Cloud computing and telecom networks
Eurescom study P1951
Eurescom study P1951 has analysed the technical and business implications of cloud services for the telecom networking business. Cloud computing, as a service delivery model, is impacting the whole computing and communication industry. This model can be considered a consequence of the long-held dream of computing as a utility. It is associated to Software-as-a-Service, Infrastructure-as-a-Service and Platform-as-a-Service as computing service market segments, and to some extent to Network-as-a-Service as an emerging network service model.
In the long term, cloud computing will lead to the era of dematerialized computing systems. Only some form of access peripherals and related subsystems will be left, with capabilities, role and usage under definition. As integrators, operators and users of large-scale distributed computing systems, telcos with their networks and data centres are directly concerned by changes arising from cloud service delivery.
Emergence and implications of cloud computing
Cloud computing relies heavily on long-term developments in high performance computing (HPC) technologies (multicore, GPU, modern cluster servers and supercomputers), high-speed networks, high-capacity storage systems and grid computing. These developments are crucial for the availability and aggregation of resources in order to realise the idea of large-scale resource sharing and on-demand usage promoted by computing as a utility. Cloud computing also relies on the success of open-source middleware and service oriented architecture in conjunction with recent generations of grid middleware platforms, e.g. Globus Toolkit 4, industrial developments of virtualization platforms such as VMware or Xen, and recent developments in autonomics and systems for the deployment and automated management of large data centres. Therefore, from a technical perspective, it is important for any organization involved in the operation of large-scale networks, data-centres and cloud services to master all these technologies.
The future of cloud computing relies on synergies between all these technical computing areas, and almost any applications domain involving computable processes. In particular, it will be possible to enhance existing IT and multimedia network applications and to investigate future network applications. This is especially important for critical computing and scientific simulation domains: emergencies, disaster recovery, stabilization of real-time global financial transaction, in-silico research in energy, bio-medical science, climate and meteorology, or complex industrial processes.
Cloud computing is the future of IT and networks. More and more corporate IT executives share this view with researchers and engineers as well as an increasing set of organisations (e.g. government agencies, research institutions or standards bodies) . For most of these organisations, the adoption process is likely to start with private clouds, and then hybrid and public clouds. On the other hand, SMEs with varying resource requirements and limited IT budget are subscribing to public cloud services, e.g. Amazon EC2, Google Apps, Microsoft Azure, and others, and hybrid clouds combining resources from relatively small scale internal data centres and public clouds. In order to take full advantage of public clouds and the synergetic features of cloud computing technologies, important research and standardization are required, in particular to ensure the dependability and the management of public clouds, openness end interoperability, global IT and networking jurisdiction and data privacy. Distributed computing assets operated by Telcos (core networks, attached data-centres and access infrastructures) and other data network operators will need to evolve to adapt to cloud computing requirements, which include basically virtualization, performance and scalability.
Expected role of telcos – recommendations
Our analysis of the cloud service environment shows that telcos are set to play a major role in cloud computing, and the positive factors of the new technical and business environment outweigh the negative ones. The environment appears to be largely in favour of telcos, which have the potential to play a key role in the operations of cloud infrastructures, to add value to existing telecom services, and explore new services. In fact, global network operators are today the only companies that have the technical skill, capital and operational experience and processes in place to run public and hybrid cloud infrastructures. This places telcos in a unique position in cloud service delivery.
However, in order to make it happen, Telcos need to reinforce their position by conducting research to enhance technical capabilities of their equipment. Telcos also need to initiate and support standardization efforts especially to promote open and interoperable clouds, as it has been the case for OSI and – to some extent – in case of the Reference Model of Open Distributed Processing systems. Furthermore, telcos need to upgrade their operational networks and data-centres, e.g. deploy very high-speed networks, virtualise and manage deep computing capacity grade data centres with virtual supercomputers to support distributed cloud services. New businesses can emerge from cloud service delivery, especially from synergies with the High-Performance Computing (HPC) community and virtually all industries and computing services requiring important resources and elasticity. There are some potentially huge opportunities in terms of revenues, research and industrial development, especially in vertical markets. For example, telcos can use their expertise of global network and service management combined with cloud services to stimulate research and industrial development of new techniques for the global management of telecom and cloud services, for example, based on network analytics. Telcos could secure cloud services with more holistic approaches to security attacks and ensure the dependability of clouds.
Cloud computing will most certainly have a deep impact on IT and networking, corporate IT and the telco business in particular. But the industry and the research community need to undertake several actions to support this development and take full advantage of it. Research efforts are needed regarding cloud service security, data location and privacy issues, large data-centre management and automation, software licensing, programming tools and techniques for high-performance distributed computing in virtualized environments, open and interoperable clouds, and more. Telcos have most of the technical know-how, capital and operational environments and processes in place that can be converted to support cloud services. However, in some cases telcos need to adapt their networks and data centres. For some specific properties, e.g. exploiting high-speed optical links, scalability and latency reduction in data-centres, telcos need to establish collaborations with the networking and distributed computing research communities, perform research and develop novel standards to enable the coverage of the full spectrum of cloud services. The issue of end-to-end latency reduction in particular seems important for critical applications and high-performance computing, and thus need further investigation.
Further information is available at http://www.eurescom.eu/Public/Projects/P1900-series/P1951. The confidential study results are only available to members of the Eurescom study programme.
Please send us your comments on this article.