CCAMP Working Group Internet Draft Jun Kyun Choi Document: draft-choi-burst-control-00.txt Min Ho Kang Expiration Date: December 2003 Jung Yul Choi ICU Tae-Gon Noh Kyoo Ryon Hahm Samsung AIT June 2003 Requirements of Burst-Level Control in Optical Networks Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC-2026. 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 obsolete 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/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Abstract This draft presents the requirements of burst-level control in order to improve channel efficiency in optical networks. By processing and control the finer granularity of burst data, the network can improve its utilization than the present circuit switching based optical networks. Implementation issues of transmission and switching for data burst are took into consideration. Conventions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC-2119. Choi, et. al. Expires - December 2003 [Page 1] draft-choi-burst-control-00.txt June 2003 Table of Contents 1. Introduction.....................................................2 2. Needs of burst-level control.....................................3 3. Consideration Issues for burst-level control.....................4 4. Enabling technologies for burst-level control....................5 4.1 Classification of data burst switching..........................5 4.2 Switch control for data burst switching.........................5 4.3 Signaling control for data burst switching......................6 4.3.1 Label for identifying data burst..............................6 4.3.2 Signaling messages for data burst switching...................6 5. Conclusion.......................................................7 6. Security Considerations..........................................7 References..........................................................7 Acknowledgement.....................................................8 Author's Addresses..................................................8 Full Copyright Statement............................................9 1. Introduction Through the emergence of wavelength division multiplexing (WDM) technology, optical networks could provide the explosively increasing transmission capacity. It enables several Terabit/sec transmission capacity in a single fiber which contains several tens or hundreds of wavelength with multiplexing [1]. Nevertheless of the huge transmission capacity of fibers, the transmission efficiency comes to decrease due to diverse applications traffic from access networks and resulting bursty characteristics of Internet traffic in a backbone network. Therefore, the needs for new efficient transmission and switching technology for bursty Internet traffic efficiently and accommodation of many subscribers are required. First of all, we will discuss the main research streams for pursuing implementation of optical networks. First, wavelength routing functions to establish a lightpath which is an all-optical data path along which data does not need to go through any O/E/O conversion before data can be sent [2]. Such lightpaths can provide a high-speed, high-bandwidth pipe that is transparent to bit rate and coding format. Some end users will require one or more lightpaths only for a relatively short period (e.g. from minutes to weeks). WDM layer MAY need to establish and tear down wavelength-routed lightpahts dynamically. One of the limitations of wavelength routing is that is inefficient for Internet traffic which is self-similar (or bursty at all time scales). Due to bursty traffic, the bandwidth utilization of a lightpath is very poor. Even though lightpaths are established dynamically, the set-up time of a lightpath based on two-way Choi, et. al. Expires - December 2003 [Page 2] draft-choi-burst-control-00.txt June 2003 reservation may be too long for a burst which contains small size of data given the high transmission rate. Second, optical packet switching, which processes data traffic in all optical domain, is recognized as an ultimate solution for transmission and switching technology in optical networks [3]. However, the doubt of possibility of implementation of optical packet switching raises due to lack of optical buffer, difficulty of header processing, and synchronization in optical domain. Because of the above mentioned inefficiency of the present switching technology and difficulty in implementation of future switching technology, we need a new transmission and switching technology as an implementable intermediate solution. Burst-level control is an emerging technology to provide finer granularity as optical packet switching and alleviate difficulty of implementation [5-9]. Burst- level control means that a network deals with burst data for transmission and switching in order to improve network utilization. It makes it possible interleaving several data burst into a single link. This draft introduces the requirements of transmission and switching per data burst in order to improve channel efficiency in bursty Internet traffic environment. Burst-level processing will be specified with the advantages, the operational mechanisms of it as well as in the aspect of switch control and signaling control. 2. Needs of burst-level control The circuit switching based wavelength routing is recognized to be not suitable for the present Internet where traffic characteristic is bursty and self-similar. Thus, even though transmission capacity is increased with WDM technology, the Internet can not process the huge burst traffic at an instant. Therefore, a new transmission and switching technology is required to deal with the bursty Internet traffic. To deliver bursty data traffic efficiently means to increase channel efficiency. So far, researches on improvement of channel efficiency have been progressed in time domain and frequency domain [1,3,5-9]. In frequency domain, wavelength division multiplexing is one of the main results [1]. In time domain, optical packet switching, which processes data packet and delivers them in all optical domain, is the key issue for pursuing the ultimate solution for optical internet [3]. However, due to the limitation of optical