The Next Generation

How to achieve a smooth transition to the next technological generation

The development of communications networks is a ­continuous process. At intervals of eight to ten years, mobile operators are however faced with a strategic decision. This is the length of the cycle for the introduction of new mobile technologies which has been seen to date. At the beginning of the nineteen-nineties it was the second generation which forced its way into the ­daily life of many people, and thus laid the groundwork for mass market success. The introduction of the third generation at the beginning of this century marked the launch of broadband data services for mobile users. The industry is presently approaching the next junction: a decision is to be made as to which technology standard will succeed in the next – the fourth – generation of mobile technology.

A number of competitors have joined the global race to become the dominant standard. Two representatives are crystallizing as the favorites in the best position to be able to provide commercial services which fulfill the requirements of the fourth generation by 2010: The LTE standard developed by 3GPP as the official successor to UMTS, and the IEEE 802.16e standard, better known as „Mobile WiMAX“, have the best odds at the moment. Without going into the details of the present discussion about these standards’ chances of success, it is interesting to consider which strategies network operators and producers are following in order to cope with the planned transition from their established third generation networks to the approaching fourth generation. This transition seems to differ from previous developments in a number of ways. 

Standardization provides the framework  

In order to assess the strategies, the basic structure of the mobile network must first be examined. The initial impression is of a rather large range of specialized network nodes which provide us with the common voice and data services on the basis of mature hardware and software technology. Some nodes can be separated out into a further functional group. A mobile network can thus be seen as a compilation of several specialized sub-networks which must be coordinated perfectly in order to provide the customers with the best possible quality of service, and to make it possible for the operator to provide these services efficiently. The networking ability of the network is inherent to the system and, in the end, an indication of the system’s efficiency resulting from the clearly defined division of work between the individual (sub-)networks.

A further aspect is that the telecommunications world is characterized by the need to provide services which are internationally compatible and function across different operators. This means that the interoperability of the products launched by different infrastructure producers must be ensured. This demand can only be fulfilled when the strict standardization of all relevant interfaces between network nodes or sub-networks is guaranteed.

The success story of the GSM standard as the worldwide dominant second generation technology provides a perfect illustration of both of these aspects. This standard, developed as a European initiative under the auspices of the original ETSI standard, is now part of the 3GPP standardization movement which includes American, Japanese and Korean representatives in addition to the ETSI members. The 3GPP’s task was and is the identification and evaluation of future mobile broadband technologies in order to then enable the technical and economic standardization of the most suitable solution. The task of the standardization is to lay the foundations to ensure that the required and sufficient interoperability of the products is achieved later, in the implementation phase.

As the name suggests the original reason for founding the 3GPP organization was to define and agree upon a mobile standard for the third generation. The very clearly stated objective was that the standard to be developed was to be able to provide ­mobile data services with bandwidths which were previously only ­possible using wireline technologies. The result was the UMTS standard. The first commercial networks based on this standard were introduced at the beginning of this century.

The timing of the market entry, which coincided with the ­bursting of the so-called Internet bubble, was certainly not optimal. But the maximum bandwidths offered in the first roll-out phase (up to 384 kbit/s to the user and up to 128 kbit/s from the user to the network) were also no real competition at a time when DSL technology was enabling 1-2 Mbit/s in the user’s direction in the fixed network. The result was that the consumer reaction to UMTS was rather reserved, and the telecommunications operators felt the consequences: the billions invested in licenses and network technology didn’t provide the promised returns.  In this situation, the standard did however win out with one major strength: it could and can be extended or upgraded. This capability is not a coincidence, but is the result of very careful standardization work based on strict principles. These include the precise definition of the service characteristics to be provided on the one hand, and on the other implementation guidelines which are kept as generic as possible so that the producers’ later product differentiation is not hindered unnecessarily. Particular attention is paid to the interfaces over which signaling and usage data is transmitted between the different producers’ products following implementation. No room for interpretation is ­allowed in order to ensure that a network of different producers’ nodes operates smoothly.   

Value can be created through the intelligent combination of network technologies   

The UMTS standard provides a good example of the benefits of this standardization philosophy, as it led to a large share of both the circuit and the packet switched domains of the core network being adapted from the GSM standard or its advanced level GPRS/EDGE and applied for use with the new standard through the upgrade of their software. These network nodes are obviously very effective at networking, as proved by their ability to take on tasks beyond the boundaries of the originally defined  network regime.

A further example of the inherent flexibility of the standard is the fact that from the very beginning onwards no single exclusive  technological standard has been defined at the transport network level. The carefully defined encapsulation of the transport and network levels, with clearly implemented service interconnection points, makes it possible in principle to organize all the communications traffic in a UMTS network using either the circuit-oriented ATM technology or the non-switched IP technology. In practice only the ATM-based transport networks were used to start with, as only this more mature technology was available on the market at the time and was able to use algorithms for prioritization and buffering to provide differential treatment of the various traffic classes defined in the UMTS standard during their transportation within the network.

