Chapter 5 - Message Transfer Part (MTP)

Chapter 5 Message Transfer Part (MTP)

		The Message Transfer Part (MTP) acts as the carrier for all SS7 messages in TDM-based networks, providing reliable transfer of messages from one signaling point to another. This function includes levels one, two, and three. In addition to providing signaling point to signaling point communications, the MTP also provides error detection and correction.

		The methodologies of MTP are very similar to those used in other bit-oriented protocols (BOPs), such as X.25. Sequence-numbering and error-checking mechanisms are very similar.

		The signal unit structure used in all SS7 messages provides all the information required by MTP level two and level three. Flow control is provided through the use of a special signal unit called the Link Status Signal Unit (LSSU), described in this chapter.

		MTP is defined in the ITU-TS documents Q.701 through Q.704, Q.706, and Q.707. They can be found in Telcordia document TR-NWT-000246, Volume One, Chapter 1.111.1 through 1.111.8. ANSI publications referring to the MTP protocol are numbered T1.111-1992, ''Functional Description of the Signaling Message Transfer Part (MTP)."

		The Telcordia recommendations add reliability and versatility to the network. The Telcordia publications are almost identical to the ANSI and ITU-TS publications, other than the additions made by Telcordia.

		Description of MTP

		The MTP provides all functions of layers one, two, and three in the OSI model. We have already discussed the types of interfaces used at level one of the SS7 network. These are industry-standard interfaces and do not necessarily require in-depth discussion here. Level two of MTP provides error detection/correction, as well as error checking through the check bit field. In addition to these two fundamental functions, the following is provided:

		Signal unit delimitation

		Signal unit alignment

		Signal unit error detection

		Signal unit error correction

		Signaling link initial alignment

		Signaling link error monitoring

		Flow control

		Signal Unit Delimitation

		Every signal unit is preceded by a flag. The flag is an eight-bit pattern, beginning with 0 and followed by six consecutive, 1s, ending in 0 (01111110). The flag is used to signify the beginning of a signal unit and the end of the preceding signal unit. While the protocol actually allows both an opening and a closing flag, in the U.S., only one flag is used.

		Signal unit delimitation is important to the upper layers. In most networks, there is traffic constantly flowing through each signaling point, even though the messages may not contain any information.

		Signal Unit Alignment

		A link is considered in alignment when signal units are received in sequence, without ones-density violations, and with the proper number of octets (based on the message type). The signal unit must be a total length of eight-bit multiples. If the signal unit is not in eight-bit multiples, or if the signaling information field (SIF) of a Message Signal Unit (MSU) exceeds the 272-octet capacity, the signal unit received is considered in error.

		The link is not taken out of service until there has been an excessive number of errors. This is determined by a counter, the signal unit error rate monitor (SUERM). This counter is used to count the total number of errors on a signaling link. Each link keeps its own unique counter.

		The purpose of the counter is to determine when an excessive number of errors has occurred (64) and take the link out of service. The type of errors is limited to alignment errors. The typical cause of alignment errors is usually clock signals not being properly synchronized on both ends of a link.

		The network management procedure at level three is responsible for realigning the link (by taking it out of service and resynchronizing the link). Level two is responsible for reporting any errors to level-three link management.

		When the link is taken out of service, the link must be tested for integrity before it is available for MSUs again. This process is known as the alignment procedure. There are two types of alignment procedures used: normal alignment procedure and emergency alignment procedure. Both these methods are discussed in greater detail later in this section.

		Signal Unit Error Detection

		Errors are detected using the check bit field and the sequence number of the signal unit. If the check bit field is in error, the signal unit is discarded and a negative acknowledgment is sent to the originating signaling point. An error is also counted by the signal unit error rate monitor (SUERM).

		The level-two timer T7, "Excessive delay of acknowledgment," prevents a signaling point from waiting too long for a positive or negative acknowledgment. Usually, an acknowledgment is sent when a signaling point becomes idle and does not have any more traffic to transmit. When congestion occurs at a signaling point, or an extreme amount of traffic is present, it is possible that T7 could time out and force retransmission of messages.

		The recommended value for T7 is 11.5 seconds. This, of course, depends on the network. While the actual value of T7 is optional, the timer is usually a nonadministrable timer, which means that once it is set, it cannot be changed by system administration.

