Ethernet Transport Over PDH Networks With Virtual Concatenation Tutorial

Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM

Ethernet Transport over PDH Networks with Virtual Concatenation Tutorial

White Paper

Steve Gorshe Principal Engineer

Issue 1.0: November, 2005 PMC- 2050380

© 2005 PMC-Sierra, Inc.

Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM

Ethernet Transport over PDH Networks with Virtual Concatenation White Paper

Abstract Virtual concatenation and LCAS were originally defined for SONET/SDH signals and G.709 OTN signals. Asynchronous hierarchy (DSn) and PDH signals lacked the overhead bandwidth required for the virtual concatenation and LCAS overhead channel. DSn and PDH signals are still leased by carriers as the means of accessing their enterprise customers through another carrier’s network. Aside from the ubiquitous availability of DSn/PDH connectivity, there are regulatory conditions in the United States that give DSn connectivity a price advantage for this application. When the virtual concatenation of DSn and PDH signals was requested by multiple carriers, modified signal frame formats were defined to create the required overhead channel. These frame formats are defined in new ITU-T Recommendation G.7043, and the topic of this white paper. This white paper also covers the mapping of GFP payloads into these signals, as defined in new ITU-T Recommendation G.8040.

About the Author Steve Gorshe, Ph.D. is a Principal Engineer in the Product Research Group and oversees ICs for SONET, optical transmission and access systems.

Currently Steve is a senior member of the IEEE and co-editor for the IEEE Communications magazine’s Broadband Access Series. He is the chief editor for the ANSI T1X1 Subcommittee, which is responsible for SONET and optical network interface standards. He is a recent recipient of the Committee T1 Alvin Lai Outstanding Achievement Award for his standards work and has been a technical editor for T1.105, T1.105.01, T1.105.02, and T1.105.07 within the SONET standard series as well as the ITU-T G.7041 (GFP) G.7043 (Virtual concatenation of PDH signals), G.8040 (GFP mapping into PDH signals), and G.8011.1 (Ethernet Private Line Service) recommendations. He has 26 patents issued or pending and several published papers.

Revision History Issue No. Issue Date Details of Change

1 November, 2005

Document created

PMC-2050380 1 © 2005 PMC-Sierra, Inc.

Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


Contents Abstract.............................................................................................................................. 1 About the Author............................................................................................................... 1 Revision History................................................................................................................ 1 Contents............................................................................................................................. 2 List of Figures ................................................................................................................... 3 List of Tables ..................................................................................................................... 3 1 Introduction ............................................................................................................... 4 2 Alternatives for creating N x PDH signal channels ............................................... 5 3 The Solution............................................................................................................... 7 4 Mapping GFP frames into PDH channels (single and virtually

concatenated)......................................................................................................... 11 5 Conclusions............................................................................................................. 12 6 References ............................................................................................................... 13 7 Notes ........................................................................................................................ 14


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


List of Figures Figure 1 Example of PDH access through an ILEC........................................................ 4

Figure 2 Multiframe format for the virtual concatenation of asynchronous / PDH signals............................................................................................................... 7

Figure 3 LCAS control packet format for PDH signals.................................................. 10

Figure 4 Examples illustrating the mapping of GFP frames into PDH channels .......... 11

List of Tables Table 1 Comparison of technologies for inverse multiplexing into NxPDH

signals............................................................................................................... 6


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


1 Introduction Since the global transport network has evolved to be based primarily on SONET/SDH technology, it may seem strange to want to add new capabilities to the previous generation asynchronous / plesiochronous digital hierarchy signal (e.g., DS1 and E1) technologies1. These networks, however, are still ubiquitously deployed and are still more common than SONET/SDH signals for enterprise access applications. Among the reasons for their ongoing prevalence in the enterprise access networks is that many of the access interfaces are still delivered over copper wires.

At least as important as the effectively ubiquitous availability of plesiochronous networks is the artificial advantage they have due to the regulatory unbundling of services. As part of unbundling, U.S. incumbent local exchange carriers (ILECs) are required to offer DS1 and DS3 access links to other carriers, such as competitive local exchange carriers (CLECs) or interexchange carriers (IECs), for lower tariff rates than equivalent SONET interfaces. The result of the tariff advantage and the availability of DS1 and DS3 connectivity is that when IECs or service providers lack their own facilities to connect to their enterprise subscribers, they typically lease DS1 or DS3 connections through the ILECs. An example of this network configuration is shown in Figure 1.