The incredible potential of the IP technology seen in the enterprise networking market, particularly its commercial potential, was already recognized at that time. Experience with the application of different traffic classes in networks of a similar caliber to national mobile networks was however still too rudimentary to be able to guarantee stable operations in that environment. In the meantime this has changed fundamentally. With the introduction of the HSPA technology representing a further evolutionary step, the 3GPP has proved that its concept to date has been correct. It wasn’t just the provision of data rates of up to 14.4 Mbit/s to the user and 5.8 Mbit/s to the network which made this mobile broadband technology into real competition for the technological solutions common in the fixed networks, but the application of the IP technology at the transport level. This forced the majority of network operators to significantly increase their capacity in order to cope with the expected ­increase in data traffic. Here existing or easy to set up IP-based transport infrastructure was used, which had the additional charm of being the most economic solution.   

A decisive point affecting the acceptance of the HSPA technology by the network operators, as well as its rapid introduction into the market, is the fact that it was always conceived as an upgrade and not as a replacement for the existing standards. In practice this means a large saving in terms of investment and operational costs, and also a significant time advantage for its introduction into the market. The commercial success of the HSPA technology, which can now be assumed as given, is thus a good example of how the intelligent combination of existing network technologies with concentrated activity in clearly defined areas can lead to significant technological and economic value added. The standardization process itself is also a good example of what can be achieved when the operator and the producer tow the same line and understand the definition of a technology standard to be a joint task.   

Evolution not revolution

Despite the wide range of options described above, and the ­flexibility included in the early definition of the 3G standard to ensure that the transition from the existing 2G networks was as smooth as possible, the introduction of the 3G networks was, seen from today’s perspective, rather abrupt. This was ­because the third generation networks were set up and operated in ­parallel, alongside the 2G networks. The synergy potential that the standard offered for the networks to network with one ­another was not used to the extent that it might have been. One of the main reasons for this was that this aspect was not very high on the operators’ list of priorities. The steep success curve expected for 3G services, and planned for by all the market ­players, was ­based on the assumption of rapid availability and not on careful integration of the new technology into the existing network. 

There are a number of indications that the transition from the current, improved 3G and the coming 4G network generation will be much less spectacular, and also much smoother, than that between 2G and 3G. All those involved – operators and infrastructure providers – have learnt their lessons and implemented new strategies as a result. These new strategies are often related to the term ‘evolution’. What does this mean in our context?    

The experience gathered during the introduction of the third generation of mobile networks as described above certainly plays a role. For the network operators in particular there was a discrepancy between the economic expectations – fired on even more by the Internet euphoria – and reality, which did not justify the investment costs made. It is therefore fair to claim that the consequence of this sobering experience is that both the standardization of UMTS and its evolutionary forms HSDPA (3GPP Rel. 5), HSUPA (3GPP Rel. 6) and HSPA-Evolution (3GPP Rel. 7), and the definition of the LTE standard (3GPP Rel. 8) are seen mainly in terms of an attempt to secure return on investments which have already been made and which are still to be made.

Before going into the details of these strategies (see Figure 1), we would like to take a closer look at the time factor, as this will play a major role in further developments towards 4G. 

According to present estimates the commercial launch of three 3GPP releases will take place in the next two years. These include not only both future HSPA evolution stages but also the first part of the fourth generation in the form of LTE. As potential users, the network operators are going to have to make a relatively fast decision as to whether and when they want to introduce the related infrastructure products into their networks. At first glance it looks as if this is a major challenge. But then it be­comes clear that the operators can actually gain more operational freedom by implementing product strategies and using network architectures which achieve integration by applying the ability to network.

Bundling operator interests

Network operators have learnt that effective networking within their own interest group is necessary if their voice is to be heard in the standardization committees. Those who should listen are the infrastructure providers, and what they will hear is a request to reflect the operators’ interests in the definition of their product strategies as well as during the implementation phase.

A good indication of the way in this can work is the foundation of the Next Generation Mobile Network Alliance (NGMN) by the big mobile operators China Mobile, NTT Docomo, Sprint Nextel, Vodafone Group, KPN, Orange und T-Mobile. Their stated objective is to bundle operator interests during the standardization phase for the next generation. This doesn’t mean making the choice of which technology to use. The implementation phase is ignored completely. The alliance puts great emphasis on the fact that they don’t intend to make recommendations for or against any of the technology candidates being considered, and it has sworn strict neutrality in this matter. Its objective is far more the identification of criteria which a future mobile standard must fulfill if it is to be a global success.

The white paper which has been produced within the framework of the NGMN initiative* reflects the operators’ motives very clearly. In its vision the objective is given as being the introduction of an integrated network based completely on IP technology. The target architecture should use a flat network hierarchy both to enable coexistence with the 2G and 3G networks, and to allow the smooth migration from these networks to the 4G architecture.

