Ipo

1. INTRODUCTION DenseWavelength Division Multiplexing (DWDM)[1] is the process of multiplexing signal of different wavelength onto a single fiber. Through this operation, it creates many virtual fibers each capable of carrying a different signal. At its simplest, WDM system can be viewed as a parallel set of optical channels, each using a slightly different light wavelength, but all sharing a single transmission medium. This new technical solution can increase the capacity of existing networks without the need for expensive re-cabling and can tremendously reduce the cost of network upgrades. 'Internet Protocol (IP) over DWDM' is the concept of sending data packets over an optical layer using WDM for its capacity and other operations. In the modern day world, the optical layer has been supplemented with more functionality, which were once in the higher layers. This creates a vision of an all-optical network where all management is carried out in the photonic layer. The optical network is proposed to provide end-toend services completely in the optical domain, without having to convert the signal to the electrical domain during transit.Transmitting IP directly over DWDM has become a reality and is able to support bit-rates of OC-192.using MPLS control plane over optical control plane (MPLambdaS) has made forwarding,routing,traffic engineering very simple.[2]

2. Why IP over DWDM is important An enormous amount of bandwidth capacity is required to provide the services demanded by consumers. For many applications, it is desirable that data be transported from one point in a network to one or more other points in the least possible time. For some applications the time sensitivity is so important that minimum delay is the overriding factor for all protocol and equipment design decisions. So, there is a need for Burst mode optical data switching and streaming data traffic as this will help cater for data centric applications as well as video applications with no draw backs at all. Internet Protocol over DWDM aims to reduce the cost per connected bit transmitted. The IP over DWDM system has inherent advantages because of the absence of many layers. It brings in the property of virtual fibers where each wavelength can be considered as a dedicated connection. The signals need not be converted onto an electrical domain for performing control operations on it. Hence, the latency in the IP/DWDM system is less compared to that encountered in the Synchronous Optical Network (SONET) system,[3] which allows data streams of different formats to be combined onto a single high-speed fiber optic synchronous data stream. The absence of a vendor specific component (SONET and other networking solutions (ATM[4], etc) are vendor specific) makes the DWDM system service transparent. IP over DWDM is overtaking ATM technology, because ATM is basically popular for the integration of the data and voice it provides. Now for people who have nothing to do with voice (data centric), question the idea of using ATM. Thus ATM is slowly loosing to DWDM because: ATM switches are the most expensive parts of the entire equipment. ATM is believed to be hopeless for handling video, which will probably represent most of the traffic on the public networks of the future. ATM is not keeping pace with the evolution of transmission technology. For example

the ATM community has only just agreed on a standard for10-gigabit interfaces.

3. IP Network Structure for Optical Networks The options shown in Figure 1 illustrates possible options for implementing an IP network over fiber optics, but now the drive is behind running IP over the best layer. At the bottom is the optical layer, based on an Optical Transport Network (OTN) that consists of the following sublayers:

Figure 1.

IP and the lower-layer evolution

Figure 1.(a) represents the structure of running IP over ATM/SONET/Optical and (b) is IP over SONET/Optical, while (c) represents the much-desired structure of IP over Optical (DWDM). In this structure the optical channel (Och) section defines an optical connection (light path) between two optical client entities. (The Och equates to a lambda in DWDM parlance.) The optical multiplex section (OMS) defines the connectivity and treatment for a multiplex or grouping of Och-level connections. (The OMS equates to a group of lambdas flowing over fiber-optic cable between two DWDM multiplexers.) The optical transmission section (OTS) defines how optical signals are transmitted over the optical media.[3] Figure 1.c (IP over Optical (DWDM)) shows the target architecture that many providers are aiming for: the SONET layer has been eliminated in favor of running IP directly over the optical layer. Eliminating the ATM and SONET layers means fewer network elements to manage [5], [6], [7]. The existing suite of MPLS traffic engineering (TE) protocols has been extended to operate over optical networks, and the IP layer, particularly the IP routers, can now interface directly with the optical layer. This intended all-optical transport network is hoped will maintain a high data rate, making the technology more attractive, since the systems can also perform well at high bandwidth speeds thereby reducing the latency. 3.1 Targeted IP Network Structure Figure 2 below shows the final intended structure of running IP over DWDM. There is no need for SONET in IP/DWDM networks because, starting with the basic

concept of layering. In most networks, a layered approach is used in building up the various pieces of equipment and various formats used in building a network. The basic idea is that each layer provides a service, hopefully in a service independent way to the layers above it. In this format it is thought that maximum flexibility to introduce new formats is maintained, while preserving the installed base of existing equipment. Hence there is a price paid for layering a system.

