In today’s market conditions, survival of a telecom service provider depends on efficient utilization of all the resources viz. network, manpower, frequency spectrum, and IT infrastructure. However, maximum RoI can be derived from the efficient use of the most precious resource of all—the network.

Most of the service providers have networks consisting of multi-domain, multi-technology and multi-vendor equipment. One big issue is of interoperability between these equipment and their integration into a single operating support system (OSS)/business support system (BSS). It is almost impossible for different proprietary implementations to interoperate unless the vendors jointly develop these implementations. Therefore, a common industry standard is important, and fortunately, various segments of the industry are working towards it. Generalized Multi-Protocol Label Switching (GMPLS) is the solution that can help service providers reduce operational expenditure as well as increase the services they offer to their
customers.

Evolution of GMPLS
Today’s transport network infrastructure provides excellent performance and reliability for voice traffic, the bulk of traffic prior to 1995. Since 1995, however, there has been a dramatic increase in data traffic primarily driven by Internet’s explosive growth. The service providers need solution that enables them to carry a large volume of voice and data traffic, and management of network devices in a most cost-efficient manner. This urge of service providers is the main driving factor in the evolution and enhancement of the MPLS suite of protocol, resulting in emergence of GMPLS.

GMPLS has evolved from MPLS—the original Internet Engineering Task Force (IETF) standard intended to enhance the forwarding performance and traffic engineering intelligence of packet-based (ATM, IP) networks. With support from the IETF and the Optical Internetworking Forum (OIF), it is fast becoming an industry standard. GMPLS extends these switch capabilities so that it is not only packet-switch capable (PSC), but also time division multiplexing capable (TDMC), fiber switch capable (FSC), and lambda switch capable (LSC). Therefore unlike MPLS, which is supported mainly by routers and data switches, GMPLS can be supported by a variety of optical platforms including SONET ADMs, Optical Cross-connects (OXCs) and DWDM systems. This will allow an entire infrastructure, extending from the access network to the core network to utilize a common control plane.

Objectives of GMPLS
Development of GMPLS began with the promise that it is possible to implement full integration of provisioning for all traffic types. GMPLS was thus developed with the goal of creating a single suite of protocols that would be applicable to all service and transport traffic. The main objectives for defining the GMPLS standards are:

n Interoperability between equipment from multiple vendors
n Automated network resource management
n Traffic engineering
n Rapid end-to-end service deployment and provisioning
n Automated network protection and restoration
n Service level agreement
n Optical virtual private networks (O-VPN)

GMPLS brings the intelligence and dynamic circuit (or path) provisioning of packet services to TDM and wavelength services. Its extensions offer a common mechanism for data forwarding, signaling, and routing on transport networks. GMPLS, thereby extends the MPLS label and label switched path (LSP) mechanisms to create generalized labels and generalized LSPs. These extensions affect routing and signaling protocols for activities, such as label distribution, traffic engineering, and protection and restoration.

GMPLS emphasizes on the control plane that performs connection management for the data plane for both packet switched capable (PSC) interfaces and non-packet switched interfaces. The non-packet switched interface includes TDM Capable, LSC, and FSC. The MPLS requires the LSP be set up between routers at both ends, while GMPLS extends the concept of LSP setup beyond routers. The LSP in GMPLS can be set up between any similar types of label switching devices at both ends. For example, the LSP can be set up between SDH add/drop multiplexers (ADM) to form a TDM LSP; the LSP can also be set up between two wavelength switching capable systems to form a LSC LSP; or the LSP can be set up between fiber switching-capable photonic cross-connect systems to form an FSC LSP. In GMPLS, different types of interfaces work together by nesting one LSP inside another. This functionality allows the system to scale better by forming a forward hierarchy.

The Building Blocks
The GMPLS control plane is made of several building blocks and these building blocks are based on well-known signaling and routing protocols that have been extended and/or modified to support GMPLS. Only one new specialized protocol is required to support the operations of GMPLS, a signaling protocol for link management.

GMPLS is indeed based on the traffic engineering (TE) extensions to MPLS, known as MPLS-TE. GMPLS extends the two signaling protocols defined for MPLS-TE signaling, i.e. RSVP-TE and CR-LDP. Following are the major enhancements to these signaling protocols in order to meet the GMPLS requirements of traffic engineering:
n Supports the set-up of bi-directional LSPs for network protection
n Signaling for the establishment of a back-up path (protection information)
n Expediting label assignment via suggested label
n Waveband switching support—set of contiguous wavelengths switched together

GMPLS also extends two traditional intra-domain link-state routing protocols already extended for TE purposes, i.e. OSPF-TE and IS-IS-TE. Following are the major enhancements to these routing protocols in order to meet the GMPLS requirements of traffic engineering:
n Advertising of link-protection type (1+1, 1:1, unprotected, extra traffic)
n Forwarding adjacency LSP for improved scalability
n Accepting and advertising links with no IP address—link ID
n Incoming and outgoing interface ID
n Shared risk link group (SRLG)—diversity routing of paths for protection and restoration that is different from the primary path

