Internet Draft Kim, et al S. W. Ryu Document: draft-kim-ccamp-gmpls-nsid-00.txt Korea University KT(Korea Telecom) J. K. Choi ICU C. H. Kang Korea University Expires: December 2002 June 2002 A Requirement of the Network State Information Database for Traffic Engineering Over GMPLS Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026 [1]. 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/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Abstract This document presents a set of requirements of the Network State Information Database (NSID) for Traffic Engineering over Generalized Multiprotocol Label Switching (GMPLS). The Network State Information Database is required to implement the network architecture for network models that introduce the control element of IP and to D. G. Kim Expires - December 2002 [Page 1] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 optimize the utilization of network resource. And this document includes discussion about the considerations and necessity of the several attributes to construct NSID for Traffic Engineering over GMPLS that are extended from the requirement for Traffic Engineering over MPLS [4]. These attributes can be used to maximize the utilization of network resources and to enhance resource oriented Traffic Engineering techniques. Conventions used in this document 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 [2]. Table of Contents 1. Introduction...................................................3 1.1 Terminology................................................3 1.2 Document organization......................................3 2. Traffic Engineering over GMPLS.................................4 2.1 Traffic Engineering Performance Objectives in GMPLS........4 3. GMPLS Architecture for Traffic Engineering.....................6 3.1 Network State Monitoring and Analysis Stage................6 3.2 Required Resource Estimation Stage.........................7 3.3 Reconfiguration Decision Stage.............................7 3.4 Logical Topology Design and Modification Stage.............7 3.5 Network Topology Migration Stage...........................7 4. GMPLS Architecture for Traffic Engineering in Overlay Model....8 5. GMPLS Architecture for Traffic Engineering in Integrated Model10 6. Network State Information Database for GMPLS..................10 6.1 Resource Attribute........................................11 6.2 Policy Attribute..........................................11 6.3 Traffic Attribute.........................................12 6.4 Adaptivity Attribute......................................13 6.5 Priority Attribute........................................13 6.6 Preemption Attribute......................................13 6.7 Resilience attribute......................................14 7. Implementation Considerations.................................14 8. Conclusion....................................................15 Security Considerations..........................................15 References.......................................................16 Author's Addresses...............................................17 D. G. Kim Expires - December 2002 [Page 2] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 1. Introduction According to the recent rapidly increasing of IP traffic, Wavelength Division Multiplexing (WDM) technology is rapidly becoming a technology-of-choice to meet the tremendous bandwidth demand of IP traffic. Traffic Engineering (TE) of an IP network is concerned with performance optimization of operational networks and mapping the flow of traffic to present a physical topology of network. A major goal of Internet Traffic Engineering is to facilitate efficient and reliable network operations while simultaneously optimizing network resource utilization and traffic performance [5]. But because the attributes of present IP routing protocols have been poor, the supplementary means for traffic engineering over GMPLS must be prepared. To overcome this problem, two schemes of research for traffic engineering in IP networks have been introduced û one scheme is able to control explicitly the data path and other scheme is the adaptive load balancing of usable path û the results of these research came to produce MPLS technology for traffic engineering in IP networks and a GMPLS technology added control plane of IP for introducing the traffic engineering over MPLS in WDM networks. For introducing the traffic engineering over GMPLS, We can consider two schemes, one is the extended routing protocol for traffic engineering and the other scheme is the allocation of traffic through a network resource database for the efficient utilization of network resource. This document describes the latter scheme that allocates the traffic to the network resource through the construction of the network state information database. In this document we define the network state information database (NSID) that support traffic engineering database(TED) for the function of traffic engineering and management information base (MIB) for the function of network maintenance. The main function of the network state information database is the control and management of network state information data. 1.1 Terminology The reader is assumed to be familiar with the GMPLS terminology as defined in [3][11]. 