Recommendations on Control Circuit Protection Scheme

Back ground: A two member committee was formed by Railway Board vide letter No. No.2010/Tele/9(1) dated 19.01.2012 with following terms of reference: Identify most suitable protection scheme for control communication circuits including SCADA. This report has been prepared in compliance to these directives. 1. Introduction and Scope The control circuits are worked on 64 kbps channels configured in conference mode and multiplexed by Primary Add-Drop Multiplexers (to be referred as PD-MUX henceforth ) and resultant E1 is carried on STM1 on OFC backbone. This E1 thread extends to the control section boundary/ division boundary. Time-slots other than those carrying omnibus control circuits carry point-to-point circuits. The PD-MUXes are in linear topology, with E1 level ring protection on STM backbone on OFC network at the end of the thread. The committee has examined protection schemes followed in various PD-MUXes. Also, certain common issues that are affecting reliability/availability of control circuits irrespective of protection scheme are discussed. Based on this study, features of the most suitable protection scheme taking care of reliability/availability as well are identified. Since a large number of PD-MUXes of various makes and versions are already working, ways to implement/migrate to the most suitable protection scheme on various types of PD-MUXes are recommended. In addition to this, certain other recommendations related to monitoring and periodic testing of control circuits to ensure availability are also included in the report. As IP based control communication is still under development, recommendations given in this report are relevant in the present context of Primary Multiplexer based control communication. 2. Protection of control circuits on OFC sections Ring/ loop protection mechanism is in general implemented in two different ways: Approach-A : Use of LPC (Loop protection card) and 2 E1s at every station for all time-slots protection Approach-B : Ring protection mechanism using spare time slots in working E1 These are discussed below 2.1 Approach-A : Use of LPC and 2 E1s at every station for all time-slots protection Please refer figure-1
Protection scheme for Control Circuits Page 1

In this approach, the E1 carrying 30 time-slots and designated as working E1 is taken from East-1 port of LPC (Loop Protection Card) of PD-MUX. This is carried over STM1 Add-drop MUX to next station and connected to West-1 port of LPC card of PD-MUX. Like-wise the E1 thread continues. Another E1, designated as protection E1, taken from East-2 port of LPC of PD-MUX is also carried via the same Add-Drop STM1 to next station and connected to West-2 port of LPC of the PD-MUX. Like-wise, the second thread of E1 continues. From the terminal station, E1 from loop-back port of LPC is carried via protection path on STM back-bone on different route to the originating station and connected on loopforward port of LPC. Protection on STM backbone via different route

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

LPC

LPC

LPC

LPC

LPC

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PD-MUX

PD-MUX

PD-MUX

PD-MUX

PD-MUX

Station-A

Station-B

Station-C

Station-D

Station-E

Fig.1 Use of LPC and 2 E1s at every station for all time-slots protection

Protection scheme for Control Circuits

Page 2

In the above figure, STM1 Add-Drop MUXes are not shown to avoid congestion In case of failure of E1 due to port failure/OFC cut between terminal station and immediate next station (for example between station-A and station-B) switching in LPC takes place as shown in fig.2 : a. At station-A, PCM2-to-E1 and LF-to-E2 connections are interrupted; instead, LFto-PCM2 is connected in LPC b. At station-B, W1-to-PCM1 and W2-to-E2 connections are interrupted; instead, PCM1-to-E2 is connected in LPC Protection on STM backbone via different route

L B W 2 W 1

LF E 2 E 1

L B W 2 W 1

LF E 2 E 1

L B W 2 W 1

LF E 2 E 1

L B W 2 W 1

LF E 2 E 1

L B W 2 W 1

LF E 2 E 1

LPC

LPC

LPC

LPC

LPC

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PD-MUX

PD-MUX

PD-MUX

PD-MUX

PD-MUX

Station-A

Station-B

Station-C

Station-D

Station-E

Fig.2 Switching in LPC in case of failure of E1/ OFC cut between terminal station and immediate next station

Protection scheme for Control Circuits

Page 3

In case of failure of E1 due to port failure/OFC cut between two intermediate stations, (for example between station-C and station-D) switching in LPC takes place as shown in fig.3 : a. At station-C, PCM2-to-E1 and W2-to-E2 connections are interrupted; instead, W2-to-PCM2 is connected in LPC b. At station-D, PCM1-to-W1 and W2-to-E2 connections are interrupted; instead, PCM1-to-E2 is connected in LPC Protection on STM backbone via different route

