UMUX from KEYMILE, covers all your communication

requirements in one system.

UMUX
Technical Description System

UMUX ATM Traffic System

(R8)

EN/LZTBU 220 102/2 RA

UMUX ATM Traffic System
Platform Release R8

Copyright and Confidentiality: Copyright in this document vests in KEYMILE AG (KEYMILE). This
document contains confidential information which is the property of
KEYMILE. It must be held in confidence by the recipient and may not
be used for any purposes except those specifically authorised by
contract or otherwise in writing by KEYMILE. This document may not
be copied in whole or in part, or any of its contents disclosed by the
recipient to any third party, without the prior written agreement of
KEYMILE.

Disclaimer: KEYMILE has taken reasonable care in compiling this document,

however KEYMILE accepts no liability whatsoever for any error or
omission in the information contained herein and gives no other
warranty or undertaking as to its accuracy.

KEYMILE reserves the right to amend this document at any time

without prior notice.

Document number: EN/LZTBU 220 102/2 RA

KEYMILE AG
Schwarzenburgstrasse 73
CH-3097 Bern-Liebefeld
Switzerland © August 2007 by KEYMILE AG
 © KEYMILE AG

Table of contents i
Summary 1- 1
General 1- 1
Services 1- 2
Circuit Emulation Service: CES 2048 kbit/s (AAL1) 1- 2
Loop Emulation Service: LES (AAL2) 1- 2
ATM UNI 1- 3
Physical layer 1- 3
ATM layer 1- 3
OAM 1- 4
Protection 1- 4
MSP / EQP 1- 4
EQP N:1 1- 4
Network synchronisation 1- 4

Architecture and implementation 2- 1

Platform 2- 1
Functional Units 2- 3
ABUS 2- 4
Topology 2- 4
ABUS characteristics 2- 4
ABUS structure 2- 4
UMUX 1500 2- 6
UMUX 1200 2- 7
UMUX 900 2- 8
ABUS access 2- 8
Traffic control 2- 9
Usage parameter control (UPC) 2- 9
Shaping 2- 10
Queuing and scheduling structures 2- 10
Ingress 2- 12
Egress multi-port 2- 13
Egress single port 2- 13
Cross connection 2- 13

EN/LZTBU 220 102/2 RA Technical Description System iii

 Table of contents © KEYMILE AG

Traffic management 3- 1
Supported service categories and conformance definitions 3- 1
Overview 3- 1
Quality of service 3- 1
Usage Parameter Control (UPC) 3- 2
Introduction 3- 2
Shaping 3- 3
Introduction 3- 3
Characteristics 3- 3
CAC 3- 4
Introduction 3- 4
Effective booked bandwidth 3- 4
Introduction 3- 4
Ingress traffic 3- 4
Egress traffic 3- 4
Available bandwidth as a resource 3- 4
Egress ports 3- 4
Over-subscription 3- 4
CAC on VP with VCCs aggregating 3- 5
Management resources 3- 5

OAM 4- 1
ATM OAM flows 4- 1
Cross connections and endpoints 4- 1
Fault management functions 4- 1
Loopback 4- 3
Cell insertion/extraction 4- 3
Loop back technique 4- 4
ATM performance monitoring 4- 4
Congestion reporting 4- 4
Statistics 4- 4

Auxiliary functions 5- 1
Network synchronisation 5- 1
General 5- 1
Synchronisation network 5- 2
Equipment timing source 5- 3
ATM network elements 5- 4
IAD 5- 4
Voice Gateway 5- 4
DSLAM 5- 4
Unit timing configurations 5- 4
Existing TDM units 5- 4
ATM units 5- 4

iv Technical Description System EN/LZTBU 220 102/2 RA

 Table of contents © KEYMILE AG

Management communications 5- 5
ECC over ATM 5- 5
ATM ECC UMUX functional blocks 5- 6
Protection 5- 7

EN/LZTBU 220 102/2 RA Technical Description System v

 © KEYMILE AG

Precautions and safety

For generic information on precautions and safety refer to [033].

Referenced documents

[033] Precautions and safety

[202] UMUX 1500 Technical Description (R8)

[203] UMUX 1200 Technical Description (R8)
[204] UMUX 900 Technical Description (R8)

[205] UMUX TDM Traffic System (R8)

[208] UMUX Operation & Maintenance (R8)
[209] UMUX Management (R8)
[210] UMUX Units (R8)
[902] UMUX Network Functions (R8)

[048] Release Note UMUX / UCST R8A

[822] UMUX ATM traffic delay (R8)

EN/LZTBU 220 102/2 RA Technical Description System vi

 © KEYMILE AG

Summary 1
General
The UMUX constitutes a modular system adaptable to combined PDH, SDH
and ATM network solutions. This document addresses the specific system
aspects of ATM.
It includes the characteristics and the measures being taken on network
element -, unit - and element manager level to accommodate ATM based
services.
A set of ATM units is available constituting services at physical, ATM and
ATM adaptation layer. All these units have at least access to the internal
ATM switch consisting of a switch module on each unit and the ABUS which
enables the interconnection between them.
The main focus of the first release is on units providing ATM UNIs at various
physical interfaces and adaptation functions to support voice and circuit
emulation services (AAL1, AAL2).

EN/LZTBU 220 102/2 RA Technical Description System 1-1

 Summary © KEYMILE AG

Services
Circuit Emulation Service: ATM Forum AF-VTOA-0078.000, Circuit Emulation Service Interoperability
CES 2048 kbit/s (AAL1) Unit: ACONV
Clock recovery
• Adaptive or SRTS in conjunction with transparent or monitored P12 cross
connection modes (PBUS).
• Synchronous, SRTS or adaptive in conjunction with terminated or clock
master P12 cross connection modes (PBUS)
Capacity

• Unit: Up to 16 P12 interfaces

• TDM cross connection: 64 x P12, corresponding to 4 ACONV units

Loop Emulation Service: ATM Forum AF-VMOA-0145.000 (07/2000), Loop Emulation Service Using
LES (AAL2) AAL2
Unit: VGATE
V5 services