		Signal Unit Error Correction

		When an error is detected in a signal unit, the signal unit is discarded. Level-two MTP counts the error (SUERM or errored interval monitor) and requests a retransmission if basic error control is being used. Preventative cyclic retransmission (PCR) treats errors differently and is explained in greater detail later in this chapter.

		When excessive errors are detected on any one link, the link is taken out of service. The link is then placed through an alignment procedure to test the link and place it back into service automatically. The link will not be placed back into service until it has passed the "proving period" of the alignment procedure.

 

Signaling Link Error Monitoring

		Three types of error rate monitors are used. Two are used when the link is in service and the other is used when the link is going through an alignment procedure. The SUERM and the errored interval monitor are used while the link is in service, and are often referred to as the "leaky bucket" technique, because of the way they decrement the counter after n number of good signal units or intervals.

		With the SUERM, each signal unit received with an error increments a counter. Every 256th signal unit received without error decrements the counter (hence the nickname "leaky bucket"). When the counter reaches a value of 64, the link is removed from service and alignment procedures begin. This method is used with 56/54-kbps links.

		When 1.536-Mbps links are used, the errored interval monitor method is used. The link is monitored for a determined period of time (defined by Telcordia as 100 milliseconds). If a flag is lost during the interval, or a signal unit received in error, then the interval is considered in error. An errored interval increments a counter, in the same fashion as the SUERM. The counter is decremented when 9308 intervals have passed without error (also defined by Telcordia). When the counter reaches a value of 144,292 intervals, the link is removed from service and alignment procedures are started. The actual values may be different, especially in international networks, but these are the recommended values defined by Telcordia.

		The alignment error rate monitor (AERM) is used during the alignment procedure and is an incremental counter. Each time an error is encountered during alignment, the counter is incremented by 1. When the AERM determines that there have been excessive errors, it causes the link to be taken out of service, and the alignment procedure begins again.

		Flow Control

		Flow control allows traffic to be throttled when level two becomes congested at a distant signaling point. The LSSU is used to send congestion indications to the transmitting nodes. When the LSSU of busy is received, the receiving signaling point then stops sending MSUs until the congestion is abated.

		Flow control also uses a priority for signal unit types to ensure that important signal units such as MSUs are transmitted, even during a congestion condition. Congestion is not the only condition indicated by flow control. Processor failures are also indicated by level two, meaning that level two can no longer communicate with levels three or four.

		Flow control at this level should not be confused with traffic management at level four. Level-two flow control is isolated to individual links and does not indicate status of the signaling point. In addition, flow control at level two provides priority consideration for signal units, with no regard to the user part.

		This is different from level-three network management, which does give consideration to the user of a signal unit. Level-three network management controls the flow of messages to a particular level-four user, while flow control at level two controls the flow of messages to a link processor.

		If the congestion condition should continue, the link will be taken out of service and realigned, using the alignment procedure. This prevents a link from becoming "locked" in the congestion state.

		Structure of MTP Level Two

		The signal unit components used by level two are shown in Figure 5.1. These components can be found in all three types of signaling units. The backward sequence number and the forward sequence number are used for sequencing packets and are used by level two to ensure that all transmitted packets are received. They are also used for positive and negative acknowledgments. The indicator bits are used to request a retransmission. The length indicator allows level two to determine the type of signal unit being sent, and the cyclic redundancy check (CRC) field is used to detect data errors in the signal unit.

		Figure 5.1 Level-two MTP is responsible for link functions, which include alignment of the link and reporting link status to MTP level three.

 

		Flag

		The flag is used to indicate the beginning of a signal unit and the end of a signal unit. As mentioned in the previous chapter, the flag in U.S. networks is used to indicate both the beginning of one signal unit and the end of another. In some other networks, there can be both an opening and a closing flag.

		The flag bit pattern can be duplicated within the information field of a MSU (ones-density violation), causing an error to occur. To prevent the data from duplicating the flag pattern, the transmitting signaling point uses bit stuffing. Bit stuffing is the process of inserting an extra bit before transmission after every series of five consecutive 1s. The bit value is always 0.

