Network Working Group Rajendra Damle Internet Draft Young Lee Expiration Date: December 2001 Iris Labs Eric Brendel Coree Networks Riad Hartani Caspian Networks Vishal Sharma Metanoia June 2001 Optical Channel Concatenation -- Need and Requirements draft-damle-optical-channel-concatenation-00.txt Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/1id-abstracts,html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Abstract This contribution identifies the need and requirements for concatenating optical channels to create multiple high bandwidth payload channels. To maximize the benefits of standardization, the concatenation methodolgy should be independant of framing protocols e.g., SONET, GFP, OCh etc.) as well as payload types (e.g., Packet, Cell, Byte Stream). A contribution for the need to concatenate optical channels was made at the T1X1.5 meeting in March 2001 and was accepted for further proposals. This contribution is based on the T1X1.5 contribution and is presented herein for information only. Damle, et. al. [Page 1] Internet Draft draft-damle-optical-channel-concatenation-00.txt Jul 2001 1. Introduction Historically, port speeds on routers and switches used in the backbone network were lower than the bandwidths used and needed for efficient transport over optical fibers. This required devices such as SONET Add Drop Multiplexers (ADMs) to multiplex multiple lower data rate sources to higher data rates for transport. Today the bandwidth requirements between the core routers/switches have increased from 2.5 Gb/s (OC-48) to 10 Gb/s (OC-192) in some cases. Routers with OC-48 or OC-192 capable port cards and a transport system that can carry that data rate over a single wavelength have supported these increases. In the future, data bandwidth demands between core switch sites are projected to quickly grow to 100s of Gb/s. To meet this demand service providers have to deploy multiple routers per switch site. This results in using multiple ports at 2.5 Gb/s or 10 Gb/s for inter router connections within a site as well as between sites [Figure 1]. ----------------------------- --------------------------- WDM Channels | | | | | router/switch v mux | Regen |demux ____ ____ ____ ___ |\ | Site | /| ___ ____ ____ ____ | || || |->| |->| | | ---- | | |->| |-| || || | ---- | |---- --- | |_ |_|\_| |_|\_|_| | --- ---- | |---- ____ | |____ ___ | | | |/ | | |/ | | | ___ ____ | |____ | || || |->| |->| | | ---- | | |->| |-| || || | ---- | |---- --- | | | optical | | | --- ---- | |---- | X | ->|/ |amplifier | \|-- | X | : | | : | | | | : | | : : | | : \ | | | / | : | | : ____ | |____ \_|____________________|__/ | ____ | |____ | || || | | | WDM System | | | || || | ---- ---- ---- | | | | ---- ---- ---- | | | | | | | | router/switch | | | | ____ ____ ____ | | | | ____ ____ ____ | || || |------| | | | | || || | ---- | |---- | | | ---- | |---- ____ | |____ | | | ____ | |____ | || || | | | ----->| || || | ---- | |---- <--- router/ | | ---- | |---- | X | switch port| | | X | : | | : | | : | | : : | | : | | : | | : ____ | |____ | | ____ | |____ | || || | | | | || || | ---- ---- ---- | | ---- ---- ---- | | ------------------------------ --------------------------- Switch Site A Switch Site B Figure 1: Multiple router ports connected across multiple WDM channels Damle, et. al. [Page 2] Internet Draft draft-damle-optical-channel-concatenation-00.txt Jul 2001 To eliminate inter router connections within a site, routers are scaling up capacity per router by developing switch fabric that can support several thousands of ports at 2.5 Gb/s or 10 Gb/s. Routers with large numbers of low-speed ports make the network difficult to manage and hence expensive. 2. Discussion For carriers to deploy manageable and stable data networks that meet the projected growth in bandwidth demands, new requirements on router designs and WDM transmission system designs are emerging. Switch Fabric Port Speed /\ || || ------------ | 100 Gbps | ------------ || || WDM Transponder || Date Rate Per Lambda || /\ || || ------------ ------------ | 10 Gbps | | 10 Gbps | ------------ ------------ || || || || ------------ ------------ | 2.5 Gbps | | 2.5 Gbps | ------------ ------------ || || > 1000 km w/o regeneration Router/Switch --- ____ ------> ____ --- ----- ------- ----- __\ | |--|>|___|--|\ /|--|___|->| | | || || | \| | | ____ | | | | ____ | |_\ | || || |__ /| |--|>|___|--| | | |--|___|->| | \ ----- | |----- / | | | ____ | |--|\--|\--| | ____ | |_ / | | ^ | |--|>|___|--| | |/ |/ | |--|___|->| | / : | | : | | | | ____ | | | | ____ | | : | X | : | | |--|>|___|--|/ \|--|___|->| | : | | : | --- --- ----- | |----- | ^ | || || | | |__ concatenated channels | || || | | ----- ------- ----- |___________ High Bandwidth Channels Figure 2: Need for transparently concatenating multiple optical channels Damle, et. al. [Page 3] Internet Draft draft-damle-optical-channel-concatenation-00.txt Jul 2001 The router designs have to change from large number of 2.5 Gb/s or 10 Gb/s switch fabric ports to a small number of 40 Gb/s and higher capacity ports. Advances in semiconductor technology and high speed packet processing are key to enabling high speed router ports at 40 Gb/s and beyond. The semiconductor technology available today is already capable of supporting aggregates of more than 40 Gb/s throughput per port. In transmission systems, the WDM systems requirements are changing to ultra long reach between terminals in order to make transport as economical as possible. The ultra long reach requirement over existing fiber plant is limiting the maximum data rate per wavelength that can be transported economically. Economical ways to mitigate fiber non- linearities, advances in optical amplification, laser modulation and optical