• Protocol resolution: Up to 4 V5.2 instances

• User ports: ATM (LES) via IAD
TDM
• Services PSTN, ISDN-BA
Capacity

• User ports
− PSTN: 2500
− ISDN-BA 1250
• Unit: Up to 16 x P12 PBUS access
(V5-links and TDM user ports)
• IAD: Up to 1024
• ATM cross connections: Up to 1024 VCCs
• TDM cross connections:
− User Ports: Up to 240 P0 (PSTN)
Up to 112 P0_nc (ISDN-BA)
− V5.links: Up to 16 P12

1-2 Technical Description System EN/LZTBU 220 102/2 RA

 Summary © KEYMILE AG

ATM UNI
Physical layer ETSI EN 301 163-2-1, Generic requirements of Asynchronous Transfer
Mode transport
ATM Forum AF-PHY-0064.000, E1 Physical Interface Specification,
Version 1.1
ATM Forum AF-PHY-0086.001, Inverse Multiplexing for ATM Specification,
Version 1.1
Unit: ATIOP
• STM1 / VC-4 Bandwidth: 149'760 kbit/s (353'207 cells/s)
Unit: ACONV
• P0-nc Minimum usable bandwidth: 1859 kbit/s
(4384 cells/s)
Number of ports: Up to 16

• IMA Bandwidth: Groups of n x 1920 kbit/s

(4528 cells/s), 2 ≤ n ≤ 16
Number of IMA Groups: 1 to 14 per unit
Max number of 2Mbit/s links per
- unit: Up to 16
- group: Up to 16
"Up to" means that the number of (P0_nc) ports and (IMA) links are
cumulative per unit up to the given number.
Unit: ADACA, ADACB
• ADSL over POTs /ISDN
Bandwidth upstream 32kbit/s, ..., 1024 kbit/s
Bandwidth downstream 32kbit/s, ..., 8064 kbit/s
Number of ports: 16

ATM layer ATM Forum AF-TM-0121.000, Traffic Management Specification,

Version 4.1
• Permanent virtual connections (PVCs) only, up to 2048 connections per
network element
• Point to point, bi-directional, asymmetric VCCs and VPCs
• UNI cell format
• Service categories supported: CBR, VBR-rt, VBR-nrt and UBR
• Conformance definitions: CBR.1, VBR.1 and UBR.1
• Ingress traffic control functions: Usage parameter control (UPC),
Connection admission control
(CAC)
• Egress traffic control functions: VP shaping (ATIOP), CAC
ATM cross connect
• Switching capacity 515'160 kbit/s (1'215'000 cells/s)
• Up to 2048 connections per network element

EN/LZTBU 220 102/2 RA Technical Description System 1-3

 Summary © KEYMILE AG

• Accessible by up to 8 units (slots), depending on type of subrack

• Although the cross connect fabric allows connections to be set up from
any port to any port or service, the UMUX has still to be considered as an
ATM multiplexer or concentrator with a maximum aggregate bandwidth
according to STM1 towards the regional ATM network.

OAM
ITU-T I.610, B-ISDN operation and maintenance principles and functions
Fault management is provided using AIS, RDI, CC and LB cells:
• AIS, RDI: F4, F5 end-to end
• CC: F4, F5 segment and end-to-end
• LB (tba): F4, F5 end-to end
F4, F5 segment any CP (LLID = CPID or all 0s))
Ingress congested seconds monitoring:
• Per unit service category buffer limit
Egress congested monitoring:
• Per unit service category and port buffer limit
• Per STM1 unit shaped VPs
Cell statistics are performed on a per connection point level.

Protection
MSP / EQP ATIOP: 1+1, non revertive: Bidirectional, fixed bridge (no extra traffic),
compatible with 1:N switching (ITU-T G.783 A.3.4.1). Interworking with
bidirectional, optimised protocol (ITU-T G.783 B.1) is possible.
Equipment failure will cause SETS/bus control (and traffic if MSP enabled)
switched to the non-failed unit.

EQP N:1 A protection group consists of N working units and one protection unit of the
same type (ACONV or VGATE).
Number of independent protection groups: up to 4
Number of working units within a protection groups: up to 4
Switching time: about 1 min

Network synchronisation
Synchronisation is a mainly physical layer aspect. Cell derived timing
information should not be used as a reference for network synchronisation.

1-4 Technical Description System EN/LZTBU 220 102/2 RA

 © KEYMILE AG

Architecture and implementation 2

Platform
The ATM function is implemented on the existing UMUX platform. Apart from
a new set of ATM units, the cell-structured ABUS has been introduced on
NE level. This bus uses the same physical bus (backplane) as the SBUS.
Some new UCST NE functions have been added to cover the specific needs
of ATM layer functions like VPC / VCC termination, VPC / VCC matrix
connection, OAM and CAC.
In general a NE consists of a mix of ATM units and TDM units, the latter
ones constituting the physical ATM layer and/or the user plane.
The example given in Fig. 2-1, should illustrate the combined use of all the
traffic buses (ABUS, SBUS, PBUS). Apart from being connected to the
ABUS, the units ATIOP and ADAC<X> feature an external ATM UNI,
whereas the units ACONV and VGATE without external front interface are
placed between ABUS and PBUS.

EN/LZTBU 220 102/2 RA Technical Description System 2-1

 Architecture and implementation © KEYMILE AG

Figure 2-1: UMUX with the traffic buses ABUS, SBUS and PBUS

xDSL STM1 STM1

ATM UNI ATM UNI SDH
(IMA over 16 x
VC-12)

ADACx
SYNIF
16x ATIOP SYNIC
xDSL SYNIO
ABUS SBUS
(VPC, VCC) (VC-12, VC-3)

AAL2
VGATE ACONV ACONV SYNAC SYNAC
CPE User
Port Access 16x CES IMA 16 x 8x P12/ 8x P12/
TDM User V5 link 2Mbit/s 2Mbit/s VC-12 VC-12
Port Access Access PBUS
(P0, P0-nc, P12)

PHLC LOMIF LOMIF

a/b 8 x E12 8 x E12

60 POTS

Note: Specific ATM units are greyed out. Arrows indicate a possible signal
flow (the drawing shows not all UMUX units but typical units for the
relevant application).
The PBUS access gives more flexibility in choosing by configuration the
external physical interface or in rearranging the aggregated traffic:
• In the case of 2Mbit/s CES (circuit emulation service), the P12 (2Mbit/s
path) signal to be sent over ATM may be composed from different
sources by switching on PBUS P0 (64kbit/s timeslot) level.
• A VC-12 over STM1 can be taken as the physical link layer for an IMA
group.
A short characterisation of the ATM units can be found in the next chapter

