Solutions to help your business Sign up for our newsletters Join our Community
  • Share

GMPLS: Intelligence for the Optical Network

For long-haul service providers, the network core is that part of the network that interconnects multiple metro networks. For metro service providers, the network core is the backbone network that interconnects multiple access networks that are also known as distribution networks. 

More on this Topic

Industry News

Blogs

Briefing Room

The term optical core network is, therefore, relative. However, these networks do have one significant attribute in common: They are all becoming more complex and more difficult to manage.

With the emergence of wavelength-based transport, today’s service providers must address yet another layer of complexity to an already complex, multi-layered network (Sonet, IP and ATM). For example, considering that: each node in an optical network may have multiple parallel links to its neighbor node, and that each node offers multiple wavelengths, carrying multiple protocols on each fiber, it has become significantly more difficult to provision and manage network resources. The challenge , therefore, is how to efficiently manage network complexity.

GMPLS
Generalized Multi-Protocol Label Switching 
a generalized version of MPLS, known as GMPLS, has been proposed to address this need and serve in a variety of network signaling layers, extending capabilities beyond those networks that are only packet or cell-based.
--from Polaris Networks' GMPLS Resource Center
Read More

From Calient Networks
Generalized Multiprotocol Label Switching:
An Overview of Routing and Management Enhancements
a pdf. file

This article presents an explanation of the merits of the intelligence offered by a new network control paradigm, known as Generalized Multi-Protocol Label Switching (GMPLS). GMPLS is seen as a key driver for managing optical networks and for automating network resource provisioning.

Today's Challenges

As emphasized above, optical networks, especially metro core, are growing increasingly complex. While the advent of dense wave division multiplexing (DWDM) systems and emerging optical cross-connects promise to dramatically increase overall bandwidth capacity, network operators are still challenged with the management of these resources. 

This complexity also creates challenges with regard to:

  • scalability issues

  • time-consuming processes related to discovery and establishment of optimal paths, and 

  • the inadequacies of offering paths with differentiated services. 

For example, even in a small network of 25 nodes, each connected with four fibers and each capable of 40 Lambdas, there are 4000 wavelengths that must be managed. In this example, even a single addition of a new link becomes tedious, as this new information must be updated throughout the network. To date, this process of keeping track and updating the available resources in the network has been done manually. 

Moreover, differentiated services are currently only offered in terms of resiliency in traditional optical networks such as Sonet. As a result, circuits may be protected, unprotected or pre-empted. However, this granularity is not adequate for the sophisticated applications that require Quality of Service (QoS) traffic parameters that could encompass the path, bandwidth, network congestion and protection level.

Overall, the industry’s response to these challenges are derived from two sources:

  • The implementation of proprietary routing and signaling protocols, and

  • Standard bodies’ specifications for a suite of protocols based on IP and originated from multiprotocol label switching (MPLS)--namely generalized MPLS (GMPLS).

GMPLS introduces the intelligence necessary to minimize the manual intervention currently required in time-consuming service provisioning cycles. 

GMPLS is better suited for deployment in a multicarrier, multi-vendor, multisystem network architecture. It provides for routing, signaling and management protocols that address the above-mentioned shortcomings.

Intelligent Optical Networking

Intelligence in an optical network is defined as the ability to offload to the network itself many of the processes that are currently carried out manually. These include: network topology discovery, bandwidth assignment, inventory control and increases in QoS granularity. Considering that user traffic usually travels through distinct segments (access, metro and long-haul) with a multitude of multi-vendor equipment, the provisioning of an end-to-end service becomes a major undertaking. 

The intelligence to accomplish these tasks has been available for some time in packet-based networks via a suite of routing (OSPF, IS-IS) and signaling protocols of MPLS [RSVP and CR-LDP]. Within the past year, the International Engineering Task Force (IETF) has extended these protocols to include devices that switch in time, wavelength and space domains via GMPLS. 

Essentially, GMPLS is intended to provide a single method for controlling devices such as digital cross-connect systems (DCS), emerging optical cross-connects (OXC), DWDM/ADM and to seamlessly operate with packet-based networks. GMPLS-based networks can find an optimal path (based on end-user requirements), for a traffic flow that potentially starts on an IP network, is then transported by Sonet and then switched through via a specific wavelength on a specific physical fiber (see Figure 1: End-to-End Flow Provisioning with G-MPLS).

Describing GMPLS

The basic objectives of MPLS are to accelerate the packet-forwarding scheme and to provide for traffic engineering in IP networks. To do this, the connection-less operation of IP networks becomes more like a connection-oriented network where the path between the source and the destination is pre-calculated, based on user specifics. To accelerate the forwarding scheme, labels are used to “map” an input port to an output port rather than relying on address matching schemes. To provide traffic engineering, forwarding tables, that represent different levels of QoS that the network can support, are used. 