1.2 Document organization The reminder of this document is organized as follows: Section 2 provides the requirement of Traffic Engineering over GMPLS and section 3 presents an overlay and integrated model in the view of Traffic Engineering. Section 4 presents the overview of the characteristics and requirements of GMPLS architecture required for an extended overlay and integrated model. Section 5 describes the necessity and structural elements of a network state information D. G. Kim Expires - December 2002 [Page 3] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 database and the implementation consideration. Finally, Section 6 contains concluding remarks. 2. Traffic Engineering over GMPLS This section describes the Traffic Engineering objectives in GMPLS and when traffic engineering of MPLS is extended to optical network, the correlation of optical resources and requirements in traffic mapping is presented. 2.1 Traffic Engineering Performance Objectives in GMPLS The key performance objectives associated with traffic engineering in GMPLS can be classified as being either in MPLS [4]: - traffic oriented or - resource oriented Traffic Oriented Performance Objectives include the aspects that enhance the QoS traffic streams. In the present Internet network providing best effort service, the key traffic oriented performance objectives can include: minimization of packet loss, minimization of delay, maximization of throughput, and enforcement of service level agreement. These elements are used in optical networks as well as in the IP network. Resource Oriented Performance Objectives include the aspects pertaining to the optimization of resource utilization. Efficient management of network resources, such that subsets of network resources do not become over utilized and congested while other subnets along alternate feasible paths remain underutilized, is the vehicle for the attainment of Resource Oriented Performance Objectives. Bandwidth is a crucial resource in contemporary networks, but in the consideration of optical networks the key resource oriented performance objectives can include: the number of wavelengths, the number of Optical Crossconnects (OXC), the number of fibers, and the number of optical transceivers. For the enhancement of resource oriented performance congestion control, load balancing can be chosen. The objective of load balancing is to minimize maximum congestion or alternatively to minimize maximum resource utilization, through efficient resource allocation. Also for the maximization of resource performance resource management is needed. For the resource management it is necessary to construct the resource control signal and network state information database that is able to manage the network resource through the control signal. Some key extensions brought by GMPLS to MPLS TE are highlighted in the following [3]. D. G. Kim Expires - December 2002 [Page 4] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 1) In MPLS-TE, links traversed by a Label Switched Path (LSP) can include an intermix of links with heterogeneous label encoding (e.g. links between routers, links between routers and ATM-LSRs, and links between ATM-LSRs). GMPLS extended this by including links where the label is encoded as a timeslot, or wavelength, or fiber. In MPLS-TE, an LSP that carries IP has to start and end on a router. GMPLS extends this by requiring LSP to start and end on several types of Label Switching Routers (LSR) capable of several kinds of labels. Therefore the type of a payload that can be carried in GMPLS by LSP is extended to allow such payloads as SONET/SDH, G.709, 1GbE or 10GbE, etc. The use of Forwarding Adjacencies (FA) that can improve bandwidth utilization is considered. When bandwidth allocation can be performed only in discrete units, FA allows the number of required labels to be reduced. GMPLS allows for a label to be suggested by an upstream node to reduce the setup latency and to limit the range of labels that is selected by the downstream node. This feature is useful in photonic networks where wavelength conversion may not be available. While traditional TE-based (and even LDP-based) LSPs are unidirectional, GMPLS supports the bi-directional LSPs. This feature will be useful in resource management. GMPLS supports the termination of LSP on a specific egress port, i.e. the port selection at the destination node. For TDM, Label Switching Capable (LSC) and Fiber Switching Capable (FSC) interfaces in GMPLS, bandwidth allocation for an LSP can be performed only in indiscrete units. There are much fewer labels on TDM, LSC, FSC than on PSC or L2SC links, because the former are physical labels instead of logical labels. Therefore resource oriented traffic engineering is needed. There are basically three fundamental problems of traffic mapping in MPLS: mapping packets onto forwarding equivalent classes, mapping forwarding equivalent classes onto traffic trunks, and mapping traffic trunks onto the physical network topology through LSP. But in GMPLS including