Figure 1 Example of PDH access through an ILEC

IEC orService Provider

Network

ILECTransport Network

(e.g., SONET)CPE CLE

Ethernetinterface

DS1, DS3, NxDS1,or NxDS3 link

UNI

This PDH situation, combined with the growing interest in providing native Ethernet connectivity, leads inevitably to a desire to map Ethernet into PDH signals. Although a number of proprietary implementations existed, there were no standards for mapping native Ethernet into DS1, DS3, NxDS1 and NxDS3 signals. In order to provide Ethernet connectivity to their enterprise customers over DS1/DS3 connections, the major U.S. IECs asked for GFP mappings into DSn and En signals. (GFP provides an encapsulation of native Ethernet frames in order to carry them through a transport network. See PMC-Sierra white paper PMC-2041083.) The resulting mappings were specified in the new ITU-T G.8040.

1 Both the North American asynchronous hierarchy and the plesiochronous digital hierarchy (PDH) will be referred to in this white paper as PDH for convenience.


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


2 Alternatives for creating N x PDH signal channels The GFP mapping into a single DS3 signal was defined first while mappings into DS1, E1, NxDS1, and NxE1 were studied. Carriers wanted to have the NxDS1 and NxE1 connections and use N=1 for mapping into single signals. Subsequently, interest developed for similar NxDS3 and NxE3 signals (e.g., for carrying data from 100Base Ethernet interfaces). Ideally, the NxDS1/E1/DS3/E3 should operate at Layer 1, providing transparent transport of Layer 2 protocol frames, independent of which Layer 2 protocol is being carried. The only existing non-proprietary solution was the Multilink Point-to-Point Protocol (ML-PPP defined in IETF RFC 1990), which performs inverse multiplexing2 at the packet level. Since ML-PPP is a Layer 2 protocol, it requires terminating the Ethernet signal in order to remap the packets into ML-PPP (i.e., change between the two different Layer 2 protocols)3. No byte level inverse multiplexing schemes such as VCAT existed since DS1 and DS3 signals lacked sufficient overhead to support VCAT, and reserving an entire payload channel for the overhead was too much capacity to lose4. Table 1 shows a comparison of the different candidate technologies that were considered.

2 Inverse multiplexing refers to taking the payload from a higher rate channel and transporting it by distributing it over multiple lower rate channels. The granularity used for assigning the payload data among the lower rate channels can be at the bit, byte, or packet/cell level.

3 Of course, ATM solutions existed, including Inverse Multiplexing over ATM (IMA). The carriers requesting the new mapping did not favor an ATM solution for this application due to its overhead inefficiency and it being another layer to provision.

4 Another potential solution existed from the Bandwidth ON Demand Interoperability Group (BONDING) consortium. Inverse multiplexing here is performed at the byte level rather than the packet level. An initialization sequence is sent on all the constituent lower-rate channels in order to synchronize the source and sink. While this technique requires no per-packet or per-link overhead, the channel must be disrupted for a long period of re-initialization when the channel size is changed.


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


Table 1 Comparison of technologies for inverse multiplexing into NxPDH signals

OPTION ADVANTAGES DISADVANTAGES Layer 2 frame inverse multiplexing

Proven technology exists for ML-PPP and Ethernet Link Aggregation No overhead required for each individual E1/DS1/DS3/E3 link Easy to add or remove links(trivial control protocol) NOTE – Layer 1 (i.e., GFP) packet interleaving was also considered, with at least one proprietary solution existing. Although it provides the Layer 2 transparency, it otherwise has the same advantages and disadvantages as Layer 2 packet interleaving.

Layer 2 technology specific – It either enforces a Layer 2 approach or requires re-mapping client data packets. Requires additional per-packet overhead (e.g., for packet sequence numbering) Egress queue management more complex due to the need to re-align the packets from the different links in the correct sequence. When there is a light load, a single link (or subset of links) is used for each packet rather than the entire set. This results in increased latency for lightly loaded cases. Under any load condition, the egress queue management will tend to introduce additional latency.

Byte inverse multiplexing with overhead barrowing

Relatively simple. Uses no additional per-link or per-packet overhead.

Changing the number of links (members) requires taking the connection down for a link re-synchronization.

Byte inverse multiplexing with permanent overhead channel

Simple (trivial) egress buffer since out-of-order packet arrival is not possible. Can directly re-use SDH virtual concatenation technology. No additional per-packet overhead. Consistency with VCAT and LCAS provides operational consistency and network predictability for the carrier.

