Title: Welcome to join Tutorial
1Welcome to join Tutorial
An Introduction to LAPS and MSR Applications
To the March 2004 SG17 Meeting
Q.7/17 Rapporteur
2Our Liaisons and Communications
3X.85/Y.1321 (IP over SDH using LAPS) milestone
1?Delay contribution, August 1998 2?It was
acceptable by ITU-T SG7(Data network and Open
System Communication) at the September meeting,
1998 3?X.85/Y.1321 on IP over SDH using LAPS was
determined at the June 1999 meeting
4?Recommendation X.85/Y.1321) was approved at
March 2000 meeting, then new version is available
Feb. 2001
4X.85/Y.1321 Comments received
1?IETF and ISOC 2?ITU-T SG15 (Optical and other
transport networks) 3?ITU-T SG11 (Signaling
requirements and protocols) 4?ITU-T SG13
(Multi-protocol and IP-based networks and their
internetworking) 6?Nortel 7?NTT 8?Swisscom
9?Lots of email from Vendors and Carriers
5Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
Note PPP is widely deployed in networks around
the world and has been updated and extended over
the last 14 years. These slices just emphases a
simple LAPS method of high-speed link (e.g. POS)
and also provide compatibility with PPP/HDLC.
6Why LAPS (X.85)
7Why LAPS (X.85)
The diagram of Router
Line Card 0
Line Card 7
Switch Fabric
Line Card 1
Line Card 8
Scheduler
Line Card 6
Line Card 10
RE
RE or LC11
8Why LAPS (X.85)
POS Line Card
POS PHY
POS Framer
STM-16c Transceiver
Network Processor
I/F
O/E
Switch Fabric
Memory
POS Line Card
16 x 16
SPI-4.2
POS Framer
STM-16c Transceiver
Network Processor
I/F
O/E
Routing Engine
Memory
POS Line Card
9Why LAPS (X.85)
Local node
BGP
RIP
SNMP
Software
Routing Engine
LCP
UDP
OSPF
TCP
Fwd/Rcv to/from
IP
Hardware
Adjacent Node
Forwarding Engine
PPP,LCP,IPCP
Filter Function
POS PHY/Utopia3 Reference Point
HDLC
SDH/SONET
Protocol processing of signaling and Traffic plane
10Why LAPS (X.85)
- The major objective of X.85 is to remove PPP
protocols including LCP and IPCP in the case of
POS. - LCP contains 10 configuration packets,16 events,
and 12 actions.
11Why LAPS (X.85)
12Why LAPS (X.85)
X.85 vs. RFC 2615, specifications including
functions and related Network management
RFC 2615RFC 1661 RFC 1662 RFC 1570 RFC 1547 RFC
1340 SNMP MIB
X.85 SNMP MIB
13Why LAPS (X.85)
LAPS or POS HDLC Framer/Deframer
functions Insertion of HDLC frame into the
SPE Framing, Inter-frame fill and transmit FIFO
error recovery. Scrambling (X43
1), Transparency processing generate a 16/32 bit
FCS. Extraction of HDLC frame, Transparency
removal, De-scrambling (if enable), FCS error
checking, Optional delete the HDLC address and
control fields.
T
R
14Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
15Why LAPS (X.85)
Comparison of Protocol states RFC 26152137,
LAPS2
16Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
17Why LAPS (X.85)
18Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
19Why LAPS (X.85)
The Diagram of Network Processor
SSRAM
SDRAM
Input stream scheduler
Search update
Receive editor
Queue mgnt.
parsing
Framer/Deframer
Switch Fabric
System I/F
Fabric I/F
Pre-search queue
Output Stream scheduler
Queue mgnt.