optical resources there are some problems of hierarchical LSP with how to map traffic trunks onto lambda, and how LSP is mapped onto lambda. The hierarchical mapping stage to optical switching is depicted in Figure 1. Packet--> FEC--> Traffic Trunk--> LSP--> Label Stacking | | | | V | -------- > Lamda < ---- | V Waveband | V Fiber D. G. Kim Expires - December 2002 [Page 5] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 FIGURE 1. The hierarchical mapping stage to optical switching 3. GMPLS Architecture for Traffic Engineering In this section, the basic framework for proposed traffic engineering can allocate optical resources on demand and wavelength assignment and network reconfiguration is focused. Main functions of these frameworks consist in the following stages [7]. +---------------------------------------+ + Network State Monitoring & Analysis + +---------------------------------------+ | V +---------------------------------------+ + Required Resource Estimation + +---------------------------------------+ | V +---------------------------------------+ + Reconfiguration Decision + +---------------------------------------+ | V +---------------------------------------+ +Logical Topology Design & Modification + +---------------------------------------+ | V +---------------------------------------+ + Network Topology Modification + +---------------------------------------+ FIGURE 2. The Framework for Traffic Engineering 3.1 Network State Monitoring and Analysis Stage This stage is responsible for collecting traffic statistics from the network elements. Then statistics are analyzed for the traffic engineering and network reconfiguration related decision-making. D. G. Kim Expires - December 2002 [Page 6] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 3.2 Required Resource Estimation Stage This stage estimates the resources, i.e. wavelength, the bandwidth of each wavelength, in the near future based on past and present measurement and collection and analysis of traffic characteristics. 3.3 Reconfiguration Decision Stage This stage consists of a series of operations that decides when a network level reconfiguration is performed for the resource allocation from the resource estimation stage. This decision element is based on traffic conditions, the number of network resources, and other operational issues, e.g., suppressing the influence of transitional factors and reserving adequate time for network to converge. Network reconfiguration can be triggered when any element value exceeds the thresholds. 3.4 Logical Topology Design and Modification Stage This stage computes a network topology based on the measurement and predictions about the traffic state and network resource. This can be considered as optimizing a layered graph (i.e. IP routers connected by lights in the WDM layers) for specific objectives (e.g. minimum hops and maximizing throughput), subject to certain constraints (e.g. nodal degree, interface capacity), for a given load matrix (i.e. traffic load applied to the network), which in general is a NP- complete problem. Since network reconfiguration can be triggered by periodically changing traffic pattern, a practical approach is to develop methods that focus on cost-effective, fast convergence, and/or minimal impacts on ongoing traffic in the view of global optimality. There are two methods considered. One method is that sub-TLV that can become aware of the optical characteristics added to the present IP routing protocol and the other method is that signaling protocol of IP network is introduced to the optical network so that network topology is changed. 3.5 Network Topology Migration Stage This stage consists of algorithms to schedule the network migration from old topology to a new topology. Even if WDM layer resources are sufficient to support the allocation of resource and the change of network resource, there are still other issues concerning the migration. For example, as WDM reconfiguration deals with large- D. G. Kim Expires - December 2002 [Page 7] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 capacity channels, changing allocation of resources in this coarse granularity has significant effects on a large number of end user traffic. The Procedure of reconfiguration resource for topology migrations is rising to critical issues. In general a migration procedure consists of a sequence of establishing and taking down individual WDM light paths. Traffic flows have to adapt to the light path changes during and after each migration stage. The algorithm is needed so that network configuration gives minimal change on present network state. 