2-2 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

Functional Units
(More details are to be found in the corresponding unit descriptions)
ATIOP
• ABUS-STM1 unit, single port
• ATM UNI at STM1 /VC-4 physical bandwidth
• ECC over ATM
ACONV
• ABUS-PBUS unit. PBUS access by 16 x 2 Mbit/s
• IMA, up to 16 links of 2Mbit/s divided into one group of 16 links up to 8
groups with two links each.
• Up to 16x CES 2048 kbit/s (AAL1)
• ECC over ATM
VGATE
• ABUS-PBUS unit. PBUS access by 16 x 2 Mbit/s
• VoDSL in AAL2/LES format
• up to 4 V5.2 interfaces with up to 16 V5.2 links
ADACx
• ADSL-UNI with 16 ADSL ports
• ADACA:
• ADSL over Pots
• Bandwidth upstream: 32kbit/s, ..., 1024 kbit/s
• Bandwidth downstream: 32kbit/s, ..., 8064 kbit/s
• ADACB:
• ADSL over ISDN
• Bandwidth upstream: 32kbit/s, ..., 1024 kbit/s
• Bandwidth downstream: 32kbit/s, ..., 8064 kbit/s

EN/LZTBU 220 102/2 RA Technical Description System 2-3

 Architecture and implementation © KEYMILE AG

ABUS
Topology
ABUS characteristics The ABUS transports ATM cells between the ABUS units of the UMUX. The
characteristics of the ABUS are
• Traffic signal formats
The traffic format on the ABUS is ATM cells. The transport capacity of the
ABUS is 1215000 cells/s (515160 kbit/s).
• Cross connection
The ABUS traffic units evaluate the address information in the cell
headers and use lookup tables for the cell routing between the ABUS
units/interfaces in the UMUX.
The UMUX can handle up to 2048 VCC/VPC connections per NE.
• Physical bus structure and access
The ABUS consists of a physical bus which transports a full ATM cell in
one bus cycle.
The ABUS has linear bus access. Any ABUS unit can read and write
cells from/to the ABUS. One of the ABUS units is the bus master which
grants the units temporary write permission for the ABUS.
• Timing
The ABUS timing is provided by a selected ABUS unit. Depending on the
ATM unit, this clock is synchronised to SETS or PETS. If both
ABUS/SBUS sectors provide SETS, the 2 sectors constitute one logical
SETS system (SETS is an NE level function).

ABUS structure The ABUS and SBUS are physically identical. The type of the first configured
unit (ABUS, SBUS unit) on the bus defines the bus use and traffic type
(SBUS or ABUS). The UCST refuses the mixture of ATM and SDH units
within the same sector.
The ABUS transfers cells from one connection point to another connection
point at a speed of 1215 kcells/s which corresponds to a bandwidth of
515160 kbit/s (for comparison: to convey ATM traffic between two STM-1
ports will cost a bus bandwidth of up to 300 Mbit/s approximately).
In contrast to the SBUS, the cell based ABUS makes no difference between
tributary and aggregate signals nor between receive and transmit direction.
The ABUS is terminated on ATM units by means of a special bus access
module. The ABUS access of the different UMUX subracks is shown in
Fig 2-2.

Figure 2-2: ABUS access of UMUX subracks and slots

NE Slots
UMUX 1500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
UMUX 1200 - - - - 5 6 7 8 9 - 11 12 - - - - - - - - 21
UMUX 900 - - - - - 6 7 8 9 - 11 - - - - - - - - - 21
Sectors Range Sector A Range Sector B

Note: Greyed out slots feature SBUS/ ABUS access. Contiguous ABUS
slots constitute an ABUS sector.

2-4 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

The ABUS/SBUS structure supports Multiplex Section Protection and

SETS Protection. The implementation of protection requires a special
relationship between the slots assigned to the working and the protecting
unit.
Slots which support the above protection types use control signals between
the units to allow monitoring of the protected functions on both units and to
enable the protection switch in case of a failure. The corresponding control
signals are hardwired within the ABUS/SBUS between 2 paired slots as
shown in the figures below.
Please note that the shaded grey structures and interconnections
in the figures below show logical connections only and represent
the complements to the UMUX 1500 ABUS implementation.

EN/LZTBU 220 102/2 RA Technical Description System 2-5

 Architecture and implementation © KEYMILE AG

UMUX 1500 The UMUX 1500 provides 2 equivalent sectors, each with its own
ABUS/SBUS:

• Sector A: Slots 2 … 9
• Sector B: Slots 15 … 20
Each of the sectors can be used independently but exclusively for ATM
(ABUS) or SDH (SBUS) traffic. A mixture of ATM and SDH traffic in the
sector is not possible!

Figure 2-3: Structure of the ABUS vs. slots of the UMUX 1500 subrack

SECTOR A SECTOR B

ABUS / SBUS Set A ABUS / SBUS Set B

Data Bus Data Bus

ATM cells or ATM cells or

STM-1 frames STM-1 frames

Interconnection of slots for protection

W2 W3 P2 P3 W6 W7 P6 P7 W15 W16 P15 P16 W19 P19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

W<p> Working <p>

You must first implement units for MS and/or SETS protection in slots
with the attribute W<p>.
P<p> Protecting <p>
You must implement units providing protection in slots with the
attribute P<p>. Protection works between the working and the
protecting unit implemented in slots with the same <p> value.

2-6 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

UMUX 1200 The UMUX 1200 supports only slots in the sector A and provides the ABUS
on the slots 5 … 9.

Figure 2-4: Structure of the ABUS vs. slots of the UMUX 1200 subrack

SBUS Set A SBUS Set B

Drop Bus A Drop Bus A

Add Bus A Add Bus A

Drop Bus B Drop Bus B

Add Bus B Add Bus B

Interconnection of
slots for protection
W2 W3 P2 P3 W6 W7 P6 P7 W15 W16 P15 P16 W19 P19

5 6 7 8 9 11 12 21

W<p> Working <p>

You must first implement units for MS and/or SETS protection in slots
with the attribute W<p>.
P<p> Protecting <p>
You must implement units providing protection in slots with the
attribute P<p>. Protection works between the working and the
protecting unit implemented in slots with the same <p> value.