		By inserting a 0 after every five consecutive 1s pattern before transmitting a signal unit, the transmitting signaling point can ensure that there is never an occurrence of six consecutive 1s except for the flag, which is inserted just before transmitting the signal unit.

		The receiving signaling point, upon receipt of a signal unit with five consecutive 1s, removes the inserted 0s from the signal unit. Because this is an absolute rule, a 0 is always going to be inserted after five consecutive 1s, and always removed after five consecutive 1s.

		When a ones-density violation occurs, the signal unit is considered out of alignment and level three is notified of a link failure. The link then enters what is referred to as octet counting mode. During this mode, every octet is counted, and the number of errors per octet is monitored rather than errors per signal unit. The link is then taken out of service (OOS) and put through an alignment procedure.

		Sequence Numbering

		The SS7 protocol uses sequence numbering like many other layer-two bit-oriented protocols. The SS7 technique is just a little different from, say, X.25, but the principle is the same. There are several ways in which sequence numbering is achieved. When 56-kbps links are used, a seven-bit sequence number is used. Both a forward sequence number (FSN) and a backward sequence number (BSN) are used on these links (explained later). The sequence number used on 1.536-Mbps links is 12 bits in length (if MTP level two is used on these links). If ATM links are used, MTP is replaced by the SAAL protocol, which uses a 24-bit sequence number.

		Here is how sequence numbering works within MTP level two. The FSN indicates the number of the signal unit now being sent. This sequence number is incremented by 1 after every signal unit transmission, except in the case of the FISU or the LSSU. The FISU and the LSSU assume the FSN of the last sent MSU or LSSU, and never increment.

		The BSN is used to acknowledge received signaling units. For example, if sequence numbers 1 through 7 have been sent and received by the distant signaling point, the next signal unit sent by the receiving signaling point could have a BSN of 7, which is acknowledging that all signal units, 1 through 7, were received without error.

		The transmitting signaling point maintains all transmitted signal units in its transmit buffer until an acknowledgment is received. When a signal unit is received, the BSN is examined to determine which signal units are being acknowledged. All acknowledged signaling units are then dropped from the transmit buffer.

		If there are signal units remaining in the buffer unacknowledged, they will remain until timer T7, "Excessive delay of acknowledgment," times out. When T7 times out, a link failure indication is given to level three. This will cause the link to be taken out of service and placed through the alignment procedure.

		Signal units received are checked for integrity (check bit field [CRC]), and they are also checked for proper length. A signal unit must be at least six octets in length, or it is discarded and the error rate monitor is incremented. A negative acknowledgment is then sent to request a retransmission of the bad signal unit. A signal unit's length must be in eight-bit multiples, or the signal unit is in error.

		The Signal Unit Error Rate Monitor (SUERM) is an incremental counter that is incremented by 1 whenever an error is encountered. An error includes signal units received out of sequence, with a bad CRC or one of improper length. After 256 signal units have been received without error (consecutive signaling units), the SUERM is decremented by 1. This technique has earned it the nickname of the "leaky bucket."

		When the SUERM reaches a value of 64 errors, a link failure is reported to level three and the link is taken out of service (OOS) and placed through an alignment procedure. Level three controls the link management function and directs level two during the alignment procedure. Level two does not initiate the alignment procedure; it simply reports the errors and takes direction from level-three link management.

 

Indicator Bits

		The indicator bits are used to request a retransmission. There are two indicator bits: a forward indicator bit (FIB) and a backward indicator bit (BIB). During normal conditions, both indicator bits should be of the same value (0 or 1). When a retransmission is being requested, the signal unit being sent by the signaling point requesting the retransmission will have an inverted BIB. The FIB retains its original value. This indicates to the distant signaling point that an error occurred and retransmission must take place.

		The backward sequence number (BSN) indicates the last signal unit received without error. The receiver of the retransmission request will then retransmit everything in its transmit buffer with a sequence number higher than the backward sequence number of the retransmission request. The indicator bits in the retransmitted message are independent of the originating exchange's indicator bits and do not provide any indication. Therefore, they will both be of the same value during the retransmission.

		When the retransmitted signal units reach the distant exchange that originated the retransmission request, an acknowledgment will be sent. The FIB in the acknowledgment will be toggled to match the BIB and will remain at this value until another retransmission request.