Figure 2 Collapse of layers in the IP over DWDM domain [8]

The multi-layer stack has a cascade of inefficiencies built up. This can be justified in view of the below stated facts that: Every layer runs at its own speed. e.g. DWDM runs in Gbps, ATM runs in Mbps and IP runs in Kbps. So, low speed devices cannot fill the wavelength bandwidth, thus creating the bottleneck for the system. Due to this new technological revolution, the core technologies are pushed to the edges of the network. For example: Now, Gigabit routers are available, so there is no need for grooming (or multiplexing), as a single router port is able to access all resources. Multiple layers also mean that there is excessive overhead data bits to be used in running the systems, which reduces the capacity of the system to deliver actual information, hence by running IP over DWDM directly all these redundancies and network problems can be half solved and taken care of.

4. MPLS (Multi Protocol Label Switching) It is an IETF framework that provide efficient forwarding, routing, designation and switching of traffic flow through network.[9,10] Components: 1. Ingress node (edge LSR-Label Switch Router) 2. core LSR 3. egress node (edge/egress LSR) To adapt MPLS to control optical CrossConnects (OXCs) four main issues need to betackled: addressing, signalling, routing and survivability. For the IP layer to communicate across a DWDM domain the ingress and the egress optical CrossConnect addresses mustbe known and they must be capable of resolving higher-level addresses [11 - 14].In an Ingress of the MPLS domain a Forwarding Equivalence Classes (FEC) is a

group of IP packets that are forwarded over the same path and treated in the same manner. The assignment may be based on e.g. host address, or a “longest match” of the destination address prefix of IP packets (thus destination). The FEC to which an IP packet is assigned is encoded with a short, fixed length label. Figure 3 shows the elements and components of an MPLS core network with three types of nodes. At the nodes of an MPLS network the labelled packets are forwarded based on the label-swapping paradigm. This means that the label associated with the IP packet is examined at each Label Switching Router (LSR) and is used as an index into the Label Information Base (LIB). The label is mapped on a next hop label forwarding entry in this In MPLS IP packets are classified in so-called Forwarding Equivalence Classes (FECs) at the table, which determine where to forward the packet. The old label is replaced with a new label and the packet is forwarded to its next hop. Thus, while an IP packet is in the MPLS domain the packet’s network header is no subject of further analysis at subsequent MPLS hops.

Fig 3 In an MPLS core network three types of nodes are used [17]: 1) Ingress node with label assignment. 2) Intermediate nodes with label lookup, label swapping and forwarding.3) Egress node with label removal

In order to establish and maintain a path according to the information collected by the routing protocols, the LSRs along this route must assign and distribute labels among neighbouring nodes. Herewith a Label Switched Path (LSPs) is created between ingress and egress points of the MPLS domain. An LSP is created by the concatenation of one or

more label switched routers, allowing a packet to be forwarded by swapping labels. Thedistribution of labels is required to allow a LSR to inform another LSR of the label/FEC binding it has made. With this the LIB in the LSRs used in the label swapping process is kept up to date. A Label Distribution Protocol (LDP) accomplishes the distribution of label/FEC bindings among participating LSRs, in order to establish an LSP.

5. Multi-Protocol Lambda Switching (MP S) Lambda switching is derived from Multi-Protocol Label Switching (MPLS),generally refered as GMPLS. A number of different approaches have been suggested to “push” as much as possible of the traffic flow down to layer 2, thus performing “switching” instead of “routing” in IP-networks.MP S is an effort from the Internet Engineering Task Force (IETF) to create a standardised solution for this. The “label” is a number, assigned at an IP router at the edge of a label switched or MP S domain, which identifies a path across the network, so that the packets may be routed more quickly without having to lookup the destination address in the IP packet, and in this case the wavelength is used as the label [11]. This label can either be appended to the IP packet, or can be stored in the encapsulation frame when a suitable field exists. With MP S no physical connection are dedicated, but a pathis defined and all the packets in a particular session are sent along that path by giving them all the same label when they enter the MP S network, figure 3 illustrates the MP S. At each node each packet is routed according to its label value and is provided with a new label for use at the next node, hence the name label switching comes in.