EN/LZTBU 220 102/2 RA Technical Description System 2-7

 Architecture and implementation © KEYMILE AG

UMUX 900 The UMUX 900 supports only slots in the sector A and provides the ABUS
on the slots 6 … 9.

Figure 2-5: Structure of the ABUS vs. slots of the UMUX 900 subrack

SBUS Set A SBUS Set B

Drop Bus A Drop Bus A

Add Bus A Add Bus A

Drop Bus B Drop Bus B

Add Bus B Add Bus B

Interconnection of
slots for protection
W2 W3 P2 P3 W6 W7 P6 P7 W15 W16 P15 P16 W19 P19

6 7 8 9 11 21

W Working
You must implement units providing interfaces for protection in slots
with the attribute W<p>.
P Protecting
You must implement units providing protection in slots with the
attribute P<p>. Protection works between the working and the
protecting unit implemented in slots with the same <p> value.

ABUS access When implementing units with ABUS access you must consider:
• Implementation of units with STM-1 UNI
To support MSP and SETS equipment protection (ABUS units with STM-
1 aggregate interfaces) the ABUS structure imposes defined unit to slot
allocations:
− ATIOP:

ATIOP Slots Sector A Slots Sector B

2 3 4 5 6 7 8 9 15 16 17 18 19 20
Working unit A B C D E F G
Protecting unit A' B' C' D' E' F' G'
Slots available:
UMUX 1500
UMUX 1200
UMUX 900

It is mandatory to start the ATIOP implementation with a working unit

(slots 2, 3, 6, 7 and 15, 16, 19)!
Only the first ATIOP (working/Protecting) pair can provide MSP and
SETS in the sector. However, all the ATIOP pairs can provide MSP:

2-8 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

You can simultaneously operate the 2 ABUS/SBUS sectors (sector A and

B) in the UMUX 1500 subrack at a time. The first unit in the sector
defines the traffic type (ATM or SDH).
• Implementation of units without STM-1 UNI
ABUS units without STM-1 UNI do not depend on the internal ABUS
structure.
You can configure the ACONV, VGATE and ADAC<X> units to any slot
within the ABUS sector(s). These units are independent from SETS.
• Capacity of the ABUS sectors
The UMUX 1500 ABUS Sectors A (8 slots) and B (6 slots) do not provide
equal slot capacity. The use of the slots depends on the application.
• Control units
The COBUL and COBUQ control units are not released for ATM
functions!
It is not possible to support ATM functions in UMUX with the COBUQ and
COBUL.
The unit with ABUS access is responsible for the ATM layer processing
including functions such as:
• Header translation
• Queuing and scheduling
• UPC
• Shaping
• Cell statistics
• OAM
These functions can be found on each ATM unit with the exception of the
traffic control functions UPC and shaping (refer to Fig. 2-6).

Traffic control
Usage parameter control UPC is located on all UMUX ATM units ingress direction providing ATM UNI
(UPC) (see Fig. 2-6). The conformance test uses dual leaky bucket implementation
based on the Generic Cell Rate Algorithm (CGRA) as defined in ATM Forum
AF-TM-0121.000 (March 99), Traffic Management Specification,
Version 4.1.
UPC is always applied on ABUS connection level.

EN/LZTBU 220 102/2 RA Technical Description System 2-9

 Architecture and implementation © KEYMILE AG

Figure 2-6: ATM Cross Connect and Traffic Control Function

"Ingress" denotes the traffic direction from port to ABUS

"Egress " denotes the traffic direction from ABUS to port

Shaping Rate shaping is available on 32 bulk CBR bulk queues in conjunction with
high capacity single port units (more details see section «Shaping»
on page 3-2).

Queuing and scheduling The Fig. 2-8 and Fig. 2-9 depict the ABUS cell switching architecture part of
structures each unit. This part covers all the ATM related processing between the
ABUS and the unit internal UTOPIA interface (ATM Forum Utopia Level 2,
Version 1.0). In this context there are two types of units depending on the
number of ports (PHY) and its capacity.

Figure 2-7: Port type and units

Type Figure Units

Multiport (up to 16) Fig. 2-8 ADAC
ACONV
VGATE 1)
Single port, high capacity Fig. 2-9 ATIOP

1)
VGATE is a one port, but low capacity unit

2-10 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

Figure 2-8: Multiport Units:

ATM queuing and scheduling structure

ABUS
single bulk queue for
each service category
Ingress CBR
UTOPIA Ingress
VBR-rt ABUS
1 IP
PHY VBR-nrt

UBR
Header lookup
2 UPC
PHY
Header translation
statistics queue for each port
OAM / FM (UTOPIA address) and
3 service category for up to
PHY
P12 16 ports Egress ABUS
1
CBR

n VBR-rt
PHY
VBR-nrt

UBR
n
Egress CBR
16 UTOPIA
PHY VBR-rt
OP
VBR-nrt

(DSL, P12, P0-nc, etc) UBR

Header lookup 16
Header translation CBR
statistics
OAM / FM VBR-rt

Legend: VBR-nrt

UBR

queues in bold lines

require high buffer capacity

EN/LZTBU 220 102/2 RA Technical Description System 2-11

 Architecture and implementation © KEYMILE AG

Figure 2-9: Single port Unit:

ATM queuing and scheduling structure

ABUS
CBR
Ingress
UTOPIA VBR-rt Ingress ABUS
IP
VBR-nrt
Header lookup
UBR
Header
translation
single bulk queue for
statistics
each service category
OAM / FM
(same as line side)
UPC
Per service category bulk queue
Egress ABUS
Egress
UTOPIA 1 CBR

PHY
user Bulk queue with shaping 1
traffic

(STM-1)
Round Bulk queue with shaping 2
Robin
2

Bulk queue with shaping 32

OP

Header lookup
Header Per service category bulk queue
translation 3 VBR-rt
statistics
OAM / FM
Per service category bulk queue

Legend: 4 VBR-nrt

Per service category bulk queue

queues in bold lines

5 UBR
require high buffer capacity

Ingress The ingress cell processing (see Fig. 2-8 and Fig. 2-9) is actually the same
for single and multi-port units.
Independent of the number of ports there are just four input queues, one for
each service category which are presented to the ABUS as follows:

Figure 2-10: Priorities of the service categories

Priority Service category

1 CBR
2 VBR-rt
3 VBR-nrt
4 UBR

The connection requests are governed in a round robin fashion by a single

ATM unit automatically elected by the common control unit COBUX.