		The process we are describing is referred to as basic error detection / correction. These procedures are discussed further in the following section.

		Length Indicator

		The length indicator is used by level two to determine which type of signal unit is being sent. The values of the length indicator can be:

		0 (Fill-In Signal Unit)

		1 or 2 (Link Status Signal Unit)

		3 or more (Message Signal Unit)

		The length indicator should match the length of the field immediately following it, before the CRC field. This field does not exist in the FISU. In the LSSU this is the Status Field (SF), which can be one or two octets in length. If the signal unit is a MSU, there are two fields, the service indicator octet (SIO) and the signaling information field (SIF). The SIO is an eight-bit field, while the SIF is a variable-length field used by level four.

 

The maximum length of the SIF is 272 octets. Obviously, the length indicator (LI) cannot accommodate such a large number. Therefore, whenever the SIF length equals 62 octets or higher, the LI is set to the value of 63. Only level two uses this LI, and its only purpose is to indicate what type of a signal unit is being received. LIs can be found throughout the level-four packet to indicate the length of the variable fields within level four.

		Cyclic Redundancy Check (CRC)

		The CRC field is the last field in the signal unit. This field is calculated using the fields immediately following the flag, up to the check bit field itself. The fundamental process is rather simple. The transmitting signaling point performs the check before bit stuffing and transmission. The remainder is carried in the transmission in the frame check sequence field. The receiving signaling point then performs a similar calculation and compares the remainder to the frame check sequence field of the received signal unit. If there is a discrepancy, the signal unit is discarded and an error is counted in the SUERM. The CRC-16 method of error checking is used in SS7.

		Basic Error Control Method

		So far, we have discussed how a link is placed into service when it is initially started and a little about negative acknowledgments and retransmission. After a link has been placed in service, error detection/correction is used at level two to maintain proper transmission of SS7 messages. There are two methods of error control: basic and preventive cyclic redundancy (PCR). First, we will discuss the basic error control method.

		Basic error detection/correction is used whenever a signaling link uses a terrestrial facility. "Land" links can be of any type of medium it makes no difference to the MTP at this level. Basic error detection/correction is the most favored method.

		Basic error control uses the indicator bits within the MTP portion of the signal unit to request retransmission of signal units received with errors (Figure 5.2). When a MSU is transmitted, the transmitting signaling point sets the FIB and the BIB to be the same. This is the state the indicator bits should always be in when there are no errors. Both the FIB and the BIB should be of the same value, but it makes no difference if the value is a 1 or a 0.

 

		Figure 5.2 This figure illustrates what a retransmission request would look like. The BIB has been changed from its previous value, while the FIB remains the same.

		When an error is detected by a signaling point, the errored signal unit is discarded and a negative acknowledgment is sent to the transmitting signaling point. The negative acknowledgment may use a FISU, a LSSU, or a MSU. The BIB is inverted to indicate a retransmission request. Again, the actual value is of no significance; just the fact that the bit is different from the FIB is of significance here.

		The backward sequence number (BSN) should acknowledge receipt of the last good MSU. The BSN is then used by level two at the transmitting signaling point to determine which signal units in the transmit buffer to retransmit. All those signal units which have been acknowledged are removed from the buffer, and the remaining signal units are then retransmitted.

		When the retransmission begins, the FIB is inverted to match the value of the BIB of the received retransmission request. Both indicator bits should now match again. They retain this value until a retransmission is requested again, in which case the BIB will invert to a value different from the FIB.

		Sequence numbers in the SS7 network can be a value from 0 to 127. This is known as module 128, meaning 128 sequence numbers can be sent before recycling again. The number of sequence numbers that can be sent before an acknowledgment must be received is referred to as the window size.

		The larger the window size, typically, the better the throughput. This is only true, however, when you have reliable transmission. When the link has too many errors, retransmissions can congest the signaling point at the link level. When there is a large number of signal units to retransmit, the problem is compounded.

		The window size of any signaling point is something that must be determined based on the type of transmissions, the capacity of the link, and the average number of retransmissions that occur.