4. GMPLS Architecture for Traffic Engineering in Overlay Model In the overlay model traffic engineering over MPLS is extended to GMPLS using the system architecture based IP control [9]. In the MPLS network, the router is required to add sub-TLV related optical attributes to traditional IGP routing protocol to perceive and control the state of optical resources [12]. The Network State Information Database (NSID) is required to control and manage the information of network topology and resources in the optical network. The update routing protocol manages the whole state of IP and optical networks and passes the change of state information to Traffic Engineering Database (TED) of MPLS [8]. +---------------------------------------------+ | +-----------+ +---------+ +-------+ | | | IGP Route | |LSP Path | |Router | | | | Selection |=> |selection| +-------+ | | | Module | | +-----+ | +-----------+ | | +-----------+ | | TED | |=> | signaling | | | ^ | +-----+ | | Module | | | | +----^----+ +-----------+ | | | | | | +----------------+-------+ | <==== | IS/IS & OSPF Routing |=================> | +------------------------+ | +--------------^------------------------------+ | | | [OXC] +--------------v---------------------------------------------------+ | +-------------------------------------------+ Control | | WDM NC & M Module |<==========> | | +-------------------------------------------+ | D. G. Kim Expires - December 2002 [Page 8] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 | Resource status Info ^ ^ ^ | | . . . | | . . connection information | | . . | | +--------------------------+ . . | | Resource control Module | . +------------+ | + | +-----------+ | . | LP | Data | | | | NSID | |<===========> | Connection |<=======>| | | +-----------+ | . | Module | | | +--------------------------+ . +------------+ | | ^ +---------------+ ^ | | Fault Notification | | Protection | | | | +---> | & Restoration |<--+ | | | Module | Fault Detection | | +---------------+ | +-------------------------------------------------------------------+ FIGURE 3.The Architecture of traffic Engineering in Overlay Model According to the architecture of GMPLS in the overlay model, the optical network consists of four modules as followings: - WDM Network Control and Management Module inform the state of the optical network to the routing protocol of MPLS and manage the optical network by using the control signal. - Resource Control Module manages optical resources and topology information for traffic engineering. - Light Path Connection Module establishes and releases the light path, selects the optical fiber. - The Protection and Restoration Module provides fast restoration through the protection and restoration algorithms. And the Resource Control Module has the function of managing limited resources attributes and network topology by using the network state information database that is created by the Routing and Wavelength Assignment (RWA) algorithm and control signal and the function of informing the change of optical resources to the MPLS domain through the WDM Network Control and Management Module. Also the Resource Control Module informs the available optical resources to Light Path Connection Module so that the table of light path can be maintained. In the optical domain the WDM Network Control and Management Module takes charge of establishing and releasing the light path from the Light Path Connection Module by using the control signal and informs the values of change related to resource information to the Resource D. G. Kim Expires - December 2002 [Page 9] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 Control Module. The WDM Network Control and Management Module reports the faults to the Protection and Restoration Module so that the restoration algorithm is active and the Protection and Restoration Module reports the fault information to the Resource Control Module so that new resources are allocated. 5. GMPLS Architecture for Traffic Engineering in Integrated Model The architecture of an integrated GMPLS model has an IP addressed system for traffic engineering and a lambda router is needed to completely support the network control and traffic management of the MPLS and optical network [9]. Also it is necessary to introduce the Wavelength-MIB to consist of information about not only fibers but also wavelength and wavelength continuity/interchange constraints of each node. And for the Wavelength-MIB network control consists of three modules [7]: wavelength routing, wavelength signaling, and wavelength access control. The wavelength routing module needs an update link-state protocol with suitable optical layer extensions. The wavelength signaling module fulfills wavelength routing decisions achieved by the wavelength routing module to perform wavelength assignment, setup/teardown optical light path, priority arbitration with preemption, and adaptive QoS with QoS negotiation. The implementation approach of wavelength routing needs to be optical layer extended RSVP or CR-LDP. The wavelength access control module manages the physical connection between the MPLS router and lambda router, and map MPLS label to wavelength. For traffic engineering, each system needs to be equipped with migration scheduling of network topology, reconfiguration algorithm, and statistics collection and analysis. Also the network state information database should be constructed to manage the integrated domain of IP and WDM for maintenance of resources change. For traffic engineering in the integrated model, the construction of centralized traffic engineering needs to manage wavelength-MIB in the optical layer and IP-MIB in the IP layer to strictly manage traffic and network control. 