Sending editor
Statistics internal registers
CPU I/F
CPU
SDRAM
20Why LAPS (X.85)
21Why LAPS (X.85)
RFC 2615(PPP/HDLC) LAPS
Cell based MPS/mPS
1600/10160 1600/4040 1 Latency
NP NP
Latency
variance 4 times -
good Packet loss 4 times
- very low Capability of
Jumbo payload processing relative
lower relative power
good Congestion processing capability low
middle relative high QoS
3 times lower normal
good
22Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
23Why LAPS (X.85)
24Why LAPS (X.85)
25Why LAPS (X.85)
26Why LAPS (X.85)
When the PPP is used to be encapsulated via SAPI
for the compatibility with RFC 2615, it is
noted (1)Both FCS-32 and FCS-16 can be set by
provisioning and is not negotiated. The 32-bit
FCS must be used for all SDH rates. For
STM-1c/VC-4 only, the 16-bit FCS may be used,
although the 32-bit FCS is recommended. (2)Regardi
ng the path signal label (C2) of SDH, for
compatibility with RFC 2615, the signal label
value of (x43 1) scrambling is changed from 24
(18 hex) to 22 (16 hex). Additionally, the LAPS
does also provide the signal label value 207 (CF
hex) to indicate PPP without scrambling. (3)The
data link will be operated as RFC 2615 defines
and the Address field is set to 11111111, the
padding field followed information field and the
functions of Link Control Protocol and Network
Control Protocol will be included.
27Why LAPS (X.85)
1?Simple implementation 2?High efficiency in
the POS line card of router 3?Function
equivalent to PPP/HDLC 4?Performance of Carrier
concern 5?Compatibility with PPP/HDLC 6?Test
equipment (share POS)
28Why LAPS (X.85)
Testing equipment 1?Smartbits are used to
throughput 2?RouterTest/Adtech is used to
traffic and routing protocols
29X.86 (Ethernet over SDH/SONET) introduction
30LAPS (X.86) milestone
1?Delay contribution from May 1999 2?It was
acceptable by ITU-T SG7(Data network and Open
System Communication) at the June meeting, 1998
3?X.86 on Ethernet over LAPS was determined at
the March 2000 meeting 4?Recommendation X.86 on
Ethernet over LAPS (TD 2046/Rev.1) was approved
at Feb. 2001 meeting, ten months earlier than
that of GFP
31LAPS (X.86), three types of application
32Latency Variance Computation
Ethernet MAC -15µs Rate Adaptation, Buffer
-15µs LAPS mapping, Buffer -15µs LAPS CRC, Buffer
-15µs ----------------------------------------- La
tency Variance Total-60µs GE on GE switch -4µs
33The competitive advantages of X.86
- Remote Trail Performance Monitoring
- Remote Fault Indication
- IEEE802.3x Active Flow Control in Burst
Traffic Condition - Low Price and Ease of Use (Compared to LANE)
- Low Latency and Low Latency Variance
- 11 redundancy based Ethernet and Gigabit
Ethernet service - Target at existing telecom transport resources
34X.86 does match Ethernet and Gigabit Ethernet
very well
35Understand GFP
(1)Payload (2)Payload Header (3)Optional Payload
FCS (4)PLI value (5)cHEC computation
36Understand GFP
- HEC inherits from ITU-T Rec. I.432,Octet based
spec. - ATM is L2/L3 technologies, based on connection
and fixed packet size (cell), IP as a ATM client
is flexible rate for the IP over ATM applications
in the most case GFP is L1/L2 technologies,
based on connectionless and variable packet size.
For the L1-GFP, FE/GE as a GFP client is rigid
rate for the Ethernet over SDH applications in
the most case - In terms of PDH(E1/T1/E3/E4), Jumbo Payload size
of IPv6 can be greater to 64Kbits length due to
GFP-PLI is 2-Octet definition (216bits64Kbits8Kb
ytes)?If PLI is changed to 3-octet or more, how
to keep compatibility with existing GFP standard - Flow control?Rate adaptation?
37Understand GFP
Why Flow control? Why Rate adaptation?
(1)FE/GE Bandwidth?VC or VCs concatenation
Bandwidth
(2)No difference is applied for the two
time-slice (TS) if Ethernet data stream is mapped
into VC overhead or/and VC payload during two
different time-slice
Mapping octet by octet
VC overhead
VC payload
TS2
TS1
Real-time
38Understand GFP
- Issue of tiny-frame (7 bytes), Client Signal Fail
(LOS of Client Signal and Loss of Character
Synchronization)? - 4-octet idle issue?