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

LPC

LPC

LPC

LPC

LPC

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PCM2

PCM1

PD-MUX

PD-MUX

PD-MUX

PD-MUX

PD-MUX

Station-A

Station-B

Station-C

Station-D

Station-E

Fig.3 Switching in LPC in case of failure of E1/ OFC cut between two intermediate stations

Protection scheme for Control Circuits

Page 4

Thus, in this arrangement of using LPC and 2 E1s at every station, all time-slots of the E1-thread are protected. Loop-back protection is end to end on STM back-bone via different route. General terms are used for description of ports. The above arrangement is used by PUNCOM PD-MUXes. Features: The scheme has following features. It offers protection to all the 30 channels. The switching to alternate E1 takes place simultaneously at two stations. (On either side of the cable cut section). Therefore the switching can be made very fast. The protected sections can be subdivided into smaller sections to take care of multiple cable faults without the use of additional PD-MUX.

Problem of overreach: In case of multiple cable cuts in a single protected section, the stations on either side of the isolated section (section isolated by two cable faults) get directly connected. This will cause extension of point to point circuits between these two non-adjacent stations which were meant between adjacent stations only. Figure 4 shows how station-C and station-C1 get isolated and also over-reach of timeslots from station-B to station-D Consider time slot 20 is used for working point to point telephone between stations A and B, and the same time slot is used for working Point to point telephone between station C and D, then in case of faults between station B and C & between stations C and C1, direct connection between point to point telephones at stations B and D will be established. This fact has to be considered if such telephones are being used for train working. Similar problem shall occur on other point to point circuits. Special precaution need to be taken if on point to point circuits, axle counters channels are worked. To avoid this, ether the time slot or the address setting for modem to modem communication must be different for axle counters in a section. For example if two axle counters are using same time slot 21, then they must not use the same address setting.

Protection scheme for Control Circuits

Page 5

Protection on STM backbone via different route

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

L B W 2 W 1

L F E 2 E 1

LPC

LPC

LPC

LPC

LPC

LPC

PCM 2

PCM 1

PCM 2

PCM 1

PCM 2

PCM 1

PCM 2

PCM 1

PCM 2

PCM 1

PDMUX

PDMUX

PDMUX

PDMUX

PDMUX

PDMUX

Station-A

Station-B

Station-C

Station-C1

Station-D

Station-E

Fig.4 Multiple Cuts Scenario : Isolation and overreach problems

2.2

Approach-B : Ring protection mechanism using spare time slots in working E1 Please refer fig.5 In this approach, all the conference circuits are always protected whereas only those point-to-point circuits get protected which are mapped to spare time slots in E1 for protection purpose. Therefore, in this approach, the protection has to be planned at time slot level for point to point circuits. This approach is followed in Webfil PD-MUX and Nokia PD-MUX having loopprotection module. In case of older versions of Nokia/ Webfil where automatic loopback is not supported, protection E1 is kept ready for patching at test-room

Protection scheme for Control Circuits

Page 6

Protection on STM backbone via different route

PDMUX P P
1 2

PDMUX P P
1 2

PDMUX P P
1 2

PDMUX P P
1 2

PDMUX P P
1 2

Station-A

Station-B

Station-C

Station-D

Station-E

Fig.5 Ring-protection using spare time-slots in Working E1

Problem of over reach: This scheme shall also lead to overreach in case of multiple cable cuts as explained in case of PUNCOM MUX. However, this scheme allows for protection of selected time slots only. Therefore the overreach in sensitive circuits like station to station hot line telephone can be avoided by disabling automatic protection for those circuits while continue to have automatic protection for control circuits. In case of applications like axle counter which require automatic protection switching for improved reliability purpose, judicious time slot allocation can prevent such problem.

3.

Common issues affecting control circuit reliability/availability Irrespective of the approach being followed for protection, certain common issues that are affecting reliability/availability of control circuits are discussed below.

