreds of years. Because the bypass switches are co-located, the cost of providing the extra ring medium for the braiding capability is negligible (implementable as printed circuit track).

The virtues of the distribution frame when used with remote bypass switches, in addition to the enhanced reliability, are the greatly increased system flexibility and maintainability that result. Deleting a station is accomplished easily, involving only the removal of a connector at the distribution frame or the manual operation of a bypass switch. The addition of new stations to the ring is accomplished simply by running the required cabling from the station out to the distribution frame and connecting to a spare bypass switch connector.

No cabling changes to other stations already on the ring is required. If a system grows beyond the capacity of a single ring, it is relatively easy to divide the ring and the frame at a suitable point to provide two separate rings each with its own frame. The two rings are then interconnected by a bridge station which accesses each ring via bypass switches at the two distribution frames. Fig. 6 illustrates this arrangement. System growth, from the addition of a single station to the addition of a new ring, is thus provided smoothly and at minimum extra cost.

Such a bridge station can follow the principles of a our Application No. 2138651 (R.T.

Swindle-R.F. Rous 7-3).

Maintainability is improved by providing simpie indicators at the distribution frame to show which stations are active and which are bypassed, requiring maintenance action. In addition, stations may be taken out of service selectively at the distribution frame to facilitate fault-finding. Finally, the distribution frame provides a secure power source for a number of remote active bypass switches in a more-effective manner than can be provided by stations for individual local bypass switches.

The ring distribution frame discussed above can be extended to form a hierarchical arrangement, see Fig. 7. Here, a number of station distribution frames (SDFs) is shown, each housing the active bypass switches for a number of stations-as described above. The stations served by a SDF can be arranged by function, e.g. department or division, or by location, e.g. floor of a building, or by any other desired criterion.

Each SDF has associated with it an active bypass switch which connects across the ring medium between the input and output to/from the SDF. The bypass switches for a number of SDFs are housed remotely at a group distribution frame (GDF). A bypass control function at the GDF, common to all the SDF bypass switches served by that GDF, is responsible for the operation of the SDF bypass switches.

Thus, any desired group of stations may be selectively removed from service by operation of the relevant SDF bypass switch at the GDF.

In its simplest form, the GDF bypass control function may provide for manual-only operation of the SDF bypass switches.

As shown the SDF/GDF hierarchy can be extended by the addition of a higher level of distribution frame. Thus a number of GDFs is served by a network distribution frame (NDF).

Bypass switches associated with the individual GDFs are located remotely at the NDF so that the NDF bypass control function is responsible for the operation of the bypass switches associated with any GDF. Any desired 'supergroup' of stations may thus selectively be removed from service by suitable action at the NDF.

The advantages conferred by the distribution frame hierarchy are enhanced security, flexibility and maintainability, especially for the largest systems. Progressively smaller groups of stations down to a single station are removed from service selectively, with ease of simplicity. Fault isolation can thus be enforced at single station, group or supergroup level, as necessary to ensure a continuing service to the rest of the system. Fault location for maintenance purposes is simplified by the selective manual bypass facility at all levels of hierarchy. The net effect is an increase in system security and a reduction in system downtime.

The optimum number of levels of hierarchy for a given system depends on the size and topology of the system. However, any number of levels (including one only) may be provided in accordance with the requirements in respect of security, flexibility and maintainability. In the event of a system growing beyond the capacity of a single ring, it is a relatively easy task to break the medium at the highest level in the hierarchy to form two separate rings. Each ring would have its own set of SDFs and GDFs and its own NDF (although NDFs could be co-located to save costs). The two rings would then be interconnected by a bridge station which accessed each ring via a bypass switch on separate SDFs. Thus, not only can new stations be introduced onto existing SDFs new SDFs onto existing GDFs and new GDFs onto existing NDFs; new rings can also be introduced to form a multiple ring structure to give any desired system topology, size and bandwidth.

It should be noted that each of the stations of the system described above may in fact serve a number, e.g. up to 32, of individual subscriber stations.

Claims

1. A communication system, in which a number of stations are interconnected in serial manner by a communication medium, in which each said station is provided with a bypass switch which is normally disabled, so that its station is connected in the medium, in which if a fault is detected in a said station or in the medium adjacent thereto that station's bypass switch is enabled so that the bypass switch switches the station out of the medium, in which all of the bypass switches are centrally located at a common distribution frame, and in which each said station is connected to its said by pass switch by physical transmission paths which extend between that station and its said switch.

2. A system as claimed in claim 1, in which each said bypass switch includes a first gate via which the incoming side of the communications medium is normally connected to the station, a second gate via which the station is normally connected to the outgoing side of the medium, and switching means responsive to a fault condition to disable the first gate and to cause the second gate to interconnect the incoming and the outgoing sides of the communications medium.

3. A system as claimed in claim 2, in which when a fault condition has been detected, the station's outgoing and incoming sides can be connected via the first gate for test purposes.