2-12 Technical Description System EN/LZTBU 220 102/2 RA

 Architecture and implementation © KEYMILE AG

Candidates are ATIOP in the first place followed by ACONV. The same unit
is also responsible for the ABUS timing.

Egress multi-port Refer to Fig. 2-8.

The cells arriving from the ABUS are waiting for getting forwarded to the port
in four per port service category bulk queues. The scheduling within the
traffic for the same port is priority based (see Fig. 2-10).
Buffer allocation occurs dynamically but there are fixed buffer limits set on
service category and on queue level. On queue level buffer over allocation is
applied. That means that the total of the capacity assigned to the queues of
the same service category exceeds the limit set for that service category.

Egress single port Refer to Fig 2-9.

The egress part of a high capacity, single port unit such as the ATIOP
contains per service category 4 bulk queues and 32 queues with shaping.
As far as priorities are concerned, the queues with shaping are placed
between CBR and VBR-rt. From a traffic point of view however, the shaped
traffic is considered to be of CBR type.
The shaped queues are primarily used to terminate a VP with VCCs of any
origin. That means that the internal switching is on VCC level whereas the
aggregated traffic will be presented as a CBR VP towards the network.
The buffer space booked by a shaped queue depends on the service
category of the VCCs being aggregated.

Cross connection Cross connection may be established on Virtual Path (VP) or on Virtual
Channel (VC) level. As the name implies the cross connection fabric allows
you to establish any connection between connection points. However the
UMUX is optimised as a multiplexer aggregating / concentrating towards the
ATM network. In order to improve use of uplink bandwidth two UMUXs may
be cascaded via an STM1 or IMA interconnection.
The connections are bi-directional featuring the same VPI /VCI and service
category in each direction. Parameters of the traffic descriptor may have
direction-specific values.
Connections between two service access points (CES, LES) are not
possible.

EN/LZTBU 220 102/2 RA Technical Description System 2-13

 © KEYMILE AG

Traffic management 3
Supported service categories and conformance definitions
Overview The UMUX supports the service categories and corresponding conformance
definitions as shown in the list below.

Figure 3-1: Supported service categories and conformance definitions

Service category Traffic descriptor parameters Conformance definition

CBR PCR, CDVT CBR.1
VBR-rt PCR, CDVT, SCR, MBS VBR.1
VBR-nrt PCR, CDVT, SCR, MBS VBR.1
UBR PCR, CDVT UBR.1

The conformance definitions are specified as follows:

• Conformance definition according to
ATM Forum AF-TM-0121.000 (March 99), Traffic Management
Specification, Version 4.1, Table 4-1.
• Conformance definition index
<Category>.1 means that the UPC is CLP transparent and no tagging is
applied (tagging: changing CLP = 0 to CLP = 1).
Non-conforming traffic will always be discarded.

Quality of service Figure 3-2: QoS classes of the different service categories

Service category QoS CTD CDV CLR0+1 CLR0

class
CBR (CBR.1) Class 1 1x10-8 none
300 µs 500 µs -10
VBR-rt (VBR.1) Class 1 3x10 none
-7
VBR-nrt (VBR.1) Class 2 Unspecified Unspecified 4x10 none
UBR (UBR.1) Class 4 Unspecified Unspecified Unspecified Unspecified

EN/LZTBU 220 102/2 RA Technical Description System 3-1

 Traffic management © KEYMILE AG

Usage Parameter Control (UPC)

Introduction UPC is defined as the set of actions taken by the network to monitor traffic
and enforce traffic contract. Its main purpose is to protect network resources
from malicious as well as unintentional misbehaviour, which can affect the
QoS of other established connections. This protection is achieved by
detecting violations of negotiated parameters and taking appropriate actions.
Connection monitoring at a UNI is referred to as UPC.
The connection traffic descriptor consists of all parameters used to
characterise unambiguously the conforming cells of the ATM connection.
The UMUX 1500 supports UPC for the service categories and conformance
definition as specified in section «Supported service categories and
conformance definitions» on page 3-1.

3-2 Technical Description System EN/LZTBU 220 102/2 RA

 Traffic management © KEYMILE AG

Shaping
Introduction With UMUX the shaping is used to enforce source traffic descriptor on
terminated, aggregating VPs. Aggregating means that a number of VC
connections are assigned to the same VP towards the network with a CBR
type traffic contract on VP level. (See also section «Egress single port» on
page 2-13). This eases network traffic management especially if ATM
signalling is not available. It may also help to characterise the traffic more
efficiently especially if the VC connections are not active all the times.

Characteristics The UMUX shaping function is characterised as follows:

• Up to 32 shaped VPs on STM1 ATM UNI (ATIOP)
• VP characteristics are
− CBR traffic type
− PCR value (selectable in increments of 4000 cells/s)
− CDVmax (fixed set to 500 µs)

• Aggregated VCCs
All the VCCs allocated to a specific VP shall be of the same service
category (CBR, VBR-rt, VBR-nrt or UBR).
• Allocated buffer space (egress traffic only)
The maximum available buffer depends on the buffer limit of the service
category of the aggregated VCC or on the maximum queue length of the
shaped VP (whichever is lower). The maximum queue length of a shaped
(aggregating) VP is fixed at approximately 14000 cells (not configurable).
The service category limits for the STM-1 ATM UNI traffic are
approximately
− CBR / VBR-rt: 500 cells
− VBR-nrt: : 14000 cells
Shaping uses a Service Interval (SI) of 250 µs for cell processing. The
slowest rate is thus 1 cell per SI which is equivalent to 4000 cells/s or 1.696
Mbit/s. With this slowest bit rate, the maximum theoretical delays for cell
traffic can be calculated as follows:
• For CBR and VBR-rt traffic the above service category limits could
theoretically create cell delays of up to
125 ms (= (500 * 53 * 8 bits) / 1.696 Mbit/s)
• For VBR-nrt traffic the bulk queue could theoretically create cell delays of
up to
3.5 s (= (14000 * 53 * 8 bits) / 1.696 Mbit/s)
However, the CAC function inhibits (normally) such excessive delays for the
above service categories.