Fig. 3 Concept of MP S The LSR, in this case, is an Optical CrossConnect (OXC). The label used in MP S is the wavelength that the data is transmitted on just as an LSR (Label Switching Router), maps an incoming label to an outgoing one, the OXC maps an incoming wavelength to an outgoing wavelength and OXCs need routing protocols and label distribution protocols. Hence for simplicity, it can be assumed that MP S to be just implementing Label Switching in the Optical domain.

6 Justification to go for MPLS/ MP S MPLS provides a number of benefits to IP network providers: Efficient forwarding: due to the use of labels, core routers/LSR no longer need to make a route lookup in very large routing tables, instead a much smaller LIB can be used to make the forwarding decision; Service differentiation: certain paths, or FECs, may be assigned different CoS. Use of the label in combination with CoS parameters allow easy identification of such traffic streams; Traffic engineering: because MPLS paths are topology based, and labels are used to identify them, a path easily can be re-routed. Also here the label is used to accomplish this. The Tabel 1 can justify the reason to go for this and the release of burden for MPLS approach from normal IP service Functions in DATAGRAM service Functions in IP router with MPLS 1. Remove and create the frame 1. create and remove frame header and header and trailer trailer. 2. check the packet header, check sum 2. decrement TTL 3. decrement TTL (Time to leave) 3. Look up the next MPLS hop in the 4. perform any packet security switching tabel (very fast) 5. process any IP options field 6. create new packet header and field values 7. compute a new header checksum 8. look up the next hop router in the routing table 9. deliver the frame header to the output interface for transport Tabel 1:functions of normal IP and IP/MPLS In MPLS there exists no buffering of packets. Hence no scheduling algorithm is required. The packets are sent as and when they arrive. Besides the numerous advantages, there are some disadvantages as well: MPLS is topology oriented, and hence a label needs to be assigned to each route. This is a weakness, since a route may not be in use, and therefore the label would be wasted.

7. IP- OPTICAL networking model 7.1 service model The optical network model considered consists of multiple Optical Crossconnects (OXCs) interconnected by optical links. each OXC is assumed to be capable of switching a data stream from a given input port to a given output port. This switching function is controlled by appropriately configuring a crossconnect table. Conceptually, the cross connect table consists of entries of the form , indicating that data stream entering input port i will be switched to output port j. An "lightpath" from an ingress port in an OXC to an egress port in a remote OXC is established by setting up suitable crossconnects in the ingress, the egress and a set of intermediate OXCs such that a continuous physical path exists from the ingress to the egress port. Lightpaths are assumed to be bi-directional, i.e., the return path from the egress port to the ingress port follows the same path as the forward path.It is assumed that one or more control channels exist between neighboring OXCs for signaling purposes.

Fig 4. domain network model In this section two possible service models are discussed in brief.The first being at the optical UNI and the second being at the optical sub-network NNI[15]. 7.1.1. Client Server Model Under this model the optical network primarily provides a set of high bandwidth pipes to the client requesting such. Standardized signaling can be used to invoke the following: Light path creation. Light path deletion. Light path modification. Light path status enquiry.

