<<

. 99
( 132 .)



>>



Payload: 1 - 65 528 bytes Pad SSCOP trailer




PL R PDU type N(S)

bits 2 2 4 24

Figure 7.58 SSCOP SD PDU


Table 7.20 shows the SSCOP PDU types, and a brief description of each.
Most of the applications will request signalling messages to be carried in the Sequenced
data (SD) PDU for assured data transfer. The SD appends a 4-octet trailer to the sig-
nalling message, as shown in Figure 7.58. The maximum permitted size of a message is
65 528 octets.
The ¬elds are de¬ned as follows:

• Pad: the pad ¬eld sizes the payload to a multiple of 4 octets
• PL: pad length, indicates the length of the pad (0“3)
• R: reserved for future use
• PDU type: the particular PDU type as de¬ned in Table 7.20; 1000 for SD
• N(S): sequence number of the SD PDU.
7.13 ATM SIGNALLING 481


For assured data transfer, a signalling connection must be established. To initiate this,
the transmitter sends a begin connection (BGN) message to the receiver, containing the
initial sequence number, a window size and the ¬rst signalling payload. The receiver
will respond with a begin connection acknowledge (BGAK) message, as illustrated in
Figure 7.59.
The transmitter then proceeds to send signalling messages, incrementing the sequence
number for each. Periodically the transmitter will send a POLL PDU to request the
receiver™s current state. The period is de¬ned by an internal timer, and is based around
either a time or a certain number of PDUs transmitted. The receiver will reply with a
STAT PDU containing the next sequence number expected and a list of missing sequence
numbers. The transmitter will then retransmit the missing PDUs. The STAT message also
contains the window size, allowing it to be adjusted during a connection. This is referred
to as a selective ARQ mechanism. An example of communication where PDUs are not
received is illustrated in Figure 7.60. Here, the receiver has not received SD PDUs 3 and
5. At sequence 6, the transmitter sends a POLL PDU, the STAT reply to which indicates
this to the sender. Once received, the sender retransmits 3 and 5.
There can be a problem with this mechanism when the time between POLLs is too
large. Here the system will have a slow recovery time when PDUs are lost. Therefore,

RNC BTS

BGN

BGAK




Figure 7.59 SSCOP signalling connection establishment


RNC BTS
Seq # Seq #
1 SD
2 1
SD
3 2
SD
4 ?
SD
5 4
SD
6 ?
POLL
7 6
SD
))
N:6 (3,5
STAT(S SD
8 7
3 8
SD
5 3
SD
5


Figure 7.60 SD PDU transfer with PDU loss
482 UMTS TRANSMISSION NETWORKS


RNC -> CN
AAL 5
CPCS-PDU: 00 00 89 E7 0A 00 01 15
POLL PDU POLL sequence no.
N(PS) : 35303 (0089E7h)
N(S) : 277 (000115h) Next seq. no. to send


CN -> RNC
AAL 5
CPCS-PDU: 00 00 89 E7 00 00
04 F3 0B 00 01 15
STAT PDU
Echoed POLL seq. no.
N(PS) : 35303 (0089E7h)
N(MR) : 1267 (0004F3h)
N(R) : 277 (000115h) Next seq. no. expected


Figure 7.61 Iu interface POLL-STAT keep-alive procedure


RNC CN

END


ENDAK




Figure 7.62 SSCOP signalling connection release


an additional PDU is de¬ned, the unsolicited STAT (USTAT). This allows the receiver to
send a status PDU (USTAT) once it detects a loss, even if it has not received a POLL
PDU. The POLL and STAT function is also used as a keep-alive procedure when there is
no network activity, with a single POLL sent and a STAT received. Figure 7.61, shows
an example of this on the Iu interface.
Upon completion of the signalling connection, the transmitter will send an END PDU,
which is acknowledged by the receiver using an ENDAK PDU (Figure 7.62).



7.13.3 Service-speci¬c coordination function (SSCF)
The role of the SSCF is to interface the protocols accessing the SAAL to the SSCOP
below; for example, mapping of Q.2931 signalling messages. The SSCF provides the
following services to the upper layers:

• Unacknowledged data transfer: no guarantee of data integrity or delivery.
• Assured data transfer: guarantee of in-sequence delivery and loss/corruption protection.
7.13 ATM SIGNALLING 483


Table 7.21 Reserved channels
for signalling
Signal VPI VCI
Meta 0 1
Broadcast 0 2
Point-to-point 0 5
ILMI 0 16
PNNI 0 18


• Transparent transfer of data: the contents of a payload is not restricted and is carried
transparently.
• Establishment and release of signalling connections: for assured data transfer.