EN/LZTBU 220 102/2 RA Technical Description System 3-3

 Traffic management © KEYMILE AG

CAC
Introduction CAC is defined as the set of actions taken by the network at establishment of
connections. Based on the CAC function, a connection request is accepted
only when sufficient resources are available in order to guarantee the QoS
objectives of the new connection and to maintain the agreed QoS of the
existing connections as well.
Resources affected in terms of bandwidth:
• Cross connection fabric (ABUS) in ingress direction
• Transmission path (port) in egress direction
• VP in the case of VCC aggregating in egress direction
Resources affected in terms of quantity:
• Number of shaped aggregating VPs
• Number of matrix connections
• VPI, VCI range

Effective booked bandwidth

Introduction The effective bandwidth booked by an ATM connection depends on its traffic
descriptor as a load on the one hand and the peak cell rate combined with
the delay capabilities of the available resource on the other hand.
The critical resources to be looked at are the cross connect fabric and the
port in egress direction (ABUS towards port).

Ingress traffic The ABUS as the critical resource in ingress direction features only little
buffering with the result that the traffic can arrive at the ABUS' entry as
almost unfiltered bursts at peak rate.
Therefore the CAC relevant effective bandwidth for the traffic of all service
categories is defined by the connection PCR value but limited by the
physical port bandwidth.

Egress traffic Since traffic concentration takes place mainly at the egress ports, most
buffer capacity is allocated to VBR-nrt and UBR egress queues.
The effective bandwidth for real time traffic such as CBR and VBR-rt with
little delay is approximately its PCR value.
For the VBR-nrt traffic experiencing higher concentration and delays,
effective bandwidth will be between PCR and SCR.
Best effort UBR traffic is not subject to the CAC procedure.

Available bandwidth as a
resource
Egress ports The available bandwidth corresponds with the physical bandwidth.

Over-subscription Over-subscription causes inflating the available resources (the booked

bandwidth is not affected) in order to pass CAC if it can be assumed that
only a fraction of the connections will be active at the same time. The feature
may be helpful above all in conjunction with permanent connections (PVC).
Over-subscriptions shall be restricted to the aggregating VP as a resource
(see Fig. 3-1).

3-4 Technical Description System EN/LZTBU 220 102/2 RA

 Traffic management © KEYMILE AG

If over-subscription is enabled, the available effective egress bandwidth of a

VP is that as defined by the traffic descriptor (PCR) multiplied by a value
between 1 and 100.
By default over-subscription is disabled, that means that the multiplier is set
to 1.

Figure 3-3: VP with over-subscription

Port

Available bandwidth
of the port

Available bandwidth
VCs
of a VP with
oversubscription
VP Egress Direction

Bandwidth as derived from the VP traffic

descriptor.

(This bandwidth is subject to port CAC)

CAC on VP with VCCs

aggregating Figure 3-4: VP with VCC aggregating

Available VPC bandwidth

equals PCR value
(VPC shall be of CBR type)

VPC

VCC

VCC

Port
CAC of VCC by the VPC

CAC of VPC by the port

For more details on VCC aggregating see also section «Shaping» on

page 3-2.

Management resources The number of connections shall be limited to 2048.

EN/LZTBU 220 102/2 RA Technical Description System 3-5

 © KEYMILE AG

OAM 4
ATM OAM flows
Cross connections and The table below shows the relation between connection layer and the
endpoints available OAM termination layers (endpoints).
The supported OAM terminations depend on whether the connection is on
VP (Virtual Path) or VC (Virtual Channel) level and whether the VC
connection is between two ATM UNI ports (VC) or between a ATM UNI port
and a service SAR port (VC-AI).

Figure 4-1: Terminations depending on connection layer

Connection layer OAM termination layers (endpoints) Notes

VC-AI VC connection VP connection VC-connection between UNI and
(VC segment) (VP segment) SAR port
VC VC segment VP connection VC-connection between UNI (s)
(if enabled) (VP segment)

VP VP segment VP-connection between UNI(s)

(if enabled)

Note: UMUX does not support layers in brackets.

Segment endpoints are enabled/disabled on a per port level.

Fault management Fault management on ATM layer is in line with the fault management on
functions TDM . This particularly implies RDI ( remote defect indication) and AIS
(alarm indication signal) as part of the user traffic overhead. However there
are differences as how this information is being inserted.
Both the AIS and RDI flows help to keep the endpoints informed if a defect
has been detected along the path, which affects the availability of the end -to
end connection.
• Permanently activated
AIS, RDI defect indication: end-to-end VP / VC
• Activated on a per VP / VC level
LOC defect indication: end-to-end VP / VC, segment VP / VC

EN/LZTBU 220 102/2 RA Technical Description System 4-1

 OAM © KEYMILE AG

Figure 4-2: Example of defects and associated AIS / RDI flows

NE1 NE2 NE3

Defect case 1

F5
VC F5-RDI F5-AIS VC

VP F4-RDI F4 F4-AIS VP

TP F3-RDI F3-AIS TP

RS/MS F1,2-RDI F1- LOS RS/MS

STM1 STM1

Defect case 2

F5
VC F5- LOC F5-RDI VC

VP VP

TP TP

RS/MS RS/MS

STM1 STM1

Transport Layers Terminated OAM Flow

VC ATM Virtual Channel F5
VP ATM Virtual Path F4
Transmission Path F3

MS Multiplex Section F2
RS Regenerator section F1

The figure above shows defect situations in an ATM network consisting of

the network elements NE1 to NE3 interconnected by STM1 links.
• Defect situation 1
LOS (loss of signal) is detected at the input of NE2. Apart from being
reported as LOS in NE2 there are actions taken in the downstream and
upstream directions.
− Downstream:
In NE2, RS layer AIS is generated which then is detected as follows:
- On NE2 as AIS in the MS and TP layer
- On NE3 as AIS in the ATM VP (F4) and ATM VC (F5) layer
− Upstream to NE1:
- RDI on RS as a consequence of LOS in NE2
- RDI on MS and TP layer as a consequence of AIS detection
in NE2
- RDI on ATM VP (F4) and ATM VC (F5) as a consequence of
AIS detection in NE3