technology such as optical buffer and header processing, optical packet switching is recognized as a future technology. The fast circuit switching has been studied to overcome the overhead of connection establishment and release of the legacy circuit switching based on three-way-handshaking [4]. In fast circuit switching, connection setup (or tear down) takes place when the start (or end) of a burst is detected by sending a control signal, and is fast since routing has already been done. Even though it tries to increase channel efficiency by reducing connection Choi, et. al. Expires - December 2003 [Page 3] draft-choi-burst-control-00.txt June 2003 establishment overhead by one-way reservation without receiving connection setup acknowledgment, its efficiency decreases when it delivers short burst traffic. As a solution of the above problems in the present Internet, a new transmission and switching technology suitable for bursty Internet traffic is required. For the purpose, optical burst transmission and switching was suggested and they have been studied [5-9]. The main feature of optical burst switching is the elimination of the round- trip waiting time before data burst is transmitted: the switching fabric inside the switches are configured for the incoming data burst as soon as the first control packet announcing the burst is received. Since the switching fabric already has been configured before data burst comes in the switch, the data burst will cut through the switch, even without buffering themselves. Another feature of optical burst switching is to transmit and switch per each data burst, contrary to circuit switching which establishes a full connection for data transmission. Once a connection has established in circuit switching, it is not possible to interleave other data into the connection even if empty space in the connection. However, in optical burst switching, other data burst can be inserted into empty space among different data burst in a link. By applying burst-level control, we can get the following advantages. - Multiplexing of data burst in a link - Dynamic provisioning for data burst - Short transmission delay owing to one-way reservation - Improvement of link utilization 3. Consideration Issues for burst-level control The following items should be taken into account for implementation of transmission and switching per data burst. - One-way-reservation for fast transmission. - Multiplexing of data burst in a link - Switch control mechanisms for setting the switch fabric to switch and transmit short bursty data traffic. - Signaling mechanisms to support dynamic provisioning for short data burst. - Route reuse scheme to use the existing established path not for calculate and setup new path for burst data. - Definition of new label to identify data burst. - Definition of service model to support data burst. Based on the above mentioned issues, the burst-level control will be developed to support burst data in all optical domain. First of all, a new label which identifies data burst SHOULD be defined. Switch Choi, et. al. Expires - December 2003 [Page 4] draft-choi-burst-control-00.txt June 2003 control and signaling mechanism also SHOULD be developed to support data burst and systems with the burst-specific characteristics. 4. Enabling technologies for burst-level control In chapter 4, classification of data burst switching according to the reservation time and types which should be supported are presented. Switch control to support them and signaling protocol to support end to end burst data transmission and switching will be discussed. 4.1 Classification of data burst switching Data burst switching can be categorized as the following types based on the beginning and releasing time of resource reservation for data burst transmission. According to how soon before the burst arrival and how soon after its departure, the switching elements are made available to route other data bursts. Explicit setup - Explicit release; After a switch receive the setup request for data burst the switch fabric is configured for the incoming data burst immediately and remains in that configuration until a release request arrives. Explicit setup - Implicit release; The setup request contains information about the length of the data burst so that a release request is not needed to indicate the end of the data burst. Implicit setup - Explicit release; The start of data burst is estimated based on information contained in a setup request while a release request notifies the switching resource to be released. Implicit setup - Implicit release; Both the start and end time of data burst are predicted based on information contained in a setup request. In the above four types, when we estimate the start and end of the burst tightly with implicit notifications, there is smaller overhead of keeping the switching resources configured and lower blocking probability in the network. However, there is a tradeoff between utilization and complexity of control. 4.2 Switch control for data burst switching Optical switch is required to establish and release a connection for data burst in order to transmit data burst fast without optical- electrical-optical conversion. For the purpose, General Switch Choi, et. al. Expires - December 2003 [Page 5] draft-choi-burst-control-00.txt June 2003 Management Protocol (GSMP) working group (WG) is working on standardization about switch control for optical switching per data burst as well as wavelength and fiber [10-11]. The work uses the Reservation Request message in order to reserve switching resource for data burst efficiently. According to different reservation mechanisms for data burst transmission mentioned in section 4.1, the switch controller tries to reserve the resource for the transmission with the message. That is, by setting the start time and burst length information contained in the message appropriately, resource on the switch is reserved and activated for the enough bandwidth. The other related to switch control follows GSMP WG documents [10-11]. 