6. Network State Information Database for GMPLS The network state information database is needed for the control and management of the MPLS and optical network. The network state information database is constructed and maintained based on the traffic-engineering model and application method described in section 4. For example, in the integrated model an entire integrated network state information database is collected from each site and is centralized; in the overlay model the MPLS traffic engineering database and optical network state information database are stored, maintained, and managed separately. The network state information database is involved with two aspects: resources and their usage. Conventional representation of network resources can be simply just the topological information. However, D. G. Kim Expires - December 2002 [Page 10] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 traffic engineering requires more information, e.g. total bandwidth, and current usage on each link. Traffic engineering in an optical layer is interested in not only the utilization status of network resources but also the optical characteristics of wavelength connections and signal quality. When overlay traffic engineering is attempted, the objectives functions at different layers may even be different. In the case of integrated traffic engineering, traffic control and resource allocation are considered together so that optimization objectives must be coordinated. Although different traffic engineering models require different design and implementation of network state information database, many common attributes are shared in both cases as discussed in this section. The attributes of MPLS traffic engineering are extended to include the attributes of an optical network. Requirement of common attributes and extended attributes are classified and elements of each attribute and application scheme are observed. 6.1 Resource Attribute When the resource attribute for state information of network topology is extended to the optical network, optical resources can be included in this attribute as followings [10]: - end to end BER - the number of wavebands and wavelengths per link - the number of LSP, traffic trunk per wavelength - wavelength, LSP coloring - bandwidth of link, wavelength, and LSP and their usage - min, max of current usage related to link, wavelength, LSP - the number of optical transceivers in optical system In the view of traffic engineering, the optimal allocation of physical and logical optical resources is an important element in determining the performance of the network and this information about resources is fundamental to the network state information database. Also Resource Class Attributes can be used to specify the class of resources and depend on the quality of the optical resources û for example, the quality of the light path, end-to-end BER attributes can be an element to determine the quality of optical resources- which become an important requirement for differentiated services and these are applied differently according to policy attributes. 6.2 Policy Attribute The Policy attribute provides differentiated services and this attribute uses a resource class attribute. The Policy attribute can have the following attributes. - Allowable hops of end to end D. G. Kim Expires - December 2002 [Page 11] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 - the number of resources offered - protection and restoration mechanism for network survivability - optical signal noise ratio (OSNR), bit error rate (BER) - network provisioning For example, if the ranges of services according to resource grade present Si, Tj, Uk, those ranges can be presented in the expressions below: S = { S1, S2, S3, ... Si} T = { T1, T2, T3, ... Tj} U = { U1, U2, U3, ... Uk} Ranges of services and same grade of services offered are classified according to the expressions listed above and differentiated services can be provided according to the network policy and optical characteristics. For example, through the limitation of end-to-end hops network management can be done. Services included in Si allow a maximum of 30 hops; Services in Tj allow a maximum of 20 hops so that the limitation of hops can be operated as an element of network management. The number of available wavelengths per link is applied for network management. For example, the number of wavelengths per link included in Si is 10; the number of wavelengths per link included in Tj is 4. According to the range of services, the finite number of wavelengths can be limited and applied to the network management. This scheme that the service range depends on service grade establishes the parameter for traffic engineering and is capable of implementation. As the provisioning that depends on the service range can construct to system, differentiated services can be provided according to the resource grade attribute of the optical network. 