4-2 Technical Description System EN/LZTBU 220 102/2 RA

 OAM © KEYMILE AG

• Defect situation 2
F5- LOC (Loss of continuity) is reported on an ATM VC in NE1.
− Downstream:
As a result of the F5 LOC a TSF (Trail Signal Fail) will be inserted
towards the higher layers
− Upstream to NE3:
RDI on ATM VC as a consequence of F5-LOC in NE1

Loopback
Cell insertion/extraction LB cells can be inserted at CPs including connection segment endpoints and
connection endpoints.
Depending on the CP (VP or VC) and the loopback location selected the
following LLID are selectable:

Figure 4-3: Examples of loopback insertion

CP type LB-cell Loopback location LLID set to

VP e-t-e VP Connection Endpoint all 1
seg VP Segment Endpoint all 1
seg VP LLID CPID or
all 0 (not supported)
VC e-t-e VC Connection Endpoint all 1
seg VC Segment Endpoint all 1
seg VC LLID CPID or
all 0 (not supported)

VC

segment

LLID =all "1"

End-to-end loopback

LLID =CPID
LLID =all "1"

CP to seg.-end loopback CP to VC endpoint loopback

LLID =CPID

LLID =CPID
CP to CP seg. loopback

LLID =CPID

VC =Virtual Connection (VPC or VCC)

CP = Connection Point
LLID = Loop Location Identifier
CPID = Connection Point Identifier

Note: Only front-end loopbacks supported

EN/LZTBU 220 102/2 RA Technical Description System 4-3

 OAM © KEYMILE AG

Loop back technique In accordance to ITU-T I.610, B-ISDN operation and maintenance principles
and functions including LLID option.

ATM performance The UMUX does not support Performance Monitoring for ATM traffic
monitoring according to ITU-T I.610. However, the UMUX provides cell statistic on the
ATM layer (unfiltered mode).
Performance monitoring is available for services (CES, LES; V5) and the
layers of the transmission interfaces according to the principles of the UMUX
platform.

Congestion reporting
Congestion reporting comprises status of all ingress and egress queues.
Ingress congested seconds monitoring:
• Per unit service category buffer limit
Egress congested monitoring:
• Per unit service category and port buffer limit
• Per STM1 unit shaped VPs
Cell statistics are performed on a per connection point level.

Statistics
Cell statistics are performed on a per connection point level.
Counters are provided for
• Total of ingress cells / egress cells, CLP 0+1 (including OAM cells, but
without non-conforming cells, tbv)
• Total of ingress cells / egress cells, CLP = 0 only
• Total of ingress cells / egress cells, OAM cells only
• Non conformant ingress cells

Please note that the cell statistic counters have counting ranges and
characteristics as follows:
• Port level counters: 64 bit
• Connection level counters: 32 bit
Exceptions are counters for
− Non conforming cells: 16 bit (roll over at 64 k counts)
− OAM cells: 16 bit (roll over at 64 k counts)

4-4 Technical Description System EN/LZTBU 220 102/2 RA

 © KEYMILE AG

Auxiliary functions 5
Network synchronisation
General Although ATM is considered to be asynchronous, the services, such as
voice over ATM, conveyed over ATM may require clock information to
restore user service information. Since ATM by its nature is confined to the
transmission of data, the associated timing signal has to be extracted either
from the data flow or by the aid of the network synchronisation.
Timing information extracted from the data flow normally exhibits rather high
jitter and wander values which disqualifies it for instance as common
reference for narrow band services. In this case provision shall be made to
derive the service clock from the ATM network synchronisation.

EN/LZTBU 220 102/2 RA Technical Description System 5-1

 Auxiliary functions © KEYMILE AG

Synchronisation network A network normally consists of nodes physically interconnected by links. To

achieve a common network clock each node shall have access to a
synchronisation trail which is composed of a number of dedicated link
connections. To protect against failures, several trails are offered to select
from. The selection algorithm shall always guarantee for the best possible
NE systems clock. The NE in turn will distribute its timing to all the adjacent
NEs independently on whether the clock is locally used as a timing source or
not. Special control mechanisms are applied to avoid timing loops.
The network clock of the PSTN is normally traceable to PRC (Primary
Reference Clock) with an accuracy of 10-11.
This accuracy allows you to run the PSTN service without significant
degradation over partly synchronised networks with plesiochronous
interfaces in between.

Figure 5-1: LES Application Network

E12/P12s V5.2
IAD STM-1
CP IWF
STM-1 E12/
DSLAM
Regional Voice GW P12s
CP-IWF Broadband CO-IWF PSTN
Network
IAD
CP-IWF DSL

OSN, RSN
xDSL OS1, RS1 E12
Synch Trail OS1, RS1

OSN, RSN

PRC

The figure above shows the elements (Regional Broadband Network, PSTN,
VoiceGW, DSLAM and IAD) of a LES network and an example of the
synchronization trail associated with. The VoiceGW, the DSLAM and the
IAD represent network elements in terms of management and
synchronisation.
The PSTN and the Broadband Network also consist of switches as their
network elements.
Ideally the synchronisation trail makes use of the physical section layer
interconnections (unprotected termination). Higher layer paths may suffer
from high jitter excursions due to coding (e.g. ATM CES), protection
switching and pointer adjustments.

5-2 Technical Description System EN/LZTBU 220 102/2 RA

 Auxiliary functions © KEYMILE AG

Equipment timing source The synchronisation model for each network element consists of a selection
and a distribution part. As an example the one for SDH is shown in the figure
below.
Simpler implementations may consist just of Select B and a PLL.

Figure 5-2: Synchronous equipment timing source

Select Squelch
A Select
T4
C

Squelch
T1

T2
T3 Select SETG T0
B (PLL)

Internal

SETG Synchronous Equipment Timing Generator function

Internal Internal Oscillator function
T0 Clock for NE internal distribution
T1 Clock derived from an STM-1 signal
T2 Clock derived from a 2Mbit/s traffic signal in accordance
to G.703 / 704
T3 Clock derived from a 2MHz clock input acc. to G.703
T4 Clock towards a 2MHz clock input acc. G.703 (external SSU)
Each type of input (T1 to T3) may have any number of instances. Not all the
types of timing sources are always available.
The selection of the actual source will be according to the following criteria
1. No signal failure (LOS, LOF etc)
2. Traceable quality level indicated by SSM (e.g. PRC, SSU, SEC)
3. Priority
The selection function helps to maintain NE synchronisation even if in case
of a synchronisation trail failure. In this way the synchronisation network has
its own protection mechanism.

