Next Generation Networks
From Blindside
Contents |
[edit] What is it
The exact technical definition of "Next Generation Network" is still developing. Generally, it's used to describe emerging telecommunications networks that, like the Internet, rely on packet switching rather than circuit switching but that incorporate the services of traditional telephone networks. Such networks separate the physical transport layer across which data travels from the applications or services layer that present the data as for example voice or video to the user. Like the Internet, NGNs treat all types of traffic - voice, video, data – as simply data. However, unlike the Internet, NGNs may be capable of guaranteeing quality of service, and ensure that each traffic type receives an appropriate class of service. There is still no clear consensus amongst Internet technologists whether such end-to-end quality of service guarantees are required for the great majority of Internet traffic.
The International Telecommunications Union defined NGN this way in 2005: "Next Generation Network (NGN): a packet-based network able to provide telecommunication services and able to make use of multiple broadband, QoS-enabled transport technologies and in which service-related functions are independent from underlying transport-related technologies. It enables unfettered access for users to networks and to competing service providers and/or services of their choice. It supports generalized mobility which will allow consistent and ubiquitous provision of services to users."
There are strong economic and practical reasons behind the development of converged IP networks. The biggest traditional telephone companies have significant problems: the high cost of legacy network designs and the extraordinary expansion of digital traffic, fuelled in part by increasing demand for new mobile and multimedia services. British Telecom, for example, whose 21CN project makes it the first major telco to convert its entire network to Internet Protocol, has not one legacy network but six; voice and ISDN are in fact separate networks because in the analog past it was simpler and cheaper to build a new network for each new service. Converting the network changes everything: all traffic, whatever its nature, can be transported using the same hardware infrastructure, and how traffic is treated can be handled in software. Before the transformation began, BT also had 4,000 operations and business support systems that it expects to be able to shrink to 500. The company will spend £10 billion on transforming its network and expects to save £1 billion a year once the change is complete.
Traditionally, telecommunications architectures relied on circuit switching; when a user's call was connected a circuit was opened between caller and receiver that was dedicated to their sole use. Internet protocols, however, divide data into packets which may be routed individually between caller and receiver and reassembled on arrival. The upshot is that many data streams can use the same bandwith simultaneously, similar to the difference between sending a package to its destination in a taxi versus sending that same package via the Post Office or a courier, where it shares a truck with many other packages going to nearby destinations.
Legacy services may be either replaced by counterparts using new technology (such as moving from ISDN to symmetrical broadband) or reengineered technically on the new network with no difference visible to the customer (such as moving from traditional voice telephony to VOIP). However, different types of traffic (voice, video, email) require different treatment to satisfy customer expectations. A few seconds' delay in the arrival of an email packet will be unnoticeable to the user, but a few seconds' lag in a phone call will be a source of huge frustration. On the other hand, just one lost packet will kill an email message but be tolerable for a TV broadcast. MPLS solves this by assigning priorities to different types of traffic, making it possible to guarantee quality of service for the services that need it. Many other mechanisms are available for dealing with these requirements, from the basic reliable transmission of TCP to Forward Error Correction and loss-tolerant codecs for the transmission of audio and video.
Instead of building a new network, as in the past, turning on a new service will be a simple matter of defining it the controlling software. The upshot should be vastly increased flexibility and responsiveness to developing trends as well as new opportunities for third-party companies such as ISPs, VOIP providers, and others who may be able to offer competing services.
[edit] Impact & Maturity assessment
Impact: 3
Maturity: 2
[edit] Information Assurance issues
NGNs are entirely controlled by software. These networks are therefore subject to the same risks and problems that come with all software: bugs, bungled upgrades, unforeseen security holes, and so on.
Because NGNs incorporate many services and many different types of users, the information assurance issues posed by them are extremely complex. Compared to traditional point-to-point communications, many of the requirements for NGNs are expensive: dynamically authenticating mobile users, managing multiple, simultaneous layers of security, and maintaining high data rates despite encryption. In addition, computing equipment is not in general manufactured to the same engineering standards of traditional public switched telephone networks (PSTNs) and therefore may not be able to match their standards of reliability.
Critics of the concept also point out that since NGNs are based on Internet protocols, they are likely to suffer from the same problems already familiar from the open Internet: denial-of-service attacks, session hijacking, viruses, and the many other forms of network abuse. NGNs are (or will be) more tightly controlled than the open Internet, but because these networks are so new it's unclear whether and which third parties might develop and run applications on them. It is therefore difficult to prepare for possible malicious applications. NGNs' distributed nature, in turn, suggests that it may be difficult to locate and eliminate problems that emerge, especially when the sources of disturbance are able to replicate (as viruses do). The newness of these networks means there is a lack of trained professionals to deal with the new technology and its problems. In addition, the industry that produces the many types of malware is increasingly commercial and professional, making the threat it poses far more substantial than that of the archetypal joyriding teenaged hacker that plagued the early days of the Internet. The shift in source also means that where yesterday's exploits were designed for bragging rights, today's are designed to be as stealthy and undetectable as possible.
Therefore, mission-critical applications and high-profit services such as ecommerce and financial transactions will be running on a vastly more complicated network, with a multi-layered, packet-based infrastructure, an open, distributed architecture, and using technology designed with no ingrained security mechanisms. Experience on the Internet has already shown that user identification carried out at the IP layer can be easily tampered with, spoofed, or used to commit fraud and only discovered by legitimate users once the abuser is long gone.
There is a risk that these systems are being deployed before they are fully mature. However, BT began piloting 21CN in Cardiff and northern Scotland in late 2006, so far with success.
An additional complication is that the law in many countries – CALEA in the US, SORM in CIS countries, RIPA in UK – requires telecommunications providers to incorporate technology that makes it possible for law enforcement to intercept communications, a simple matter on PSTNs.
[edit] Implications for UK Government
[edit] Timescale
NGN is one of the most rapidly emerging technology worldwide. BT expects the rollout of 21CN to be complete by 2011, and the impact of this technology is likely to be felt over the next 25 years.
[edit] Examples
ntl:Telewest plans next-generation services
[edit] Comments (attributed)
“In the 21st century, it won’t be who has the best hardware, but the best software. That’s where the pressure is for innovation with IP.” Hossein Eslambolchi CTO and President of AT&T Labs
“Since now, IP technology is widely applied and the IP network has become an important part of national information infrastructure, network security becomes a huge concern. In design and orientation of NGN, security should be a very important factor in the consideration of the architecture.” Jiang Lintao is Chief Engineer of China Academy of Telecommunication Research, MII
[edit] Organisations
International Telecommunications Union NGN Global Standards Initiative
[edit] Documents & research papers
Basic NGN Architecture Principles and Issues, by Keith Knightson (PDF), ITU/IETF workshop on NGN, Geneva, May 1-2, 2005.
VoIP streaming over packet-based networks (PDF)
Next Generation Networks - Future arrangements for access and interconnection
Next Generation Network (NGN) Services
NGN security: Next Generation Nightmare (PPT)
Next Generation Network and Reliability (PDF), by Pertti Raatikainen, IPLU Project, Technical Research Centre of Finland (VTT).
Next Generation Network Services (PDF) Telcordia white paper. Experts (academic, practitioner) Ralph Cochrane, leader of 21CN development for BT
Alan Clark, PhD, founder of Telchemy Incorporated