		Figure 5.3 shows a sequence of good transmissions, with positive acknowledgments. The positive acknowledgment is the sending of a signal unit in the opposite direction with an acknowledgment (the BSN) for previously transmitted signal units. When the acknowledgment is received, the signal units can be dropped from the transmission buffer.

 

		Figure 5.3 This figure depicts a successful transmission sequence between two exchanges. The BSN acknowledges receipt of the previously sent messages. Unlike most protocols that indicate what sequence number they expect next, the BSN value is of the last received sequence number.

		It should be mentioned here that an acknowledgment does not have to be received after every signal unit. A common practice in all asynchronous protocols is to allow several signal units to be acknowledged in one acknowledgment.

		Likewise, an acknowledgment could be received for only some of the transmitted signal units. This does not indicate an error. As seen in Figure 5.4, the acknowledgment could come a little later. Sometimes processors get busy and cannot acknowledge everything at once. As long as the timer T7, ''Excessive delay of acknowledgment," does not expire, this does not pose a problem.

		Preventative Cyclic Retransmission (PCR)

		PCR is used whenever satellite transmission is required for signaling links. This is not the favored method, since it uses a higher number of retransmissions than basic error detection/correction does.

		PCR does not use the indicator bits when a retransmission is needed. Instead, all unacknowledged signal units are automatically retransmitted during an idle period. Indicator bits could prove to be a frustrating tool with satellite transmission, because of the propagation delay introduced with this method of transmission.

 

		Figure 5.4 This figure appears as if there may have been an error; however, no error occurred. Signaling point B sent an acknowledgment for sequence number 24 and, then sent acknowledgment of 25 and 26. Timers allow this asynchronous acknowledgment of messages without causing retransmissions.

		The PCR method waits until there are no more MSUs to be sent, and then it retransmits everything in its transmit buffer. Only unacknowledged signal units will be in the buffer, so, in essence, the signaling point is retransmitting what it perceives to be signal units that have not been received.

		If a signal unit has been received, and another of the same sequence number is received, the retransmitted signal unit is discarded. If a signal unit is received with a new signal unit, it is processed as normal. Eventually, the distant signaling point will send an acknowledgment for all received signal units.

		If a signaling point is sending retransmissions and it receives additional MSUs to transmit, it stops the retransmission and sends the newly received MSUs. The distant signaling point must be capable of determining which are new MSUs and which are retransmissions.

		If the transmit buffer of a signaling point should become full, it stops sending any MSUs and sends the entire contents of the buffer. No new MSUs can be sent during this period. If an acknowledgment is still not received, the buffer is retransmitted again, until an acknowledgment is received. This is known as a forced retransmission.

 

To prevent this condition from occurring, a counter is typically implemented to allow the forced retransmission before the buffer becomes full and MSUs become lost. If the buffer is full, MSUs received are going to be discarded, and a busy status will be sent to adjacent signaling points. By using an administrable counter, a threshold can be set to force retransmission when the buffer reaches a predetermined capacity.

		The only rule in PCR is the number of signal units which can be sent without acknowledgment and the number of octets which can be sent without receiving an acknowledgment. No more than 127 signal units can be sent without acknowledgment. The number of octets must not exceed the time to send a signal unit and receive an acknowledgment.

		Figure 5.5 depicts a sequence of events in which several MSUs have been sent, but not yet acknowledged. The transmitting signaling point does not have any more MSUs to transmit, so it automatically begins retransmitting what is in its transmit buffer.

		Figure 5.5 With PCR, the transmitting signaling point maintains all messages in its transmit buffer until an acknowledgment is received. The signaling point retransmits all unacknowledged messages after expiration of a timer, or when there is nothing else to transmit.

 

During the retransmission, a group of new MSUs is generated. This results in the retransmission being stopped and the new MSUs being transmitted. Once these have been transmitted, the retransmission begins again.

		Finally, the distant signaling point has reached an idle state and begins sending acknowledgments back to the originating signaling point. The theory behind this method is that eventually a signal unit will reach the other signaling point, with or without retransmission. The retransmission may flood the link and push the link beyond its 40 percent capacity, but the messages are guaranteed a higher rate of success if continually transmitted in this fashion.