The use of technologies like dense wave division multiplexing (DWDM) implies that we can now have a very large number of parallel links between two directly adjacent nodes (hundreds of wavelengths, or even thousands of wavelengths if multiple fibers are used). Such a large number of links was not originally considered for an IP or MPLS control plane, although it could be done. Some adaptations of that control plane are thus required in the GMPLS context. The link management protocol (LMP) was specified for this purpose. The main functions of LMP are as follows:
n IP control channel maintenance: Mechanism to maintain control channel connectivity
n Link verification: Verify the physical connectivity of the data bearing link between the neighboring nodes
n Link-property correlation: Identification of the link properties of the adjacent nodes
n Managing link failures: Fault localization and fault notification

Benefits of GMPLS
A GMPLS network offers improved network efficiency and flexibility. GMPLS allows each network layer to be managed according to its unique attributes. It enables utilization of the inherent differences of the network layers to ensure optimal use of network resources. New service offerings enabled by GMPLS mean new revenue opportunities for service providers.

Network provisioning: GMPLS enables faster and more accurate provisioning. In a GMPLS network, edge devices can become peers of GMPLS core devices, for dynamic end-to-end provisioning. If the edge devices are not GMPLS-aware, GMPLS can still be used to ease the provisioning burden in the core network in the same style as switched PVCs in ATM.

Traffic engineering: Effective traffic engineering is one of the keys for maximizing return on investment while improving service offerings. Implementation of GMPLS Traffic Engineering (TE) and optical extensions for routing and signaling protocols provide enhanced network information, intelligent path computation and common signaling to packet, TDM, and wavelength services.

Bandwidth on demand: Services with quality of service (QoS) constraints and the large bandwidth increments they need are extremely difficult to provision in real-time on a layered network. GMPLS-enabled architecture allows any combination of fine grain packet LSPs to coarse grain STM-64 (OC-192) LSPs. Thus, service providers deploying GMPLS networks in a consolidated network will be able to offer a customer, for example, an STM-4 (OC-12) from Mumbai to Singapore in the morning and an STM-64 (OC-192) from Mumbai to New Delhi the same afternoon, while making optimal use of all network resources.

Differentiated services: A GMPLS network enables service providers to manage traffic across all layers. They can view all the network layers and efficiently allocate resources to the most appropriate layer, thereby tailoring differentiated services to their customers’ varied and changing needs, while reducing costs through optimal resource usage.

Service level agreements: GMPLS enables service providers to reap greater benefits through more comprehensive and flexible service level agreement (SLA) offerings to customers. With resource allocation not longer restricted to a specific network layer, service providers have greater flexibility in designing and enforcing SLAs, which can be used to generate new revenues.

Cost effective services: GMPLS offers service providers a vehicle to help them migrate their networks from the current complex and costly architectures to simpler, more efficient models. Service providers who deploy GMPLS will not only see significant savings through improved network efficiencies, but will be able to offer advanced, revenue-rich services to existing and new customers.

GMPLS Issues
There are some issues with GMPLS that deserve attention. These issues are briefly discussed below:

Network management systems: The most important parameter in managing a traditional IP network e.g., the Internet, is address reachability. GMPLS network-management system needs to keep track of thousands of LSPs for their operational status, routing paths, and traffic engineering. This makes the GMPLS network management system more complex as compared to the traditional ones.

Interworking: Interworking in the control plane is very complicated as different suites of protocols are used in separate networks such as routing, private network-to-network interface in ATM versus OSPF–TE in GMPLS networks. GMPLS switching can be packet-based, TDM-based, wavelength-based, waveband-based, or fiber-based. Several industry forums are currently addressing the specifics of interworking between these networks. Some of the prominent forums are MPLS Forum, the ATM Forum, and the Frame Relay Forum.

Security: Traditional IP routing reads the contents of the header of a received packet to determine the next hop for it. Though time-consuming, this step helps in establishment of firewalls, as the source and destination information is available in the packet headers that are globally unique. In contrast, in GMPLS/MPLS, labels are used which are understood and used internally only by the GMPLS device itself. As such, these labels cannot be used for network-security purposes. One way to establish security in a GMPLS network is to enforce access security during the connection set-up time, like other connection-oriented networks like X.25 or ATM.

Network stability: When a new resource is deleted or added in a GMPLS network, the set of control information that is exchanged is larger than that of a traditional IP network. While not tested, theoretically, an MPLS/GMPLS network would take a relatively longer time to achieve a stable state than would a traditional IP network when the network is disrupted.

Lalit Bansal and Mehul Sanghavi consultants, Infosys Technologies