The continued operation of the system requires that the client systems continuously register with the optical network. Signalling extensions need to be added to allow clients to register, deregister and query other clients for an optical-networked administered address so that lightpaths can be established with other clients across the optical network. Along with these signaling extensions a service discovery mechanism needs to be added which will allow the client to discover the static parameters of the link along with the UNI signaling protocol being used on the link. In this service model the routing protocols inside the optical network are exclusive of what is followed inside the client network. Only a minimal set of messages need to be defined between the router and the optical network. RSVP-TE, LDP or a TCP based control channel can be used for the same. Within the optical cloud NNI interface is defined between the various optical subnetworks. 7.1.2 Integrated Service Model In the Integrated Service Model the IP and the optical networks are treated as a single network and there is no distinction between the optical switches and the IP routers as far as the control plane goes. MPLS would be the preferred method for control and routing and there is no distinction between the UNI, NNI or any other routerrouter interface. Under this model, optical network services are provisioned using MPLS signaling as specified in [GMPLS]. In this service model the edge router can do the creation and modification of the label switched paths across the optical network. In some sense this resembles the client server model just presented, but it seems to promise seamless integration when compared to the client server model. OSPF with TE extensions to support optical networks could be used to exchange topology information and do the routing .It might happen in an optical network that a LSP across the optical network may be a conduit for a lot of other LSPs. This can be advertised as a virtual link inside a forward adjacency in protocols like OSPF. Thus from the point of view of the data plane an overlay is created between two edge routers across the optical network. Some of the proposed models for interaction between IP and optical components in a hybrid network are (1) Augmented model (2) Overlay model (3) Peer model Augmented Model In the augmented model, there are actually separate routing instances in the IP and optical domains but information from one routing instance is leaked into the other routing instance. For example IP addresses could be assigned to optical network elements and carried by optical routing protocols to allow reachability information to be shared with the IP domain to support some degree of automated discovery. Peer Model In the peer model(fig5), the same signaling protocol (GMPLS) is used to set up the whole path, including both the packet and transport elements, with the client.s requests being spliced into the requests needed to set up the internal tunnel. Whilst this has the advantage of providing a single management layer to the Service Provider, there are some significant considerations:

Fig 5 GMPLS Peer Model • The Service Provider may wish to keep topology information of the optical core private. • Much of the equipment available for the optical core today uses proprietary signaling rather than GMPLS. • Service Providers may wish to keep the paths set in the transport core as stable as possible, managed on a overall network basis for protection and restoration purposes. Thus they would not want the service requests of the packet network to be fulfilled automatically. Overlay Model In this model the optical and packet networks are managed independently and may potentially be signaled by different protocols. One solution is for the optical transport network (OTN) to provide a User-to-Network Interface (UNI) that allows client network devices to request connections across it dynamically.

Fig. 6: The OTN and UNI Access A signaling protocol needs to be defined to bridge the UNI. An internal tunnel between endpoint UNI-N nodes is set up using either GMPLS or some

proprietary signaling mechanism. This tunnel can be provisioned ahead of time, or on demand in response to the client request. The data path between the client nodes is then set up by overlaying it upon this internal tunnel.

Fig. 7 : UNI Overlay Model

An internal tunnel between endpoint UNI-N nodes is set up using either GMPLS or some proprietary signaling mechanism. This tunnel can be provisioned ahead of time, or on demand in response to the client request. The data path between the client nodes is then set up by overlaying it upon this internal tunnel. 8.UNI Services The client and OTN edge nodes that support UNI are not traditional signaling peers, in that the OTN service provider agrees a contract (in the billable sense!) with the client to provide a certain level of service across the network. A UNI message from client to OTN is best viewed as a request for service. Given the contractual angle, it is important for there to be a defined method for identifying and authenticating clients across the UNI. This allows the OTN to check that a client.s account is in good standing before granting a connection request. For billing and auditing purposes, the OTN may also need to keep track of the connection by assigning a connection identifier that is valid beyond the lifetime of the connection and returning this to the client. In order to set up a connection, a client needs to discover what services are available from the OTN and, that done, signal its requirements across the UNI. The types of service parameters signaled across the UNI are • requested bandwidth of the connection. • class of service (e.g. the network restoration/protection strategy required) • diversity (discussed below) • data plane specific characteristics (e.g. for SONET/SDH, port, transparency and concatenation information). A fundamental requirement of UNI is that clients are not allowed access to OTN internal