		Structure of the Link Status Signal Unit (LSSU)

		The LSSU as shown in Figure 5.6 is also used at level two to notify adjacent nodes of the status of level two at the transmitting signaling point. Link status is carried over the same link for which it applies, and is not carried over other links (assuming, of course, that the link is functional at level two).

		If level two fails completely, then nothing is received over the link, including FISUs. Level three is notified that acknowledgments have not been received, and level three initiates a link failure recovery. Link recovery is accomplished through the alignment procedure.

		The LSSU is transmitted by the hardware and software associated with a particular link. The link itself may or may not be at fault. The trouble most usually can be found at the termination point of the link-for instance, a link interface card. Since this is where level-two software resides, this should be the first point of maintenance testing. A hardware failure at the link interface card will cause level-two software to initiate LSSUs, as well as software.

		The LSSU provides three types of status information:

		Level two is congested.

		Level two cannot access levels three or four (processor outage).

		The alignment procedure has been implemented.

		Figure 5.6 Components of the LSSU.

 

The status field (SF) is used to provide the information to the adjacent signaling point. The information does not include the link number, which means that this signal unit cannot be used to notify nodes about link status over other signaling links. It pertains only to the link on which it was received.

		Another important concept with the LSSU is that the status is really level two and three status in the transmitting signaling point, rather than the transmission facility. Throughout the various SS7 publications, when they talk about the signaling link status and signaling link functions, they are really talking about the level-two functionality in the signaling point itself.

		The status field is an eight- or 16-bit field, although only three bits are actually used. The three most significant bits provide the actual status information, while the remaining bits are set to 0.

		The status indicator "busy" (SIB) indicates that level two is congested at the transmitting signaling point. When a signaling point receives an SIB, it stops the transmission of all MSUs and begins sending FISUs. If the condition persists for three to six seconds, level three is informed of a link failure. Level three then initiates the alignment procedure to begin on the affected link.

		The congested signaling point will continue sending SIBs at regular intervals (timer T5) until the condition abates. Timer T5 is a value of 80 to 120 milliseconds. Any MSUs already received by the transmitting signaling point will be acknowledged, but no new MSUs should be sent.

		At the receiving signaling point, the timer T7 is reset every time an SIB is received. Timer T7 is used for excessive delay of acknowledgment. If T7 should time out, the link is considered at fault, and level three initiates the alignment procedure. To prevent this from occurring, the T7 timer gets reset after every SIB.

		To prevent an excessive delay caused by a received SIB, upon receipt of an SIB, a signaling point begins timer T6. This timer is used to time the congestion period, preventing congestion at a remote signaling point from causing the network to bottleneck. If T6 should time out, the link is considered at fault, and level three initiates the alignment procedure on the affected link. The value for T6 is one to six seconds.

		To indicate that congestion has subsided and normal processing can begin, the affected signaling point will begin transmitting MSUs again and stop the transmission of SIBs. The receiving signaling point recognizes the absence of SIBs and begins receiving MSUs as normal.

		The status indicator "processor outage" (SIPO) indicates that the transmitting signaling point cannot communicate with levels three and four. This could be caused by a central processor unit (CPU) failure or a complete nodal failure. If maintenance personnel have manually placed a link out of service, the link will send SIPOs to the adjacent signaling point as well. Level two is usually functional, since it resides within the link interface hardware. The signaling point sends an SIPO to notify the distant signaling point to stop sending MSUs.

		Upon receipt of an SIPO, transmission of all MSUs is stopped, and FISUs are sent to the affected signaling point. The transmitting and receiving nodes stop level-two timers T5, T6, and T7. If the condition persists for too long, then the link is failed, and level three initiates the alignment procedure on the affected link.

		Level three, even though there has been a processor outage, may still be capable of controlling level-two functions. When this is the case, level three may request level two to empty its buffer (both receive buffer and transmission buffer). When this is the case, all MSUs in the buffer are discarded. When an MSU is received from a remote signaling point, level two of the affected signaling point sends an FISU with the forward sequence number (FSN) and the FIB set to the same value as the BSN and the BIB of the last MSU received from the remote signaling point. Normal processing of all messages then resumes.