mux/demux technologies will enable higher data rates per lambda. The semiconductor technology is already enabling router/switch fabric ports at 40 Gb/s and greater. However deployable transmission systems are far from being ready to transport 40 Gb/s and greater per wavelength over ultra long reach distances. Therefore there is a clear need to concatenate lower speed WDM optical channels (sub-channels) that use any framing protocol to form one or more higher bandwidth interfaces (super channels) to the routers/switches. In the future semiconductor technologies are expected to develop faster than high speed optical transmission technologies creating a sustained need for concatenating multiple optical channels to create high bandwidth channels. Standardizing a methodology to concatenate multiple optical channels that is agnostic to the transport framing protocols and payload types will have a significant impact on the both the equipment manufacturers and the carriers. The standardized methodology will completely de-couple the router/switch equipment from the transmission equipment. This de-coupling will allow the development of new equipment in both domains such that carriers can extract huge savings by deploying large but manageable routers/switches directly over a cost effective ultra long haul WDM system. 3. Requirements There are proposals at T1X1 and ITU [1]-[3] to virtually concatenate VTs and STSs to make efficient use of SONET based transport for bursty traffic. However, they do not cover the emerging applications and requirements described in this document. These requirements for concatenating optical channels to enable a carrier grade backbone data network are as follows: A. In order for the concatenation to be truly transparent today and in the future, it should be agnostic to: - Payload types: work with Cell/Packet/TDM or byte stream as the input. - Framing protocols used on the optical channels (sub-channels) - capability to concatenate optical channels that use any of the standardized framing protocols (SONET, OCh, GFP etc.) into one or more higher bandwidth super channels. Damle, et. al. [Page 4] Internet Draft draft-damle-optical-channel-concatenation-00.txt Jul 2001 B. The concatenation should be truly scalable by being independent of: - Payload data rates - Transmission data rates - capability to concatenate a set of channels at any data rate as long as the data rate is the same within a set. Variation in optical channel data rate within a concatenated set is unlikely in a real network hence the added complexity to accommodate a set of variable data rate channels would not be necessary C. Keep the overhead to the minimum by optimizing for point-to-point network topology since the data traffic demands are essentially point-to-point. Low overhead also ensures that there is minimal penalty for the concatenation function. D. Should have all the carrier class survivability features required to make the high bandwidth channels ultra reliable across the long haul transmission system and have graceful degradation. We believe the following features support carrier class survivability: - Capability to uniquely identify a concatenated channel as well as optical sub-channels contained within the concatenated channel - Capability to monitor degradation per concatenated optical channel through BER monitoring and CRCs - Generalized arrival-time variation compensation - Capability to add/remove sub-channels automatically without disrupting the super channel - Capability to communicate individual optical channel status to the transmit end without the use of separate messages so as to minimize the delay in this critical communication - Capability to de-couple individual optical channel errors from the concatenated superchannels and evenly distribute the available bandwidth amongst all the payload streams - Provide capability for extended burn-in testing of individual sub-channels - Provide hooks to support concatenated channel level protection schemes under the control of a higher layer - Capability to guarantee payload arrival sequence (e.g., packet order) E. Should have the capability to allow the carriers to easily manage and service the concatenated super channel and the sub-channels through automatic as well as manual provisioning features. 5. References [1] T1X1.5/2000-157R1 "A Justification for a Variable Bandwidth Allocation Methodology for SONET Virtually Concatenated SPEs" [2] T1X1.5/2000-156 "A Proposal for Variable Bandwidth Allocation (VBA) Methodology for SONET Virtually Concatenated SPEs" [3] T1X1.5/2000-199 "A Proposed Link Capacity Adjustment Scheme (LCAS) for SONET Virtually Concatenated SPEs" [4] T1X1.5/2001-090 "Need for Concatenating Optical Channels to Create a Transparent High Bandwidth Channels" [5] T1X1.5/2001-103 "Clarification of T1X1.5/2001-090" Damle, et. al. [Page 5] Internet Draft draft-damle-optical-channel-concatenation-00.txt Jul 2001 5. Security Considerations This draft does not introduce any new security issues. 6. Authors' Addresses Rajendra Damle Iris Labs Inc. 101 E. Park Blvd 855 Plano, TX 75025 Phone: 972 943 2963 Email: rdamle@irislabs.com Young Lee Iris Labs Inc. 101 E. Park Blvd 855 Plano, TX 75025 Phone: 972 943 2964 Email: ylee@irislabs.com Eric Brendel Coree Networks 56 Park Road Tinton Falls, NJ 07724 Phone: 732 380 2800 Email: brendel@coreenetworks.com Riad Hartani Caspian Networks 170 Baytech Drive San Jose, CA 95143 Phone: 408 382 5216 Email: riad@caspiannetworks.com Vishal Sharma Metanoia, Inc. 335 Elan Village Lane Unit 203 San Jose, CA 95134-2539 Phone: 408-943-1794 Email: v.sharma@ieee.org Lee, et. al. [Page 6] Internet Draft draft-ylee-optical-channel-concatenation-00.txt July 2001 Expiration Date: January 2002