addresses or topology information. This means that UNI connection requests are not allowed to specify explicit routes. The .Diversity. parameter allows a UNI client to request that a new connection follows a different route to a set of previous connections without requiring internal knowledge of the OTN. This is necessary to allow the client to request a disjoint backup to a primary route. 9.Routing Approaches 9.1Fully peered routing model This routing model is used for the peer model described above. Under this approach there is only one instance of the routing protocol running in the IP and Optical domains. An IGP like OSPF or IS-IS with suitable optical extensions is used to exchange topology information. These optical extensions will capture the unique optical link parameters. The OXCs and the routers maintain the same link state database. The routers can then compute end-to-end paths to other routers across the OXCs. Such a Label Switched Path (LSP) can then be signaled using MPLS signaling protocols like RSVP-TE or CR-LDP. This lighpath is always a tunnel across the optical network between edge routers. Once created such lightpaths are treated as virtual links and are used in traffic engineering and route computation. As and when forwarding adjacencies (FAs) are introduced in the link state corresponding links over the IP Optical interface are removed from the link state advertisements. Finally the details of the optical network are completely replaced by the FAs advertised in the link state. 9.2 Domain Specific Routing This routing model supports the augmented routing model. In this model the routing between the optical and the IP domains is separated with a specific routing protocol running between the domains. The focus is on the routing information to be exchanged at the IP optical interface. Interdomain routing protocols like BGP may be used to exchange information between the IP and optical domain. OSPF areas may also be used to exchange routing information across the two domains. 9.3 Routing using BGP BGP will allow IP networks to advertise IP addresses within its network to external optical networks while receiving external IP prefixes from the optical network. Edge routers and OXCs can run External BGP (EBGP). Within the optical network EBGP can be used between optical subnetworks across the NNI and Internal BGP (IBGP) can be used within the optical network. Using this scheme it is essential to identify the optical network corresponding to the egress IP addresses. The reason is as follows. Whenever an edge router wants to setup a LSP across an optical network it is just going to specify the destination IP. Now if the edge router has to request another path to the destination it must know if there already exist lightpaths with residual capacity to the destination. To determine this it needs to know which ingress ports in an OXC correspond to which external destination. Thus a border OXC receiving external IP addresses by way of EBGP must include information about its IP address and pass it on to the edge router. The edge router must store this association between the OXCs and the external IP addresses and need not propagate the egress address further. Specific mechanisms to propagate the BGP egress addresses are yet to be determined.

9.4 Routing using OSPF OSPF supports the concept of hierarchical routing using OSPF areas.Information across a UNI can be exchanged using this concept of a hierarchy. Routing within each area is flat. Routers attached to more than one areas are called Area Border Routers (ABR). An ABR propogates IP addressing information from one area to another using a summary LSA. Domain specific routing can be done within each area.to other routers across the OXCs. Such a Label Switched Path (LSP) can then be signaled using MPLS signaling protocols like RSVP-TE or CR-LDP. This lighpath is always a tunnel across the optical network between edge routers. Once created such lightpaths are treated as virtual links and are used in traffic engineering and route computation. As and when forwarding adjacencies (FAs) are introduced in the link state corresponding links over the IP Optical interface are removed from the link state advertisements. Finally the details of the optical network are completely replaced by the FAs advertised in the link state. 9.5 Overlay Routing Overlay routing is much like the IP over ATM and supports the overlay connection model. IP overlays are setup across the optical network. Address resolution similar to that in IP over ATM is used. The optical network can maintain a registry of IP addresses and VPN identifiers it is connected to. On querying the database for an external IP address it would return the appropriate egress port address on the OXC. Once an initial set of lightpaths are created VPN wide routing adjacencies can be formed using OSPF. The IP VPN would then be "overlayed" on the underlying optical network which could have an independent way of routing. 9.6 Path Selection CR-LDP

PATH SELECTOR

RSVP-TE

TE DATA BASE

OSPF TE EXT

IS-IS TE EXT

Fig 8.Lightpath Selection A possible scenario for path selection is presented in fig 8. These systems use CR-LDP or RSVP-TE to signal MPLS paths. These protocols can source route by consulting a traffic engineering database, which is maintained along with the IGP database. This information is carried opaquely by the IGP for constraint based routing. If RSVP-TE or CR-LDP is used solely for label provisioning, the IP router functionality must be present at every label switch hop along the way. Once the label has been provisioned by the protocol then at each hop the traffic is switched using the native capabilities of the device to the eventual egress LSR. Path selection can be online or offline. An offline computation is normally centralized while as an online computation is normally distributed. An offline