		The LSSU also sends a status indicator out-of-alignment (SIO). This condition occurs when a signal unit is received that has a ones-density violation (the data field simulated a flag) or the signaling information field exceeded its maximum capacity of 272 octets. The SIO is sent when a link is failed and the alignment procedure is initiated.

		An LSSU out of service (SIOS) indicates that the sending signaling point cannot receive or transmit any MSUs for reasons other than a processor outage. Upon receipt of an SIOS, the receiving signaling point stops the transmission of MSUs and begins transmitting FISUs. The SIOS is also sent at the beginning of the alignment procedure.

		Link status of normal (SIN) or emergency (SIE) indicates that the transmitting signaling point has initiated the alignment procedure. The link is made unavailable for MSUs and only FISUs are transmitted to the affected signaling point until a proving period has been passed (see the section on proving periods further in this chapter). After successful completion of a proving period, MSUs can be transmitted over the link to the affected signaling point.

		In the event that a LSSU is received with errors, the receiving signaling point discards the signal unit. Retransmission is not requested of the LSSU. The FSN of an LSSU does not increment, but assumes the value of the last transmitted MSU. The BSN does increment when an acknowledgment is being sent to the distant signaling point.

		The LSSU is processed within level two and does not get passed to level three. However, level two may pass control information to level three, depending on the status of the LSSU. For example, if an LSSU with a "busy" status is received, level two notifies level three to stop transmission of MSUs.

		It is also important to remember that the LSSU works independently of network management, which is a level-three function. In fact, network management uses the MSU to send management information to other nodes and can use any link or route to reach adjacent nodes. Level three is used when links have failed altogether and LSSUs cannot be transmitted over the affected link.

		Most level-two problems are caused by hardware failures. Therefore, they do not require the same level of sophistication as software problems. Level-three and -four errors are normally software related, and they require more sophisticated reporting mechanisms. The recovery procedures used by maintenance personnel will also be very different for level-three and -four problems from those at level two.

		Signal Unit Alignment Procedure

		Alignment is the sending of correct signal units with correct length and without ones-density violations. When an error occurs due to either of these events, the affected signaling point will begin the alignment procedure.

		The purpose of the alignment procedure is to reestablish the timing and alignment of signal units so that the affected signaling point can determine where signaling units begin and end. As previously mentioned, the out-of-alignment condition occurs when the flag has been simulated within the data (ones-density violation) or the signaling information field is too long (longer than 272 octets), which would indicate that a flag was missed.

		The procedure resets both the transmitting and receiving nodes at level two and does not affect other links at either signaling point. The procedure also provides testing for a given period of time to ensure that the link transmission is reliable, preventing further errors from occurring.

		There are two alignment procedures used: normal alignment and emergency alignment. Normal alignment is used when there are other links associated with the affected link (such as in a linkset). The other links must be to the same destination. An emergency alignment is used when there are no other links to the adjacent signaling point within the linkset. The emergency alignment goes through the same procedure, but within a shorter time period.

 

Level three is responsible for determining which alignment procedure to use.

		There are four states entered during alignment. Timers associated with each state ensure that the signaling point does not get stuck in any one state. When any of the timers times out, the alignment starts over again. The following explanation describes each state and the events that occur during these states.

		State 00 Idle

		This state indicates that the procedure is suspended, and it is the first state entered. State 00 is resumed whenever the alignment procedure is aborted (due to excessive errors). During the time a link is in "proving," level-three network management reroutes signal units to other links. If a link should fail the proving period, level three places the link back into state 00 for a specified time period (level-three timer T17).

		To prevent rapid link oscillation between in-service states and out-of-service states, level-three timer T32 is used. When a link is placed back into the alignment procedure, timer T32 is started. If the link fails during T32, the link is placed back into state 00, until T32 expires. Attempts by the remote signaling point to begin alignment of the link are ignored during this time. LSSUs with status indication of out of service (SIOS) are sent by the signaling point until the T32 expires.

		State 01 Not Aligned

		This is entered when initiated by level three. The LSSU is used to send a status of out of alignment (SIO). When this state is entered, the level-two timer T2 is started. Timer T2 ("not aligned") for normal alignment is set to 11.5 or 23 seconds. If the signaling nodes at either end use a mechanism for the automatic allocation of signaling links, the timer T2 must be set differently at each end.