Protection scheme for Control Circuits

Page 7

3.1

Common E1 thread for adjacent control sections In some of the sections, the E1 thread carrying control circuits are not coinciding with control section boundaries. For example, in figure 6, A-P is one control section and PB is another, but they are worked on common E1-thread from A to B. Therefore, any failure of PD-MUX on section A to B is affecting entire thread.

Figure 6 Problem of Common E1 thread for adjacent control sections

3.2

Delay in Protection Switching on PD MUX NMS Automatic Protection Switching on STM backbone is within 50 msec. However, there is no laid down time for automatic protection switching for PD MUX resulting in such time being in excess of 1.5 second, which may result in failure of axle counters using point to point voice channel.

3.3

Non-uniformity of PD-MUXes on multi-segment control section On multi-segment control sections, variety of PD-MUXes are in place. For example, on control sections where spur links take-off or on sections involving junction stations, different segments of the same control in different directions from the junction station may have different PD-MUXes : on one segment-Nokia, on anotherWebfil; on yet another segment PUNCOM . Such a situation arises due to commissioning of PD-MUXes on these segments under different works. Such an arrangement complicates time-slot management on the control section.

Protection scheme for Control Circuits

Page 8

3.4

Inadequate/Improper monitoring of PD-MUX NMS In some of the divisions, PD-MUX NMS is not located in test room and in some cases, it is not being monitored on 24x7 basis. Consequently, interpretation of alarms and corrective action, pin-pointing the faulty section are delayed.

3.5

Over-reach of point-to-point circuit in case of PD-MUX failure/switch-off In case of failure of PD-MUX/switch-off or pass-through of KLM carrying the E1 thread on STM1, over-reach of time-slots is inevitable. Consequently, point-to-point circuits across stations may be carried to unwanted destination.

4.

Features of most suitable protection scheme Considering the features of the two approaches discussed in section-2 and reliability/availability issues discussed in section-3, features of most suitable protection scheme are identified as : 1. E1 thread carrying control circuits on work-path shall terminate at control section boundary; it shall not be continued on adjacent control section. 2. However, other omnibus circuits like EC, SCADA etc. are to be patched at VF level across control section boundaries and picked-up in the fresh E1 thread. 3. E1 level ring protection on STM backbone on OFC network at the end of the thread shall be via different path. 4. Also, wherever feasible, additional protection path (3rd path) shall be provided on STM backbone to ensure availability of circuits in case of simultaneous cuts on Working Path and Protection Path. 5. Protection has to be automatic through PD-MUX NMS 6. Overreach problem is to be avoided in case of PD-MUX failures/switch-off/KLM pass through on STM1 7. Overreach problem is to be avoided in case of multiple OFC cuts 8. Smaller sections for loop protection to be implemented to the extent possible so that multiple cuts scenarios have better availability of control circuits 9. Uniformity of PD-MUXes on all segments to be ensured on multi-segment control section Since there is large installation base of PD-MUXes of various types/versions and no further technical developments are taking place in this segment, migration to the above scheme of most suitable protection is discussed in the next section for all existing types/versions of PD-MUXes

Protection scheme for Control Circuits

Page 9

5. 5.1

Migration to Most Suitable Protection Scheme Sections with PUNCOM PD-MUX or WEBFIL PD MUXes or Nokia PD MUXes as per IRS TC 68/04 The protection scheme followed in these MUXes creates overreach problem of pointto-point circuits in case of multiple OFC cuts. Hence, channeling plan with enough spare slots and judicious staggering of channels should be followed. If numbers of channels are more, 2nd PD-MUX chain can be planned.

5.2

Sections with old versions of MUXes which does not support automatic loop protection In case of old version of Webfil MUX as per (IRS TC 68/97), automatic protection scheme can be implemented by replacing NIM and TRIB cards in the terminal MUXEs only. The NIM and TRIB cards of Webfil make are to be taken as per newer version of MUX i.e. as per IRS TC 68/2012. OR External LPM (Loop protection module) can be procured and protection implemented as shown in Fig.7. Such modules are available with M/s PUNCOM and M/s Webfil. However, as this scheme provides complete time-slot protection, which can cause overreach of point-to-point circuits, time-slots shall be programmed leaving spare slots to avoid overreach.
WEST DIRECTION EAST DIRECTION

Protection via STM backbone on different route

Stn A STM-1

Stn B STM-1

Stn C STM-1

Stn D STM-1

Stn E STM-1

Stn F STM-1

N S(1)