computation is facilitated by simulation or network planning tools and can be used to provide guidance to subsequent real time computations. An online computation may be done whenever a connection request comes in. A combination of offline and online computations may be used by a network operator. Offline computations are used when complicated traffic engineering, demand planning, cost planning and global optimization is a priority. In case of online computations there can be two choices when it comes to routing. 1) Explicit routing using a global view of the network can be used to calculate the most optimal solution taking into consideration constraints other than link metrics. 2) Hop by hop routing using path calculation at every node. This may not be able to provide an optimal solution taking into consideration constraints other than non additive link metrics. 9.7 Constraints on Routing The constraints highlighted here apply to any circuit switched networks but differences with an optical network are explained where applicable[GMPLS-CONTROL]. One of the main services provided by any transport network is restoration. Restoration introduces the constraint of physically diverse routing. Restoration can be provided by pre-computed paths or computing the backup path in real time. The backup path has to be diverse from the primary path at least in the failed link or completely physically diverse. A logical attribute like the Shared Risk Link Group (SRLG) is abstracted by the operators from various physical attributes like trench ID and destructive areas. Such an attribute may be needed to be considered when making a decision about which path to take in a network. Two links which share a SRLG cant be the backup for one another because they both may go down at the same time. In order to satisfy such constraints path selection algorithms are needed to find two disjoint paths in a graph. Suurballe’s algorithm as discussed [16] is a good example of an algorithm to find two node disjoint paths in a network. Another restoration mechanism is restoration in a shared mesh architecture wherein backup bandwidth may be shared among circuits. It may be the case that two link disjoint paths share a backup path in the network. This may be possible because a single failure scenario is assumed. A few heuristics to optimize the bandwidth allocated to a backup path in a mesh architecture have already been proposed [Bell-Labs]. Another constraint of interest is the concept of node, link, LSP inclusion or exclusion, propagation delay, wavelength convertibility and connection bandwidth among other things. A service provider may want to exclude a set of nodes due to the geographic location of the nodes. An example would be nodes lying in an area which is earthquake prone. Propagation delay may be another constraint for a large global network. Traffic from the US to Europe, shouldn’t normally be routed over links across the Pacific ocean but instead should use links over the Atlantic ocean since propagation delay in this case would be much less. Wavelength convertibility is a problem encountered in waveband networks. It refers to ability of OXC to crossconnect two different wavelengths. The wavelengths may be completely different or slightly different. Since wavelength convertibility currently involves an optical-electrical-optical (OEO) conversion, vendors may selectively deploy these converters inside the network. Therein lies the problem of routing a circuit over a network using the same wavelength. This requires that the path selection algorithms

know the availability of each wavelength on each link along the route. With link bundling, this is difficult since information about all the wavelengths may be included in the same bundle. Link probing may have to be employed at the source router to find out the number of wavelengths available along the path. Bandwidth availability is another consideration in routing. This is simplified in a wavelength optical network since requests are end to end. 10.PROTOCOLS FOR GMPLS PROTOCOLS DESCRIPTIONS ROUTING: Routing protocols for the auto-discovery of OSPF-TE,IS-IS-TE network topology, advertise resource availability (e.g. bandwidth or protection type). The major enhancement are as follows: Advertising of link protection type( 1+1,1:1,unprotected,extra traffic) Implementing derived links (FA) for improved scalability. Accepting and advertising with no IP address – link ID Incoming and outgoing interface ID

SIGNALING: RSVP-TE, CR-LDP.

Route discovery for back up that is different from the primary path. Signaling protocols for Traffic engineered LSP’s. The major enhancements are : Label exchange to include non-packet networks(generalized labels) Estaablishment of bi-directional LSP Signaling for establishment of back up path Waveband switching support-set of contiguous wavelength switched together

LINK MANAGEMENT: LMP

Control channel management: Established by negotiating link parameters and ensuring the health of a link (hello protocol) Link connectivity Verification: Ensures the physical connectivity of the link between the neighboring nodes using a PING-like test message

11. Future N/W architecture The MPLS based network architecture (fig 9)has been found to be favorable for the IP/DWDM system. It shall host features like: • MPLS control plane will handle costraints like Data rate, Attentuation, Dispersion, Length, Delay. • Using Next hop forwarding label entry (NHFLE) for determining the output port (or path) of a packet. • The control plane shall be enforced by network management by having a dedicated supervisory channel. Each OXC shall be an IP addressable device. • The subnets will acts as a single abstract node performing restorations within itself Fig 9: future network architecture

The required protocol stack, for achieving those goals, would be as shown in Fig.7. The MPLS control plane shall regulate all the connections. The switch fabric would perform packet switching based on the wavelengths. The data plane shall be responsible for transmitting the packets. The switch shall maintain mapping between the wavelength + port at input and the wavelength + port at output. The table ensures that the packets reach the proper destination eventually. This is similar to how wavelength routing is done.