The tables and the labels are used together to provide an end-to-end traffic-engineered path called a label switched path (LSP). Signaling and routing protocols are used to establish these paths. Traditional IP routing protocols (OSPF, ISIS) and extensions to existing signaling protocol, resource reservation protocol with traffic engineering (RSVP-TE) and/or new signaling protocol constraint-based routing label distribution protocol (CR-LDP) are also used.

Extending the concepts of MPLS to networks that are inherently connection-oriented becomes an exercise in the calculation and management of optimal end-to-end routes. This is a natural fit for optical networks that will be expected to carry the increasing volume growth in data traffic. 

To do so, routing and signaling protocols are needed and a protocol is required to address the following unique issues inherent to optical networks:

  • Localization and resolution of signal problems at the optical level;

  • Ensuring the level of protection required by the application;

  • Ensuring the health of the link that carries the control (signaling – GMPLS – RSVP-TE, GMPLS – CR-LDP and routing GMPLS – OSPF-TE, GMPLS-ISIS - TE) information; and

  • Correlation of link property between the adjacent nodes.

The link management protocol (LMP) is the new protocol that will address these issues.

What's in It for Providers?

So what does all this do for service providers? Why should an operator configure and plan a network with multiple devices that are inherently more complex than their predecessors?  The answers lie in the economic realities and drivers of today’s networks. The basic objectives are:

  • reduction of operational expenditures,

  • increasing  revenues and competitiveness,

  • providing simplicity and efficiency for network operations

  • accelerating time to market through providing “on-demand” services

So how does GMPLS address these objectives? GMPLS introduces the intelligence necessary to minimize the manual intervention currently required in time-consuming service provisioning cycles. End-to-end service provisioning requires coordination among multiple operators, assessing capacity at each network and node, establishing the connection, and, finally, testing the connection. While it is too early to concretely quantify the cost saving measures of GMPLS, industry reports indicate that a provider can expect to save over 70% of operating cost and reduce service delivery time cycle by 95% with an intelligent optical mesh metro network (see chart). 

The savings in service delivery time cycles translates to one day of work instead of 22 days (assuming eight hours of work per day and 22 working days per month on the average). GMPLS provides the intelligence to automate many of the provisioning steps that currently cause delays. The goal is to decrease the cycle time from several months to days or even minutes.

Cost savings also come from resolving disruptions in the network topology when a resource (node, link and amount of available bandwidth) is added or removed. By automating the advertisement (OSPF-TE, ISIS-TE) of the availability or absence of a resource throughout the network, the network not only reaches a state of equilibrium faster, it is also less prone to human error.

Another major enhancement enabled by the implementation of the GMPLS suite of protocols is the opportunity to provide new, sophisticated services to the customer. Data storage, video streaming, virtual private networks (VPNs) for enterprises are some of these examples. Many of these new services will be data-centric and consequently, bursty, and non-deterministic in nature, which may not necessarily require the level of protection that a voice application must have. 

The operator must have the tools to assign a more granular level of QoS to client services for more flexibility. Specifically, a GMPLS-capable, end-to-end network will allow the operator to automate path provisioning that includes traffic parameters for differentiated services or guaranteed QoS. These parameters are carried by signaling protocols (GMPLS-RSVP-TE or CR-LDP) at the time of connection establishment and can be based on network congestion, protection level, bandwidth use or other routing policies.

With the ability to offer customized level of service a carrier may elect to structure different grades of services, for new sources of revenue (see table at left). 

Grades of service
Service Protection
Platinum Very high reliability (50 milliseconds)
Gold Reroute upon failure
Silver Interruptible (provisioned on working bandwidth)
Bronze Interruptible (provisioned on protection bandwidth)

Conclusion

Given the increasing level and variety of protocols (Sonet, IP and ATM) carried by optical networks, and given that wavelengths are becoming the basic unit of transport, the operator requires new tools to effectively manage the next generation optical network. Even though GMPLS-based network face challenges such as multi-vendor interoperability and interworking with legacy networks, GMPLS has been recognized as an important tool in advancing the evolution of optical networks. GMPLS holds the key ingredients for a path to continued operational cost savings while creating the opportunity for new sources of revenue.
Robert Haim is Senior Product Marketing Manager for Polaris Networks, Inc.


Visit Polaris Networks online.

Want to use this article? Click here for options!
© 2012 Penton Media Inc.

Learning Library

Featured Content

A time and money saving approach to fiber deployment

Service providers are under tremendous pressure to turn up new services faster then before and, at the same time, to do it at less expense - and intra-office fiber is one of the biggest challenges in terms of both cost and service turn-up.

The Latest

News

From the Blog

Briefingroom

Join the Discussion

Resources

Get more out of Connected Planet by visiting our related resources below:

Connected Planet highlights the next generation of service providers, as well as how their customers use services in new ways.

Subscribe Now

Back to Top