|Issue:||North America 2008|
|Topic:||Converged networks – just a pretty name?|
|Author:||Yossi Ben Harosh|
|Title:||President and CEO|
Yossi Ben Harosh is the President and CEO of Telrad Networks. Mr Ben-Harosh has two decades of experience in telecommunications, operations and management, including over 16 years at Telrad. Prior to serving as CEO, he was Division President of Global Operations, responsible for all Telrad Operations Infrastructure and R&D of OEM/ODM products and solutions. Previously, Mr Ben-Harosh acted as Operations Manager in the areas of Global Supply Chain Management, New Product Introduction & Manufacturing, R&D Engineering Support and System Integration Fulfilment. Prior to joining Telrad, Mr Ben-Harosh was a project manager at TEVA Pharmaceutical Industries Ltd. Mr Ben-Harosh holds an MBA from Bar Ilan University and a BSc in Industrial Engineering & Management from Ben Gurion University.
IP-based networks are gradually replacing digital switches, but TDM equipment installations are still growing at incumbents. New operators are building converged IP networks to offer the latest services, but most ‘converged networks’ are built with co-existing legacy and next-generation equipment since it takes as long as five years to changeover a network and the devices that use it. New converged networks will make it possible to provide, and charge for, variable quality-on-demand in accordance with the needs of each service.
Technical discussions abound in the telecommunications industry about the most effective technologies for different networks. By and large, fixed line incumbents are making gradual changes from digital switched architecture to IP-based architecture. The need to replace equipment that has reached end-of-life and the need to increase capacity drive upgrades of the operator’s core network equipment. Despite grandiose plans regarding IMS and other IP-based schemes, in the field, TDM equipment installations are still growing. At the other extreme, in much of the world deregulation and low equipment costs are enabling new operators. The new operators can use the newest technologies, in fact skipping generations of equipment development, and bypassing incumbents regarding the kinds of services they offer. Networks are still fragmented This rush is putting pressure on incumbents to provide competitive pricing, applications, and services. However, the technological challenges are overwhelming. Many customers purchase a single bundle, but still get separate bills for data, voice, and other communications services. This reflects the true complexity of converging these networks. In fact, the terminology ‘converged network’ is basically a euphemism for a mashed-up network with legacy and next generation equipment co-existing. When operators talk about upgrading to a 3G network, for example, it is not just the core that needs an upgrade. In fact, each base station needs upgrading and, in practical terms, it takes time to change every base station in a network. The same goes for today’s Metro Ethernet fixed line networks. When looking at the number of installations required, it is clear that any change has to be gradual. Furthermore, even the phrase, ‘the operator’s network’, is a misnomer. In fact, the physical infrastructure may have one or more owners; the operator might only have core equipment and end customers, having leased the infrastructure from the owners. Much of today’s wireless backhaul uses microwave to reduce the costs of leasing lines. So it is obvious that many different technologies are involved in most communications today. Consumer equipment replacement One of the issues that gets little attention, but is obviously of huge importance, is the issue of end-point devices – the handsets, telephones and computers used for communication. Today, there are still very few IP enabled telephones in use. In large corporations, the move to IP telephony can justify the replacement of all the end-user equipment. However, for consumers, this is going to happen a lot more slowly than a core upgrade. Consumers have gotten accustomed to the fact that they need to have an electrical line connected to their phones; however having an IP input as well is just another complication as far as consumers are concerned. Likewise, with mobile handsets, the typical lifetime of a handset is 2-4 years, and mass replacement is expensive and impractical. No matter how quickly an operator wishes to upgrade, even with proven technology, a complete transition – from the network core to the consumer device – takes a minimum of five years. Quality of service When looking at the network infrastructure, one of the major issues is how to manage the quality of service on the ever-expanding networks. Under the legacy time-division protocols, quality of service was guaranteed. Likewise, the legacy video broadcast networks are able to provide high-quality, real-time video with very few glitches. However, as networks expand, move to IP-based protocols, and become mobile, providing quality of service becomes the major issue. While end-users can tolerate delay and poor quality on some kinds of media, for example, humorous video clips, there are areas where glitches in quality are simply unacceptable, for example, for a stock brokerage following real-time stock prices. Until now, the demand for quality of service used to be associated with enterprise customers who are willing to pay, such as financial organizations needing real-time, highly accurate, tracking of all of their financial movements. Distributed organizations, with a deep understanding of their own needs for security and quality were the main customers in mind when discussing how quality of service would look going into the future. However, a major change is taking place; typical end users are increasingly concerned about the quality of service they need, will want to be able to control it. For example, a subscriber viewing a soccer game might be willing to pay for high-quality video broadcast service for the 90 minutes of the game, but not at other times of the day. On the one hand, the consumer’s changing needs give operators opportunities for new revenue streams, even at the simplest ‘bit-pipe’ level of service. However, a highly granular service provisioning resolution creates network management challenges. First, service providers need to be able to provide quality of service on an individual basis, for specific services or specific periods of time. In order to be manageable, the service has to offer an easy-to-use self-provisioning interface, so that the user can control the experience. It is not practical to offer this kind of service unless it is automated. Wonks talking bits and bytes One of the major debates regarding quality of service in converged networks revolves around the protocols used. Operators, vendors and the media have been debating the benefits of the existing MPLS networks, and the limitations of MPLS (Multi Protocol Label Switching) when it comes to managing a large number of IP addresses or Virtual LANs (VLANs). MPLS is a natural choice for many operators, because they already have installed bases of MPLS networks and routers. Furthermore, many operators have already developed management systems that allow them to overcome the limitations of MPLS, albeit in an expensive and relatively complex way. However, once those systems are in place, it makes little sense to replace them. Other operators are pushing for a hybrid approach, with MPLS at the core of the IP network, but towards the edge of the network. New protocols, predominantly PBB-TE (PBT) [PBT refers to a package of Ethernet enhancements defined by the IEEE’s Provider Backbone Bridging Traffic Engineering (PBBTE)], are used to provide an easier management paradigm for handling of large numbers of users. PBT uses a methodology known as MAC-in-MAC, which includes a dual-level of MAC (Media Access Control) addressing. One MAC address is the unique MAC of the end point on the edge network, and the second level of MAC address is unique only within the sub-network. As the packets travel between networks, the external MAC address is stripped off, and a new external MAC address for the next network is added. The figure below illustrates this dual-addressing scheme. Hybrid networks by any other name While the infrastructure wonks discuss in detail the different types of addressing schemes and management systems that will make the most sense in the future, most neutral experts admit that the protocols are going to live side-by-side for some time. In fact, even the TDM infrastructure is still on the rise, despite the move to IP in the networks. The bottom line is that the pure economics of the networks are going to dictate the future. Within that framework, operators are hesitant to replace existing infrastructure, especially where the infrastructure is perfectly sufficient for the current network needs. Therefore, legacy and IP networks will demand interoperability, as will PBT and MPLS networks. Within any given network, a variety of different types of equipment will co-exist. While the two networks may not provide unified management or transparent interoperability, data will surely pass through a vast sea of different types of equipment without having any impact on the users. From the consumer’s point of view, none of this is particularly relevant. Users will go wherever they get the best service and the best value for their money. Infrastructure issues will remain the domain of the vendors and operators to hash out, while consumers reap the benefits of these new technologies.