Requires per-link overhead. Control protocol for adding and removing links is more complex (same complexity as LCAS).


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


3 The Solution The solution, specified in the new ITU-T G.7043, was to adopt byte-level inverse multiplexing, but to borrow one octet from the signal’s payload area once per multiframe to carry the per-link overhead information rather than permanently reserving the entire time slot. Figure 2 shows the resulting multiframe formats for the DS1, E1, DS3, and E3 signals.

Figure 2 Multiframe format for the virtual concatenation of asynchronous / PDH signals

FFFFF

FF

12345

2324

Concatenation overhead octet

Fram

e nu

mbe

r

sµ125

a) DS1 multiframe format for virtual concatenation

12345

1516


Fram

e nu

mbe

r

T.S.0T.S.0T.S.0T.S.0

T.S.0T.S.0

T.S.0

sµ125

b) E1 multiframe format for virtual concatenation


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


Figure 2 – continued

X1X2P1P2M1M2M3

F1 F2C11C21C31C41C51C61C71

F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4

C12C22C32C42C52C62C72

C13C23C33C43C53C63C73

680 bit subframe

84 bits

7subframemaster-frame


c) DS3 multiframe format for virtual concatenation

FA1 FA2 VLI

EM

TR

MA

NR

GC

529 octetpayload

59 colums

9 ro

ws

Concatenation overhead

d) E3 multiframe format for virtual concatenation


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


As shown in Figure 3, the overhead carries the same type of LCAS5 control packet as is used in the SONET/SDH H4 byte. (See PMC white paper PMC-2030895.) The PDH and SONET/SDH control packet format and bit definitions are identical except for the number of bits used in the sequence number (SQ) and the specific multiplexing of the member status information into the MST bits of each control packet. The transmission order of the control packet in Figure 3 is left to right for the bits, and top to bottom for the octets. In all of the fields, the MSB is the first bit to be transmitted. In the case of the SQ, SONET/SDH allows a maximum of 256 members and hence uses a two-nibble (8-bit) SQ field. The maximum number of members is 16 for DS1/E1, and is eight for DS3/E3. Hence, they require a SQ field of 4-bits and 3-bits, respectively. Since the SQ values are justified to the LSBs with the upper, unused SQ field bits set to 0, the SQ field use is still consistent for SONET/SDH (H4), DS1/E1, and DS3/E3 member types.

5 The LCAS (Link Capacity Adjustment Scheme) for virtually concatenated channels is specified in ITU-T Rec. G.7042.


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


Figure 3 LCAS control packet format for PDH signals

Concatenation overhead octet definition

Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8

Control Packet MFI1

MST (bits 1-4) 1 0 0 0

MST (bits 5-8) 1 0 0 1

0 0 0 RS-ACK

1 0 1 0

Reserved (0000) 1 0 1 1

Reserved (0000) 1 1 0 0

Reserved (0000) 1 1 0 1

Reserved (0000) 1 1 1 0

SQ bits 1-4 1 1 1 1

MFI2 MSBs (bits 1-4) 0 0 0 0

MFI2 LSBs (bits 5-8) 0 0 0 1

CTRL 0 0 1 0

0 0 0 GID 0 0 1 1

Reserved (0000) 0 1 0 0

Reserved (0000) 0 1 0 1

C1 C2 C3 C4 0 1 1 0

C5 C6 C7 C8 0 1 1 1


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


4 Mapping GFP frames into PDH channels (single and virtually concatenated) The encapsulation mechanism currently defined for carrying data frames in these virtually concatenated DS1, E1, DS3, and E3 signals is the Generic Framing Procedure (GFP). GFP is presented in the next section of this chapter. The mapping, which is defined in ITU-T Rec. G.8040, is described here.

Figure 4 Examples illustrating the mapping of GFP frames into PDH channels

FFFFF

FF

12345

2324


Fram

e nu

mbe

r

GFP Frame Overhead

a) Octet-aligned mapping for GFP into the DS1 signal

GFP Frame Overhead

X1X2P1P2M1M2M3

F1 F2C11C21C31C41C51C61C71

F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4F1 F2 F3 F4

C12C22C32C42C52C62C72

C13C23C33C43C53C63C73

680 bit subframe (84 payload bytes)