		Nodes have the option of assigning links for signaling but making them inactive (at which time, they are available for other applications, such as voice transmission). The signaling point has the capability through level three of interrupting the inactive links from their present applications and reassigning them to carry signaling traffic on an as-needed basis (such as in the event of signaling link failures within a linkset). It is with this type of application that T2 must be set differently at either end.

		State 01 remains until level three initiates the next state transition. During this state, the signaling point transmits LSSUs with a status indication of SIO.

 

State 02 Aligned

		This state indicates that the link is aligned and is capable of detecting flags and signal units without error. Remember that "out of alignment" means that the signaling point can no longer delineate signal units based upon receipt of a flag. What this really implies is that the link has lost its timing, and it no longer recognizes the beginning and end of a signal unit. This state is indicating that the link is now capable of detecting flags and recognizes the boundaries within the signal unit itself.

		During the time that the link is in state 02, level-two timer T3 is started. When the link leaves state 02 and enters state 03, T3 is stopped. If T3 should time out, the link is returned to state 00, and the process begins all over again.

		State 03 Proving

		The proving period is used to test the integrity of the link and level two at the signaling point. During the proving period, LSSUs with the value of SIN or SIE are sent, and errors are counted. There are two proving periods: normal and emergency. The proving period of normal lasts for 2.3 seconds, during which time, no more than four errors may occur while in state 03. The alignment error rate monitor keeps count of all errors received during the proving period. The AERM is an incremental counter, which counts all transmission errors, including CRC errors and ones-density violations.

		During the proving period, FISUs are sent on the link. The LSSU of normal or emergency (SIN or SIE) is also sent to indicate that the link is in proving.

		The emergency proving period lasts for a duration of 0.6 seconds, during which time, no more than one error may occur. This is also monitored by the AERM. When excessive errors have occurred according to the procedure, the link is returned to state 00, and the process begins all over again.

		If a link cannot be restored, it is returned to the idle state and the alignment procedure is repeated, until either the link is restored or maintenance personnel detect the repetitive failure and take corrective action. When a link is found to be continually failing or in constant alignment, it usually indicates an equipment failure, which can be resolved by manual intervention (replacement of the link interface card usually fixes the problem). Maintenance personnel should always be monitoring the status of signaling links and watching for links that continuously fail. Traffic measurements provided by most signaling points can be a useful tool in determining the number of failures during a given period of time.

 

Level Three Alignment Processes

		After a link has successfully passed the alignment procedure, the link is returned to an in-service state, where MSUs are transmitted, and normal processing is allowed. The link is placed into a probationary period by level three, which lasts a period of level-three timer T33. If the link fails during the probation, it is placed back into suspension (state 00), during which time, level two sends LSSUs with a SIOS. All attempts to place the link into alignment are ignored until level-three timer T34 expires.

		Timer T34, "Suspension timer for link oscillation," is also used to prevent links from rapidly oscillating from in-service to out-of-service state. This set of level-three timers, T32, T33, and T34, are all used to filter link oscillation and are only required in Bell System networks at Signal Transfer Points (STPs).

		Upon completion of the alignment procedure at level two, level three begins a signaling link test, to determine if the link is capable of carrying traffic. This is an option in the ANSI recommendations, but it is a requirement within the Bell System networks.

		The Bell System networks also use a mechanism for the automatic allocation of signaling links in the event that there is a signaling link failure within the linkset and additional links are required. This procedure uses a predefined set of links which are used for other applications (such as voice transmission) and, when needed, places them into the alignment procedure for preparation as signaling links within the SS7 network.

		This process ensures that there are always links available, even when the designated active links have failed and cannot be restored. Level three is responsible for the activation and restoration of these links, and for assigning them to the proper linkset. Link management at level three is the function that looks after this.

		In summary, MTP is a vital part of the SS7 network. This protocol provides the transport services for messages between signaling points. All messages traveling through the network rely on MTP for reaching the adjacent signaling point without error.

		The level-three network management functions of MTP ensure the integrity of the network, managing the various functions of routing and link alignment. This vital part of the network allows the SS7 entities to work around failures and congestion conditions without human intervention, and is what makes the SS7 network so robust.