The SSCF can accept SDU sizes of up to a maximum of 4096 octets.
The ITU-T provides two recommendations for the SSCF: one for SSCF at the UNI
(Q.2130) and one for SSCF at the NNI (Q.2140). In UMTS, the Node B or BTS is
considered to be the host and therefore the Iub interface between the Node B and the
¬rst switch, the RNC, is a UNI interface. Therefore, SSCF-UNI is used on the Iub for
both NBAP and AAL2 signalling (Q.2630.1). The other ATM interfaces, Iu and Iur, are
considered NNI and use SSCF-NNI for RANAP, RNSAP and AAL2 signalling.
ATM signalling is sent over reserved channels. Simple point-to-point signalling uses
VPI = 0, VCI = 5. The remaining reserved channels are outlined in Table 7.21.
Meta-signalling is used to establish signalling channels. It sends a one-cell signal, which
can set up three different types of channel:

• Point-to-point
• General broadcast
• Selective broadcast.

It contains procedures to set up new channels, verify channels and release existing
channels. For example, a packet sent on the meta-signalling channel, VPI 0/VCI 1, could
indicate that point-to-point signalling will use VPI 0/VCI 25 instead of the default.
ILMI is the integrated local management interface, used for network management, and
PNNI is the private network node interface, responsible for dynamic routing.



7.13.4 ATM addressing format
Within ATM, there are three different address formats currently speci¬ed. The formats
are all based on the network service access point (NSAP) scheme outlined by the ISO.
An example of an ATM address is:

47.246F.00.0E7C9C.0312.0001.000014362758.00
484 UMTS TRANSMISSION NETWORKS


Network no. Host no.

AFI IDI HO DSP ESI SEL

Initial domain ID Domain specific part (DSP)

Figure 7.63 NSAP address format


Table 7.22 ATM addressing standards
AFI Issuing authority Format
39 ISO Data country code (DCC)
45 E.164, B-ISDN Public addresses
47 British Standards Institute International code designator (ICD)
49 N/a “ private addressing Implementation speci¬c


The address format of a NSAP is shown in Figure 7.63.
The initial domain part is globally unique, with the domain-speci¬c part assigned
locally. The AFI ¬eld is the authority and format indicator. It is a unique number allo-
cated to each addressing authority. IDI, the initial domain identi¬er, is a unique identi¬er
assigned by the addressing authority. HO-DSP is the high-order bits of the domain-speci¬c
part. It is similar to the network number in an IP address and can be used for subnetting.
ESI, the end system identi¬er, speci¬es the end system, and corresponds to a host number
in IP. SEL, the selector, selects a protocol within an end system. The three address formats
used are allocated three unique AFIs for the organizations that de¬ne the address schemes,
as shown in Table 7.22. Like IP, if the ATM addressing is completely private, then an
organization need not apply for an address. This will be the case within the network for
most mobile operators implementing an ATM transport layer for UMTS. However, the
core network element, which bridges the divide between the ATM network of the RAN
and the ISDN network of the circuit core, will most likely be given an E.164 format for
compatibility with the ISDN network.
The address format for each is 20 bytes long, as shown in Figure 7.64. Note that the
format for AFI = 49 is decided by the organization using this as an internal, private
addressing structure.

• DFI: domain and format identi¬er
• AA: administrative authority
• RD: routing domain
• AREA: area identi¬er.

The ESI is typically derived from the 48-bit MAC address of the destination system.
In the E.164 format, an extension is made to the standard N-ISDN E.164 number, which
is up to 15 digits. Here, the 15-digit binary coded decimal (BCD) number is padded out
with 0xFs to 8 bytes long.
7.13 ATM SIGNALLING 485


Network supplied End system supplied

DCC ATM Address Format
AFI DCC DFI AA Resvd. RD AREA ESI SEL


Initial domain ID Domain specific part (DSP)
ICD ATM Address Format
AFI ICD DFI AA Resvd. RD AREA ESI SEL

Initial domain ID Domain specific part (DSP)
E.164 ATM Address Format
AFI E.164 RD AREA ESI SEL

Initial domain ID Domain specific part (DSP)
20 bytes

Figure 7.64 ATM address formats

Within UMTS, the ATM address formats E.164 and ICD are speci¬ed for use at both
the ATM and AAL2 layers.
The ATM Forum has also de¬ned an address naming system with resolution, similar
to the domain name system (DNS) used with IP. The system is called the ATM Name
Service (ANS). ATM automatic address registration is provided by the ILMI network
management protocol.
An organization is free to implement its own addressing scheme either using format
0—49 or by applying for its own ICD. However, since ATM is generally only being
used as a backbone, most commercial ATM equipment can automatically con¬gure ATM
addresses to interfaces connected to the switch. A typical example of how this is performed
is Cisco ATM auto address con¬guration. Cisco has been assigned the ICD of 0x00.91.

<<

. 99
( 132 .)



>>