4.3 Signaling control for data burst switching Since the length of data burst is assumed to be short, one-way- reservation scheme is recommended to reserve resource for data burst under the assumption that the out-of-band control channel is reliable to deliver control message without transmission error or blocking. The setup request to reserve resource for data burst is delivered from an ingress node to an egress node. The intermediate nodes, which receive the setup request, reserve the switching resource according to information contained the request for data burst without sending back acknowledge about the setup request. At this time, the switch controller configures the switch element for the data burst switching as well as reserves the switching resource. If there is inefficient resource and no available output port the controller sends the failure of the setup request to the ingress node. Generalized Multi-Protocol Label Switching (GMPLS) signaling protocol, such as RSVP-TE or CR-LDP, can be used as a signaling protocol to transmit and switch data burst with some modification or extension. GMPLS supports multiple types of switching: Packet, Layer-2, TDM, Lambda, and Fiber level switch capable [12-13]. For delivery of data burst, GMPLS SHOULD support optical burst switch capable interface. Since the existing signaling protocols use two-way reservation for establishing label switched path (LSP), the protocols SHOULD be modified to support one-way reservation for data burst. A new identifier is required to identify data burst as a new switching unit in the signal protocol. For the purpose, "Label for optical burst" SHOULD be defined to identify data burst. The label is used to identify data burst of which granularity is coarser than packet (PSC) and finer than wavelength (LSC). 4.3.1 Label for identifying data burst This label is used to identify data burst. The format and its semantic is TBD. 4.3.2 Signaling messages for data burst switching Choi, et. al. Expires - December 2003 [Page 6] draft-choi-burst-control-00.txt June 2003 Signaling messages are used to establish a connection for data burst in one-way reservation mechanism, not like the existing signaling mechanisms which is based on two-way reservation. Therefore, the existing signaling mechanisms SHOULD be modified to fast transmit and switch data burst. The candidate protocols are CR-LDP and RSVP-TE. The format and semantic of signaling message is TBD. 5. Conclusion This draft specifies the requirements and basic technology for burst- level control in order to improvement channel efficiency in optical networks. For achieving higher utilization of channel, data burst with a finer granularity SHOULD be processed in all optical domain. Burst-level control can be implemented with GSMP for switch control and GMPLS for signaling protocol with some modification and extension. 6. Security Considerations This document does not have any security concerns. The security requirements using this document are described in the referenced documents. References [1] C.A. Bracket, "Dense Wavelength Division Multiplexing Networks, Principles and Applications", IEEE JSAC, August, 1990. [2] Zang, H., Jue, J. P., Mukherjee, B., "A Review of routing and wavelength assignment approaches for wavelength routed optical WDM networks", Optical Networks Magazine, 2000 [3] D. J. Blumenthal, et. at., "Photonic packet switches: Architectures and experimental implementations", Proceedings of the IEEE, Nov., 1994 [4] Christer, B., Markus, H., Per, L., Lars, R., Peter, S., "Fast Circuit Switching for the Next Generation of High Performance Networks", JSAC, 1996 [5] C. Qiao, M. Yoo, "Choice, and Feature and Issues in Optical Burst Switching", Optical Net. Mag., vol.1, No.2, April 2000, pp.36-44. [6] C. Qiao, M. Yoo, "Optical Burst Switching (OBS) - a new paradigm for an optical Internet", Journal of High Speed Networks, 1999 Choi, et. al. Expires - December 2003 [Page 7] draft-choi-burst-control-00.txt June 2003 [7] J. Turner, "Terabit burst switcing", Journal of High Speed Networks, 1999 [8] Ilia Baldine, George N. Rouskas, Harry G. Perros, Dan Stevension, "JumpStart: A Just-in-time Signaling Architecture for WDM Burst- Switching Networks", IEEE Comm. Mag., Feb. 2002. [9] Illia Baldine, et. al., "JumpStart: A Just-in-Time Signaling Architecture for WDM Burst-Switched Networks", IEEE Comm. Mag., Feb., 2002 [10] Doria, A, "GSMPv3 Base Specification", draft-ieft-gsmp-v3-base- spec-02.txt, June 2003. [11] Junkyun Choi, JungYul Choi, et. al., "General Switch Management Protocol (GSMP) v3 for Optical Support", draft-ietf-gsmp-optical- 02.txt (work in progress), June 2003. [12] Mannie, E., et. al., "Generalized Multi-Protocol Label Switching (GMPLS) Architecture", draft-ietf-ccamp-gmpls-architecture-03.txt (work in progress), August 2002. [13] Ashwood-Smith, D., et. al., "Generalized MPLS - Signaling Functional Description", RFC3471, Jan. 2003. Acknowledgement This work was supported in part by the Korean Science and Engineering Foundation (KOSEF) through OIRC project Author's Addresses Jun Kyun Choi Information and Communications University (ICU) 58-4 Hwa Ahm Dong, Yusong, Daejon Korea 305-732 Phone: +82-42-866-6122 Email: jkchoi@icu.ac.kr Min Ho Kang Information and Communications University (ICU) 58-4 Hwa Ahm Dong, Yusong, Daejon Korea 305-732 Phone: +82-42-866-6136 Email: mhkang@icu.ac.kr Jung Yul Choi Information and Communications University (ICU) Choi, et. al. Expires - December 2003 [Page 8] draft-choi-burst-control-00.txt June 2003 58-4 Hwa Ahm Dong, Yusong, Daejon Korea 305-732 Phone: +82-42-866-6208 Email: passjay@icu.ac.kr Tae-Gon Noh Samsung Advanced Institute of Technology P.O. Box 111, Suwon, Kyoungki Korea, 440-600 Phone: +82-31-280-9621 Email: tgnoh@samsung.com Kyoo Ryon Hahm Samsung Advanced Institute of Technology P.O. Box 111, Suwon, Kyoungki Korea, 440-600 Phone: +82-31-280-9549 Email: ryonhahm@samsung.com Full Copyright Statement Copyright (C) The Internet Society (2002). All Rights Reserved. This document and translations of it MAY be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation MAY be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself MAY not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process MUST be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Choi, et. al. Expires - December 2003 [Page 9]