CONCLUSIONS: This paper has been discussing several issues arising in an IP-optical network. The basic concepts underlying an IP over DWDM system .The IP/DWDM systems shall support the Open architecture & provide complete service transparency. It shall host a MPLS control plane in the OXCs for providing much of the services. The future holds many a challenges to the all-optical networks. But, the commercial implementations for IP over DWDM are not far away. It opens the pathway to Terabit networking and unleashes the enormous bandwidth potential of the silica fiber. The trend of IP/DWDM solutions over the last few years seems to have taken an exponential growth. DWDM acts as the stepping stone towards a true optical networking era.

LIST OF ACRONYMS BGP - Border Gateway Protocol CR-LPD - Constraint-Based Routing LDP DWDM – Dense Wavelength Division Multiplexing FA - Forwarding Adjacency FEC - Forwarding Equivalence Class GMPLS-Generalized Multi Protocol Label Switching IGP - Interior Gateway Protocol IS-IS – Intermediate System to Intermediate System Protocol LDP - Label Distribution Protocol LIB- Label Information Base LMP - Link Management Protocol LSP - Label Switched Path LSR - Label Switched Router MPLS - Multi-Protocol Lambda Switching NHRP - Next Hop Resolution Protocol NNI – Network to Network interface OCT - Optical Channel Trail OLXC - Optical layer crossconnect OMS - Optical Multiplex Section OSPF - Open Shortest Path First OTN - Optical Transport Network OTS - Optical Transmission Section OXC - Optical Crossconnect PSC - Packet Switch Capable PVC - Permanent Virtual Circuit QoS - Quality of Service RSVP - Resource reSerVation Protocol RSVP-TE - Resource reSerVation Protocol with Traffic Engineering TDM - Time Division Multiplexing TE - Traffic Engineering TTL - Time to Live UNI - User to Network Interface VC - Virtual Circuit

REFERENCES [1] optical networking and DWDM by walter goralski [2] optical networks 3rd generation by ulysses black [3] optical network:a practical perspective by Rajiv ramaswamy [4] High performances communication networks by jean varland [5] D. O. Awduche, Y. Rekhter, J. Drake, and R. Coltun, “Multi-protocol lambda switching: Combining MPLS traffic engineering control with optical crossconnects,” IETF Internet Draft, 1999. [6] H. Zang, J. P. Jue, and B. Mukherjee, “A review of routing and wavelength assignment approaches for wavelength-routed optical WDM networks,” Optical Networks Magazine, vol. 1 Jan 2000. [7] L. Zhang, S. deering, D. Estrin, S. Shenker, and D. Zappala, “RSVP: A New Resource ReSerVation Protocol,” IEEE Comm. Mag., May 2002 [8] IP Over DWDM by Suresh Lanka [9] “ www.ietf.org” [10] “www.mplsrc.com” [11] A. Banerjee et al., “Generalized Multiprotocol Label Switching: An Overview of Routing and Management Enhancements,” IEEE Commun. Mag., vol 39. No. 1, Jan. 2001. [12] A. Banerjee et al., “Generalized Multiprotocol Label Switching: An Overview of Routing and Management Enhancements,” IEEE Comm. Mag., Jan 2001 [13] P.Ashwood-Smith et al., “Generalized MPLS Signaling – CR-LDP Extensions,” Interent draft, draft-ietf-mpls-generalized-cr ldp-01.txt, Mar. 2001. [14] P.Ashwood-Smith et al., “Generalized MPLS – Signaling Functional Description,” Internet draft, draft-ietf-mpls-generalized-signaling-02.txt, Mar. 2001. [15] J. Luciani, B. Rajagopalan, D. Awuduche, B. Cain, Bilel Jamoussi, Debanjan Saha, "IP Over Optical Networks - A Framework," Internet Draft draft-many-ip-optical-framework-01.txt, November 2000. [16] J. Suurballe, "Disjoint Paths in a Network," Networks, vol. 4, 1974. [17] MPLS/GMPLS tutorial by IEC “www.iec.org”