84 bits

7subframemaster-frame

CONCATENATION OVERHEAD OCTET

b) Octet-aligned mapping for GFP into the DS3 signal


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


When GFP frames are mapped into DS1, DS3, E1, or E3 signals, the GFP frame octets are mapped octet-by-octet into the signal payload container octets. DS1 and E1 signals were originally designed to have octet-structured payload containers, and an octet-oriented payload container structure was added to the E3 signals in ITU-T Rec. G.832. In the case of DS3 signals, the payload structure was originally bit-structured. When ATM cell mappings were defined for DS3 payloads, they exploited the fact that there are 84 payload bits between DS3 framing bits and defined a nibble-wise mapping with the ATM bytes divided into 4-bit nibbles and 21 nibbles mapped into the 84 payload bits. The GFP mapping took advantage of there being an even number of nibbles per subframe and chose an octet-oriented structure for the payload mapping in order to be consistent with the octet-oriented concatenation overhead channel6. These mappings are illustrated in Figure 4-9 for the DS1 and DS3 signals.

5 Conclusions Due to their continuing nearly universal availability, PDH networks will continue to play an important role as carriers roll out new data services. This will be especially true in North America as long as DS1s and DS3s have a cost advantage due to unbundling of tariffs in the access networks. The new ITU-T Rec. G.8040 provides an efficient, robust GFP-based mapping into PDH signals and the new ITU-T Rec. G.7043 provides the virtual concatenation capability with LCAS to flexibly choose and adjust the channel size in a PDH network. The combination of these two technology enhancements to PDH networks provides carriers with powerful tools to offer new, higher-rate Ethernet connectivity services while continuing to derive benefit from their existing PDH infrastructures. Providing new services, and maximizing the return on existing capital investment, provide compelling business drivers for carriers to invest in best-in-class implementations of G.8040 and G.7043 technologies.

6 The nibble-oriented mapping was originally chosen for the GFP mapping into a DS3 signal. Since virtual concatenation is less complex to implement with an octet-oriented structure than with the nibble-oriented structure, the mapping switched to octet-orientation after virtual concatenation was defined for DS3s. .


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM


6 References [1] M. Elanti, S. Gorshe, L. Raman, and W. Grover, Next Generation Transport Networks – Data,

Management, and Control Plane Technologies, Springer, 2005.

[2] ITU-T Recommendation G.7043 (2004), Virtual Concatenation of Plesiochronous Digital Hierarchy (PDH) signals (S. Gorshe – technical editor)

[3] ITU-T Recommendation G.7042/Y.1305 (2001), Link Capacity Adjustment Scheme for Virtual Concatenated Signals

[4] ITU-T Recommendation G.8040 (2004), GFP frame mapping into Plesiochronous Digital Hierarchy (PDH) (S. Gorshe – technical editor)

[5] S. Gorshe, A Tutorial on SONET/SDH Technology White Paper, PMC-Sierra, PMC-2030895.

[6] T1.105-2001 Synchronous Optical Network (SONET) Basic Description including Multiplex Structure, Rates, and Formats (S. Gorshe – technical editor)

[7] ITU-T Recommendation G.707 (1996), Synchronous Digital Hierarchy Bit Rates.

[8] ITU-T Recommendation G.7041/Y.1303 (2001), Generic Framing Procedure. (S. Gorshe – technical editor)

[9] S. Gorshe, Generic Framing Procedure (GFP), PMC-Sierra, PMC-2041083.

[10] IETF RFC 1990, The PPP Multilink Protocol (MP). K. Sklower, B. Lloyd, G. McGregor, D. Carr, T. Coradetti. August 1996.

[11] Interoperability Requirements for Nx56/64 kbit/s Calls, version 1.0, from the Bandwidth ON Demand INteroperability Group (BONDING) Consortium, 1992

.


Downl

oade

d [c

ontro

lled]

by

Anoua

r L o

f No

on T

hurs

day,

15

April,

201

0 0

2:11

:40

PM

Resilient Packet Ring (RPR) Technology White Paper

7 Notes

14

Head Office: PMC-Sierra, Inc. #105 - 8555 Baxter Place Burnaby, B.C. V5A 4V7 Canada

To order documentation, send email to: [email protected] or contact the head office, Attn: Document Coordinator

All product documentation is available on our web site at: http://www.pmc-sierra.com/ http://www.pmc-sierra.com/processors/ http://www.pmc-sierra.com/serdes/

© Copyright PMC-Sierra, Inc. 2005.

PMC-2050380 (1.0)

Tel: 604.415.6000 Fax: 604.415.6200

For corporate information, send email to: mailto:[email protected]

All rights reserved.

Documents

Ethernet Transport Over PDH Networks With Virtual Concatenation Tutorial