4. A system as claimed in claim 1, 2 or 3, in which said operation in response to a fault condition can be in response to a signal from an adjacent downstream station, and which the detection of a fault condition at the station can be used to effect said disablement at the next upstream station.

5. A system as claimed in claim 2, 3 or 4, and in which the bypass switch has a manual control to enable the bypass switch for test purposes or to remove a station from the system.

6. A system as claimed in claim 1, 2, 3, 4 or 5, and which includes three communications media interconnecting the bypass switching, the first of said media interconnecting all of said bypass switches, the second of said media interconnecting the odd-numbered bypass switches and the third of said media interconnecting the even-numbered switches, and in which when a said fault condition is detected in respect of the first medium one or more bypass switches are operated as appropriate to use the second or third medium in place of the part of the first medium on which the fault condition has been detected.

7. A network which includes two or more systems each as claimed in any one of claims 1 to 6, and one or more bridge stations each of which interconnects two bypass switches in different ones of said systems, so that inter-system communication is possible.

8. A hierarchial network which include networks as claimed in claim 7 forming a first layer of the network, and one or more subsequent layers of the hierarchial network each of which interconnects a plurality of lower layers.

9. A communications system substantially as described with reference to Figs. 2 and 3, or Figs. 4 and 5 and 6 of the accompanying drawings.

10. A communications systems network substantially as described with reference to Fig. 7 of the accomoanying drawings.

New claims or amendments to claim filed on 17 April 1985 New or amended claims:

11. A communications system, in which a number of stations are interconnected in a serial manner by a communication medium, in which each said station is provided with a bypass switch which is normally disabled so that its station is normally connected in the medium, and in which if a fault occurs in a said station or in the medium adjacent thereto which is not detected by that station, the station's downstream neighbour-station recognises an erroneous input from the said faulty station or from the adjacent medium and produces a signal to operate the said faulty station's bypass switch so that the bypass switch switches the faulty station out of the medium.

12. A system as claimed in claim 11, in which all of the bypass switches are centrally located at a common distribution frame, and in which each said station is connected to its said bypass switch by physical transmission paths which extend between that station and its said bypass switch.

13. A communication system, in which a number of stations are interconnected in serial manner by a closed-loop communication medium, in which each said station is provided with a bypass switch which is normally disabled, so that its station is normally connected in the medium, in which if a fault occurs in a said station or in the medium adjacent thereto and that fault is detected in said station that station's bypass switch is enabled so that the bypass switch switches the station out of the medium, in which if a fault occurs in a said station or in the medium adjacent thereto which is not detected by that station that station's downstream neighbour station recognises an erroneous input from the said faulty station or its adjacent medium, that downstream neighbour station produces a signal to operate the faulty station's bypass switch, so that that bypass switch switches the faulty station out of the medium, in which all of the bypass switches are centrally located at a common distribution frame, and in which each said station is connected to its said bypass switch by physical transmission paths which extend between that station and its said switch.

14. A system as claimed in claim 12 or 13, and in which the said physical transmission paths which connect the bypass switches to their respective stations are physically separate from the communication medium.

15. A system as claimed in claim 11, 12, 13 or 14, in which each said bypass switch includes a first gate via which the incoming side of the communications medium is normally connected to the station, a second gate via which the station is normally connected to the outgoing side of the medium, and switching means responsive to the detection of a fault condition to disable the first gate and to cause the second gate to interconnect the incoming and the outgoing sides of the communication medium.

16. A system as claimed in claim 15, and in which a fault condition at a said station has been detected, the station's outgoing and incoming sides can be connected for test purposes via the first gate.

17. A system as claimed in claim 11, 12, 13, 14, 15 or 16, and in which the bypass switch has a manual control to enable the bypass switch for test purposes for the removal of a station from the system.

18. A system as claimed in claim 11, 12, 13, 14, 15, 16 or 17, and which includes three communications media interconnecting the bypass switching, the first of said media interconnecting all of said bypass switches, the second of said media interconnecting the odd-numbered bypass switches and the third of said media interconnecting the even-numbered switches, and in which when a said fault condition is detected in respect of the first medium one or more bypass switches are operated as appropriate to use the second or third medium in place of the part of the first medium on which the fault condition has been detected.

19. A network which includes two or more systems each as claimed in any one of claims 11 to 18, and one or more bridge stations each of which interconnects two bypass switches in different ones of said systems, so that inter-system communication is possible.

20. A hierarchial network which include networks as claimed in claim 19 forming a first layer of the network, and one or more subsequent layers of the hierarchial network each of which interconnects a plurality of lower layers, in which a plurality of stations of a lower layer of the network may be bypassed simultaneously by the operation of a single bypass switch on a higher layer of the network, e.g. in response to the detection at the higher layer of a persistent fault condition within the said plurality of stations, and in which such bypassing may be effected by the operation of one or more bypass switches on successively higher layers of the network, if called for by fault conditions.