EN/LZTBU 220 102/2 RA Technical Description System 5-3

 Auxiliary functions © KEYMILE AG

ATM network elements

IAD The IAD will preferably be synchronized to the VoiceGW via the DSLAM and
the xDSL physical link, the xDSL clock being synchronized to the DSLAM's
NE clock
Note: If the xDSL contains a P12s client layer synchronized by the DSLAM's
NE clock, it is basically possible to use it to recover the timing information.
But this signal is rather sensitive to phase shifting as a result timing source
switching.

Voice Gateway The VoiceGW will be synchronized to the PSTN network via the E12 V5.2
links or by the Regional Broadband Network via the STM-1 link if a timing
reference traceable to PRC is available.
The timing information of the E12 or the STM-1 input signal is used as an
input to the SETS which re-clocks all the output signals.
The VoiceGW is a UMUX based NE with the network timing relevant units:
• ATIOP T1 (STM-1 ATM UNI)
• LOMIF T2 (E12/ V5)
(For the definition of T1 and T2 see Fig. 5-2).

DSLAM The DSLAM should be part of the synchronization network to provide

network clock for the restoring of synchronous user service interfaces either
on the IADs associated with or as part of the DSLAM equipment itself.
Note: A pure DSLAM (without CP-IWF) actually need not to be
synchronized because the ATM layer is not terminated.
The DSLAM is synchronized via the ATM user network interface. In the case
of an STM-1 UNI the SETS configuration looks similar to the one of the
VoiceGW. (apart from the 2Mbit/s interfaces the to the PSTN)
If an E12 ATM network interface is used only PETS will be involved.
The DSLAM is a UMUX based NE with the network timing relevant units:
• ATIOP T1 (STM-1 ATM UNI)
• ACONV T2 (P12s CES) (adaptive clock recovery implies high
phase variations)
• ADAC T2 (xDSL, future release)
• LOMIF T2 (E12 ATM IMA/UNI)

Unit timing configurations

Existing TDM units The timing configurations of existing TDM units (e.g. LOMIF) used in
conjunction with ATM are not especially looked at. But in many applications
they accommodate the physical ATM interface which is best qualified as a
timing source.

ATM units Please refer to the specific unit descriptions

5-4 Technical Description System EN/LZTBU 220 102/2 RA

 Auxiliary functions © KEYMILE AG

Management communications
ECC over ATM The figure below shows two cascaded UMUX ATM network elements
managed by the Element Manager EM via the embedded communication
channel ECC.
In non ATM network the ECC is conveyed over specific overhead bytes or
over a normal payload TDM path. Even in ATM networks the TDM layer
might be still available for the transport of management information. E.g. the
ECC between NE1 and NE2 can either make use of the D1 to D12 bytes of
the STM1 section overhead or of a ATM layer VCC.
Since the main focus of this document is the ECC implementation on ATM,
the already existing TDM approach is not discussed here in details.
The implementation of the ECC should also be looked at in the context of
protection switching.

Figure 5-3: Typical ECCoATM Network

UMUX NE 1 UMUX NE 2
(ATM) (ATM)
Qx

UCST
Qx TCP/IP ATM Network

PPPoAAL5 PPPoAAL5 PPPoAAL5

IP ATM ATM ATM
Ethernet ATM PHY ATM PHY ATM PHY
UNEM

PPP
P0-nc

Alternative stack if section

overhead bytes can be used
instead of an ATM VCC

EN/LZTBU 220 102/2 RA Technical Description System 5-5

 Auxiliary functions © KEYMILE AG

ATM ECC UMUX functional The additional feature to be implemented for the ECCoATM is the PPP
blocks relaying function as shown in the figure below within dotted lines.
The ECC over ATM uses an AAL5 <-> P0_nc PPP relay function.

Figure 5-4: UMUX internal ECCoATM processing

ATM Interface Unit Control Unit

ATIOP /ACONV COBUX

IP Relay
PPP Relay IP IP
PPPoAAL5 PPP PPP Ethernet
ATM P0-nc P0-nc

UTOPIA TTP UBUS /PBUS Qx

CTP
10BaseT
ATM PHY

STM1
IMA

Each of the ECC over ATM channels has the following parameters:
• Configurable
− Bandwidth: 64, 192 and 576 kbit/s
− VPI
− VCI

• System defined (automatically created)

− QoS: CBR (CBR.1)
− PCR: 80.1 (70.8) kbit/s
240.4 (212) kbit/s
720.4 (636) kbit/s
− CDVT: 5000 (600) µs
(Rates in brackets apply up to the UCST R6A SP01)

The ECC over ATM is available on STM1 and IMA UNI as implemented on
ATIOP and ACONV.

5-6 Technical Description System EN/LZTBU 220 102/2 RA

 Auxiliary functions © KEYMILE AG

Protection
For ATM units the following types of protection are provided:
• Multiplex section protection MSP, to protect STM1 traffic against traffic
failures and equipment failures
• Equipment failure protection EQP, to protect traffic, ABUS master and
SETS against equipment failures.

Figure 5-5: Protection configuration with ATM units

Units Type of protection Configuration

ATIOP MSP 1:1 (bi-directional) Protection function enabled on working unit. Requires
STM1 protecting unit in associated slot.
Switching controlled by APS protocol.
ATIOP EQP 1:1 SETS/ABUS Master is assigned automatically to the
SETS/ABUS Master first unit being configured.
Protection available if protecting unit added in
dedicated slot.
May be located on the same units featuring MSP.
However switching actions are not necessarily
synchronized. (MSP switchover caused by a traffic
failure will have no effect on SETS protection).
ACONV EQP N:1 Configuration of protection group consisting of up to 4
VGATE ABUS - PBUS units working units associated to a single protecting unit. Up
to four groups.
ACONV EQP Inherent Automatic selection between active (non failed) units.
ABUS Master Protection against equipment failures inherent to
selection mechanism, no specific protection function.
In case of N:1 EQP working units are of higher priority
than protecting units, irrespective of its grouping.

EN/LZTBU 220 102/2 RA Technical Description System 5-7