6.3 Traffic Attribute The Traffic Attribute is the element that captures the characteristics of the traffic streams. In MPLS the bandwidth of LSP, maximum throughput allowed, and minimum data rate guaranteed are included in classification of attributes [4]. If this definition is extended to the optical network, traffic attributes can be extended to physical and logical values of wavelength, waveband, and fiber. There are important traffic attributes as follows: - the characteristics of fiber: fiber type(single mode, or multimode), maximum transmission distance per fiber - the number of wavelengths per fiber and existence of fiber or not - the number of lambda channels included in a waveband - the bandwidth of a lambda channel - the maximum or minimum of data rate D. G. Kim Expires - December 2002 [Page 12] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 These traffic attributes include common elements of resource attributes and utilize the characteristics of resources. 6.4 Adaptivity Attribute The states and resources of network change over time. These changes result from the availability of new resources and reactivation of failed resources, and de-allocation of allocated resources. In the view of traffic engineering, an administrative function that can control the resources is required because of this dynamism of network. An adaptivity attribute is a part of the path maintenance parameters and can be presented as re-optimization [8]. According to this adaptivity attribute, the execution of an optimization algorithm depends on the network state in considering stability. This adaptivity attribute is the necessary attribute at the stage of Network Topology Migration described in section 4.1.5. The adaptive algorithm that has a minimum effect on the network state should support this adaptivity attribute and can provide provisioning at the time of implementation, and control the network state. 6.5 Priority Attribute Priority attribute gives priority to the emergency data and real time data and should be considered with the above resource class attribute and policy attribute. For the implementation of the priority attribute optical resources with excellent quality characteristics- for example resource that its end-to-end BER is lower is allocated or backup path leased for restoration is allocated- are allocated to high priority. And routing and signaling protocol that are updated to occupy optical characteristics can be considered to include the function that can allocate and control the priority. Also for the priority services the preemption that depends on priority should be allowed. 6.6 Preemption Attribute The preemption attribute should be considered with the priority attribute and resilience attribute and it is the function that can give itself using resources to high priority data according to network priority. This function is useful for the efficiency of resources and costs and especially for the implementation of various prioritized restoration policies. However low priority traffic should be used for the optical link and path with preemption attribute, and the provisioning function should be able to control the preemption attribute. D. G. Kim Expires - December 2002 [Page 13] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 6.7 Resilience attribute The resilience attribute for network stability can be defined as a function that when a fault of link or node occurs, a used traffic stream is transferred to a new path through an alternate path and the path generation algorithm. The type of recovery mechanism of the link can be classified as below according to the time of capacity allocation and the decision about backup capacity [11]. 1) Dedicated Protection is that capacity allocation and decision of backup that is concluded before a fault occurs and for example, a 1+1 recovery scheme gives a dedicated protection path for a link or wavelength. This schemeÆs advantages are that it is simple to implement and has fast recovery, but its disadvantage is that it is poorer in the utilization of resources. 2) Preplanned Restoration is that backup capacity that is decided before a failure but allocation of backup capacity is decided after a failure. For example, when high priority traffic and low priority traffic are used together and high priority traffic has failed, low priority traffic is removed. A 1:1 recovery scheme is a link of low priority traffic that is given to high priority traffic. And 1:N recovery scheme is the other example that recovery depends on priority traffic and one backup path is established for the N link through the control signal assuming that a failure in all links does not occur simultaneously. This scheme is achieved by signaling so that recovery time is slower than dedicated recovery and more complex but is excellent in resource utilization 3) Dynamic Restoration is the backup capacity and decision that are executed after failure and is most excellent but has the disadvantage of having recovery time is low and the implementation is complex. In the optical network fast recovery is required and therefore a fast path generation algorithm is required for introducing Dynamic Restoration into the optical network. Also the Protection Link Group (PLG) of concept is required for providing the differentiated services of protection according to the priority of the link or data. The Protection Link Group provides differentiated recovery time when the priority link and its backup path are established by using the three recovery schemes above. This PLG scheme can provide differentiated services in the view of resource utilization and network policy because the N: 1 or N: M protection group based on priority can be created. 