A further important aspect for the network operator is the need for flexibility concerning the time at which the new technology standard is to be introduced. The correct timing for and implementation of the introduction of new services is of strategic importance to the telecommunications provider. However, if the new services can be supported by the existing network technology, and can even co-exist with existing services, then this gives the operator an incredible advantage in that they are free to choose the point in time for market entry flexibly. The ability to introduce services rapidly as demanded, and the reduced risk of making investments into potentially redundant network infrastructure in a poor market environment are two strong arguments in favor of this strategy.  

New producer strategies 

And how do things look for the infrastructure producers? At first glance things look fine. The introduction of a new technology standard provides the opportunity of new turnover with existing customers, and maybe also with new ones. In the past the challenge for the producers was that they not only had to provide a convincing and stable technological solution, but also an aggressive development strategy to be able to achieve market entry as quickly as possible.

These two objectives are not necessarily in line with one another. If the market launch is too early then the price may be poor quality and network instability, leading to negative effects on the company’s profit and loss account which are no less significant than those which would result from the introduction of a perfect technical product too late, and missing the boat in the race for customers. However, the evolutionary transition which is now demanded has changed this situation and provides, as so often, new opportunities but also new challenges.

Generally speaking, producers want to find ways of convincing operators to invest in new, more advanced network technology. The ability to integrate the operator’s installed network technology has always been a strong sales argument. But, particularly in the case of the transition to a new technological generation, the complex market requirements have nevertheless led to the introduction of rather specific product solutions. Aggressive time plans for the launch of new networks have represented a further barrier to the search for and application of solutions which enable more effective interaction between different technological generations.

In the end this will also be a worthwhile option for the providers. If they manage to convince their customers of the sense behind their long term and consistent technology strategy, then improved customer loyalty will be the implicit result. More consistent turnover over a longer period of time will then be able to be booked as a positive secondary effect.

But which measures have to be taken to actually achieve the network described above? For some time now, the producers have been working on production platforms which should make it possible to support a number of different technical standards in parallel, or at least to enable a simple switch between them through the relatively easy application of new software. In the past this has already been possible for individual network nodes and has also been applied in a few isolated cases. The ­challenge is however the application of this method for all network nodes, particularly those in the radio access network, as here there are hardware components which have been optimized to suit a ­specific standard at the air interface. The rapid developments in processor performance do however make it possible to ­achieve complete functionality using software, which should run on ­generic hardware platforms. The consequence is that the network functionality will no longer be defined by the installed hardware, but purely by the software version in use at any point in time.

A further area of activity is the gradual introduction of the IP technology into existing mobile networks. In order to achieve the long term objective of a unified transport network using IP technology the network nodes must be able to support this transport standard. But this doesn’t mean that we have to wait for the introduction of the new mobile standard, the platforms being used for newly installed 3G networks and capacity extensions in existing networks already support the IP protocol. This is vital if the equipment is to be used beyond the market entry of the 4G standard, and thus the return on its investment ensured.

To round off the measures to be taken, the existing network architectures should be gradually adapted to suit the target architecture. The standard is structured to allow this type of optimization as an optional solution, and thus makes it possible for the producers to aim for a consistent, flat network hierarchy at an early stage in the transition. In this way the operator not only profits from the related economic optimization potential quickly, but can also enjoy the flexibility needed to determine the timing of the migration so that this is in line with the over­all business strategy. The difficult part is the coordination of the necessary work on the network with operational business activities.

Reduction to the main principles is the name of the game

This makes it very clear that the demands being made on the design and the operation of mobile networks are changing. They are the subject of a continuous transformation process which has to be very well prepared and implemented in a controlled fashion. The transformation process itself can be characterized by the fact that it, in contrast to earlier transitions to new generations, will be drawn out over a long period of time.

A further demand that this process will make on those involved is that they will have to define the core principles of the transformation at an early stage. These principles include the choice, evaluation and application of a few powerful and scalable technologies, the identification of those parts of the network which should be the starting point for the transformation and from which the new network should gradually develop, and also a time plan which reflects both the availability and maturity of new technologies as well as expected market developments.

Technologically speaking the reduction to a number of important core principles will be reflected in an accompanying notice­able reduction in the number of functional network nodes. Functionalities will be bundled more closely together in a flatter network hierarchy. The application of the widely used and now mature protocol architecture based on the IP technology will, due to its inherent ability to network flexibly, increase the options open to the operators in their search for an optimal transformation solution which suits their specific timing and geographic requirements.

There are a lot of arguments which suggest that the new network resulting from this concept of a few, but powerful principles, will be much better at networking (see Figure 2).

This characteristic will make it possible for the operators to implement a much more gentle transition process than was the case in the past. One could say that the operators and producers have made use of improved networking between and amongst themselves to help ensure the success of the next generation of mobile broadband communications.