39Understand GFP
tiny-frame issue for IPv4 over SDH using GFP(1)
L7
Application
End system
L6
Presentation
L5
Session
L4
Transport
Network
Intermediate system,40-1600 octets
L3
L2
GFP
Intermediate system ,7 bytes
L1
Physical Layer
Open System Interconnection
40Understand GFP
tiny-frame issue for IPv4 over SDH using GFP(2)
Line Card
POS-PHY bus/SPI-4.2 bus
Routing and signaling
155M POS ---8?25M 622M POS --- 16?50M 2.5G POS
--- 32?100M,64?50M
41Understand GFP
tiny-frame issue for IPv4 over SDH using GFP(3)
GFP LAPS
Cell based MPS/mPS
1600/7228 1600/4040 1
Latency variance
great middle little
MPS Maximum packet Size, 1600 octets specified
by IETF mPS Minimum Packet Size , 40 octets
specified by IETF
42Understand GFP
4-octet idle issue
Mapping relationship and timing, FE/GE is mapped
into GFP, LAPS and RPR(LAPS/GFP)
43Understand GFP
Ethernet Frame
IFG
PreambleSFD
IFG
PreambleSFD
8
64-1518
8
12
12
64-1522 if VLAN
GFP frame
GFP Payload
4
4
4
2
2
2
2
4
2
2
4
4
2
Idle
Idle
Idle
tHEC
PLI
Idle
Idle
PLI
cHEC
Idle
eHEC
Type
CID Spare
FCS
FCS
Real Time
Case 1 - Octets Mapping of Ethernet using GFP
44LAPS GFP
68-1522 if VLAN
45(No Transcript)
46LAPS GFP
68-1522 if VLAN
PreambleSFD
PreambleSFD
IFG
IFG
Ethernet Frame
12
64-1518
8
8
12
64
Next frame
RPR Frame
2
2
RPR Payload
2
6
6
16
Header CRC
DA
FCS
SA
Ring Control
Protocol Type
?
Packet Loss
GFP frame
GFP Payload
4
4
2
2
2
4
4
2
4
4
2
4
4
2
2
tHEC
Idle
Idle
Idle
PLI
CID Spare
Type
cHEC
PLI
Idle
Idle
Idle
eHEC
Timing border
FCS
FCS
4
4
Variable
Idle
Idle
Real Time
Case 3 - Octets Mapping of Ethernet/RPR/GFP
47LAPS GFP
68-1522 if VLAN
IFG
IFG
PreambleSFD
PreambleSFD
Ethernet Frame
8
12
64-1518
8
12
64
Next frame
RPR Frame
2
6
6
2
16
RPR Payload
2
Header CRC
DA
FCS
SA
Ring Control
Protocol Type
Good
No Packet Loss
GFP frame
LAPS Payload
1
1
4
N
1
1
1
N
1
1
2
Flag
Idle
Flag
Flag
Flag
Ctrl
Addr
Flag
Flag
SAPI
Timing border
FCS
FCS
N
1
1
Fixed
Flag
Flag
Real Time
Case 4 - Bytes Mapping of Ethernet/RPR/LAPS
48LAPS GFP
Comparison of Measurement
GFP LAPS/X.86 Percentage
64bytes 10.520 µs 9.658 µs 8.9
1518bytes 203.620 µs 133.967 µs 51.9
9.6Kbytes - 769.567 µs
49Advantage of using LAPS
- Flow-control and rate adaptation are very useful
for BW mismatch between Ethernet and SDH VC (or
concatenation) and time-slice difference between
VC overhead and VC payload - Application of Both SDH and PDH(E1/T1/E3/E4) for
LAPS, Both SDH and PDH(DS3) for GFP - High efficiency
- LAPS is compatible with POS and PPP/HDLC
- LAPS Idle(Flag) has good performance for
line-speed Ethernet over SDH/SONET applications - X.86 was approved ten months earlier than that
of GFP
50LAPS GFP
Disadvantage of LAPS and PPP/HDLC (1) Escape Code
51An Introduction to MSR Applications
- What is MSR?
- Why MSR?
- MSR Scope
- Multi-service capability
52What is MSR(1)
TCE (1)
TCE (2)
MSR (2)
53What is MSR(2)
Station A
Station B
Small pipe tributary (just like PVC)
Big pipe Aggregate
54An Introduction to MSR Applications
- What is MSR?
- Why MSR?