7. Implementation Considerations If considering the common information for the construction of the GMPLS network based on these attributes in IP and optical domain, next the requirements of each layer should be considered [6]. 1)IP layer D. G. Kim Expires - December 2002 [Page 14] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 - information about virtual topology (set of light path connected between routers) of IP network - information about link state of IP network 2) WDM layer - information about Physical Topology ( set of fiber connected between WDM devices) - information about the continuity of wavelength allocated to each link, e.g. wavelength converter is equipped or not - information about switching capability and port usage (information of capability of fiber, waveband, and wavelength switching for each port) - information about fiber, the physical link, for example the number of available wavelengths per fiber, directionality, optical SNR of wavelength, BER - information about the light path, for example, the identity (ID) of the component at the destination network, port ID of adding light path, port ID of dropping, directionality, Fiber ID, throughput per wavelength, end-to-end SNR, Shared Risk Link Group (SRLG) ID, Protection Link Group (PLG) ID etc. 8. Conclusion This manuscript presented a set of requirements for traffic engineering over GMPLS and attributes that should be considered. Especially a network state information database for traffic engineering is a necessary requirement that should be considered when extended to an optical network. Through the definition and application scheme of attributes that consist of network state information database, the model of traffic engineering can be proposed for optimal utilization of resources when a real network is presented. When GMPLS network is popular in the future, the scheme that adds and controls information of optical characteristics to present the IP control element will be an important scheme to improve traffic engineering. Security Considerations This document does not introduce new security issues beyond those inherent in GMPLS. It is, however, possible that using the suggested network management attributes provisioning can be done as administrative usage. D. G. Kim Expires - December 2002 [Page 15] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 References 1 Brander, s.,"The Internet Standards Process -- Revision 3",RFC 2026, October 1996. 2 Brander, s.,"Key words for use in RFCs to Indicate Requirement Levels",BCP 14, RFC 2119, March 1997. 3 E. Mannie,"Generalized Multiprotocol Label Switching (GMPLS) Architecture",Internet Draft, work in progress. 4 D. Awduche, J. Malcom, J. Agogbua, M. O'Dell, J. McManus, " Requirement for Traffic Engineering over MPLS",ITEF RFC 2702, September 1999. 5 D. Awduche, A. Chiu, I. Widjaja, X. xiao,"Overview and Principles of Internet Traffic Engineering",IETF RFC 3272 May 2002. 6 D. Awduche, et al, "Multi-protocol Lamda Switching: combining MPLS traffic engineering control with Optical Cross Connects",IETF Internet Draft Nov 1999. 7 Kevin H.Liu, chang dong Liu and John Y.Wei, "Overlay vs. Integrated Traffic Engineering for IP/WDM Networks",GLOBECOM '00, IEEE, vol.2, 2000. 8 Hang Lu, Ruifeng wang, yugeng sun, "An architecture of Traffic Engineering", Circuit and systems, IEEE, APCCAS 2000, pp.70- 73, 2000. 9 R.M. Antonio, B. Paul, K, Murali,"The Optical Internet: Architecture and Protocols for Global Infrastructure of Tomorrow", IEEE Communication Magazine, Vol. 39, No. 7, pp.152-159, July 2001. 10 Nada Golmie, Thomas D, Ndousse,"A Differentiated Optical Service Model for WDM networks",IEEE Communication Magazine, Vol. 38, No. 2, pp.68-73, February 2000. 11 Eric Mannie,D. Papadimitriou "Recovery (Protection and Restotation) Terminology for GMPLS",IETF Draft, draft-mannie- gmpls-recovery-terminlogy-00.txt, February 2002. 12 Germano Gasparini, Gert Grammel, Dimitri Papadimitriou, "Traffic Engineering Extensions to OSPF and ISIS for GMPLS Control of G.709 Optical Transport Networks",ITEF Draft, draft-gasparini-ccamp- gmpls-g709-ospf-isis-03.txt, June 2002. D. G. Kim Expires - December 2002 [Page 16] draft-kim-ccamp-gmpls-nsid-00.txt June 2002 Author's Addresses Dae-gun Kim Korea Telecom, Korea University 1,5-ka, Anam-dong, Sungbuk-ku, Seoul,136-701, Korea Phone : 82-2-927-6116 e-mail : dkim@kt.co.kr Sung Woo Ryu Korea Telecom, Korea University 206 Jungja-dong Bundang-gu, Songnam-city Kyonggi-do ,463-711, Korea Phone : 82-2-929-5625 e-mail : isdn@kt.co.kr Jun Kyun Choi Information and Communication University (ICU) 58-4, Hwaam-dong, Yuseong-gu, Daejeon, 305-732, Korea Phone: 042-866-6122 e-mail : jkchoi@icu.ac.kr Chul-Hee Kang Korea University 1,5-ka, Anam-dong, Sungbuk-ku, Seoul, 136-701, Korea Phone: 82-2-927-6116 e-mail: chkang@widecomm.korea.ac.kr D. G. Kim Expires - December 2002 [Page 17]