S(1)

N S(1)

S(1)

N S(1)

S(1)

N S(1)

S(1)

N S(1)

S(1)

N S(1)

S(1)

P4 P6 P5 P3

P4 P6P5 P3

P4 P6 P5 P3

P4 P6 P5 P3

P4 P6 P5 P3

P4 P6 P5 P3

P2

LPM

P1

P2

LPM

P1

P2

LPM

P1

P2

LPM

P1

P2

LPM

P1

P2

LPM

P1

P8

P7
P-1

P8
P-2

P7
P-1

P8
P-2

P7
P-1

P8
P-2

P7
P-1

P8
P-2

P7
P-1

P8
P-1

P7

E/T Mux

D/I Mux

D/I Mux

D/I Mux

D/I Mux

E/T Mux

Fig.7 Use of External LPM Module on PD-MUX not supporting loop-protection

5.3

Sections with Coral PD-MUX As mechanism for loop protection is not available in Coral PD-Mux, external LPM module of Webfil or LPC module of PUNCOM can be procured and protection implemented as shown in Fig.7. However, as this scheme provides complete time-slot protection, which can cause overreach of point-to-point circuits, time-slots shall be programmed leaving spare slots to avoid overreach.

6.

Summary of recommendations Considering the different protection schemes followed as described in section-2 and the issues reliability/availability as described in section-3, the committee recommends the most suitable protection scheme with features as described in section-4 and implementation strategy described in section-5. Accordingly, following action-plans by Zonal Railways, RailTel and RDSO are recommended:

6.1

Action by Zonal Railways a. All Zonal Railways shall ensure uniformity of PD-MUX equipment on one control segment. b. Instead of single thread E1 covering entire section, additional PD-MUX shall be provided at control circuit boundaries (please refer figure 8). By providing additional PD-MUX, different E1s can be used for two different control sections. This work should be carried out by the divisions themselves as PD-MUX NMS and channeling plan are under their control.

Figure 8 Additional PD-MUX at Control Boundary c. Migration to most suitable protection scheme as described in section 5 to be implemented as per the make of PD-MUXes. d. PD-MUX NMS of every control section of the division should necessarily be located in Test room and be manned round the clock by staff trained in the NMS.
Protection scheme for Control Circuits Page 11

e. Smaller sections for loop protection may be implemented to the extent possible so that multiple cuts scenarios have better availability of control circuits f. PD-MUX and its NMS to be maintained by Divisions only as many point-to-point circuits are concerned with safety of train operations. g. Since it is only E1 level handover to PD-MUX and no scope for channel-wise visibility / alterations within E1 by RailTel and also SDH NMS is with RailTel with various options on STM backbone for protection of E1s, Railways may consider handing over of STM1 maintenance to RailTel on mutually agreeable terms and conditions. 6.2 Action by RailTel a. RailTel should provide protection path via SDH backbone from each control circuit boundary to test-room. This needs to be provided on a path different than working path of the control-E1 thread so that OFC cut shall not affect the protection E1. RailTel NOC should share protection path details of control circuit E1s with concerned Zonal Railways. Wherever feasible, RailTel shall provide alternate path in addition to back-bone protection path for the control circuits E1s so that availability of control circuits is greatly improved RailTel should share fault escalation numbers and contact persons to respond in case of control circuit failures. Top most priority shall be given to restoration of control circuits. Monthly testing program of control circuits protection shall be prepared and implemented considering the number of divisions being served by RailTel-NOC. This can be carried out from Test Rooms by asking for port-down on WP (Working Path) in the E1 thread and checking for protection-E1 availability. RailTel should provide required STM backbone protection when Railways are carrying out the work of split-up of E1 thread at control boundaries, split-up into smaller loops etc.

b. c.

d.

e.

f.

6.3

Action by RDSO: Telecom Directorate-RDSO should standardize the protection switching time for PD MUX in consultation with Signal Directorate-RDSO and amend the specification suitably. RDSO to issue typical channeling plan for various cases of PD-MUX deployment once the present report is accepted by Railway Board.

(P.V.Sreekanth) General Manager(NPM)RailTel

(Yashpal Singh Tomar) Director Tele/RDSO