- MSR Scope
- Multi-service capability
55Basic Consideration(1)
- Focus on Metro
- Sync. to async., TDM to packet, Physical
transmission from SDH/SONET to MAC, High-order VC
and Lower-order VC to tributary (packet) - MSPP/MSTP from TDM based to packet based
- Push multi-service transport issue to LLC(L2)
56Basic Consideration(2)
- Differentiated SLA, controllable and
configurable service operation - Static and dynamic resource management
- Both connection and connectionless
- Bridge and back-to-back connection between
MSR/RPRs - Lower cost and little maintenance
57Why MSR(1)
differentiates RPR transport capabilities and
provides multi-congeneric-service and
multi-heterogeneous-service transport including
Ethernet, Gigabit Ethernet, FR, G.702 PDH circuit
-- Synchronous and asynchronous circuit
transport, Video signal, voice-band signal,
Digital channel etc.
58Why MSR(2)
Differentiates RPR ring protection function and
provides service (or tributary) based standby
(protection) of 11, 11, and 1N models within
50 ms.
59Why MSR(3)
Reassembly a single packet based multicast and
broadcast of RPR to form service (stream)
multicast and broadcast capabilities, and provide
Service (or tributary) based multicast and
broadcast.
60Why MSR(4)
Adding bandwidth management function for each
service (tributary) and provides service
bandwidth management with either symmetry or
asymmetry.
61Why MSR(5)
Adding bandwidth merging function for
congeneric-service (congeneric-tributary) and
provides tributary merging with either symmetry
or asymmetry.
62Why MSR(6)
Adding the simple security filtering function for
each tributary service and provides line-rate
filtering to monitor and manage service security.
63Why MSR(7)
Adding the performance management function for
each tributary service and provides tributary
performance monitoring in 15-minute and 24-hour.
64Why MSR(8)
- QoS Guarantee
- Customers separation
- Addresses separation
- Topologies separation
65Why MSR(9)
- Topology supported
- Two fiber ring
- Link
- Broadcast topology
66Why MSR(10)
Dynamic resource management
Next job
Static resource management
Tributary transport
67An Introduction to MSR Applications
- What is MSR?
- Why MSR?
- MSR Scope
- Multi-service capability
68X.87 scope
IP
X.87 layer client
Client
X.87
XP UNACK Data Req.
Reference Point T1/T2
Ind.
RPR MAC client X.87 layer
X.87
MSR
MAC Data Req.
MAC Ctrl Req.
Reference Point G1/G2
Ind.
Ind.
RPR MAC
RPR
GFP/HDLC
LAPS
GFP/HDLC
LAPS
SDH/SONET
WEST PHY
EAST PHY
S4
S0
S1
S2
Ringlet 1
RPR-MSR protocol stack
S5
S7
S6
S62
S8
Ringlet 0
69Layered Model (1)
70Layered Model (2)
IP
Other
X.87
Multi-service
MSR
802.3 MAC
802.3 PHY
802.3 MAC-MSR protocol stack
71Layered model of Multi-service node
72Agg
Agg
MSR node
MSR node
MSR node
Trib
LinkADM
Trib
MSR node
MSR node
73(No Transcript)
74An Introduction to MSR Applications
- What is MSR?
- Why MSR?
- MSR Scope
- Multi-service capability
75Differentiate RPR (1)
76Differentiate RPR (2)
77Differentiate RPR (3)
78Differentiate RPR (4)
79Differentiate RPR (5)
80Differentiate RPR (6)
81Applications
Backbone
Backbone MSR
Access MSR
MAN/STM-64
Access MSR
82Applications POP Interconnect
RPR 2X2.5 Gbps
STM-16 2.5 Gbps
POS/EOS
RPRMSR 2X2.5 Gbps
83Further study in Q.C/17 (old Q.7/17)
- Maintenance of X.85/Y.1321 and X.86/Y.1323,
Update and extension of X.87/Y.1324 - Progress draft Recommendation X.msr, including
associated Ethernet/Gigabit Ethernet/10G Ethernet
aspects - Develop Requirements for MSDN, areas of study
and development include - - Identification of market needs
- - architectural considerations, for L2 data
networks - - multi-service multicast aspects
- - Ethernet UNI and NNI aspects.
- Enhance existing packet protocols or, if
required, develop new packet protocols to support
the developed MSDN requirements, including
service mechanisms - Develop associated MIBs (Management Information
Base) to support item (iv)
84Thank You !