Discussion Papers

Franz Büllingen, Annette Hillebrand, Peter Stamm:

Recent developments in Cloud Computing services and their impacts on the development of network traffic

No. 385/April 2014

(Full version only available in German)

Summary

The present study aims at the analysis of recent market developments in Cloud Computing services and their impacts on IT-network infrastructure with special regard to increasing volumes of traffic. Within the last years it has become obvious that Cloud Computing offers a new market model to provide IT-related services. This paradigm shift implies that e.g. computing power, software, storage as well as managing services can be accessed any time by any user and that they can be billed individually due to the scope of services, the duration or the number of users. On top a broad variety of IT-specific applications, processes and functions can be delivered by Service Orientated Architectures (SOA).

Regarding the long lasting and intensive discussion on the availability of Broadband Internet and its bandwidth provided in Germany, it is quite an astonishing fact, that this discussion looks to be com-pletely unaffected by the question whether Cloud Computing services can be delivered without being restricted by the existing network related bottlenecks. In fact many experts expect that Cloud Computing might exceed existing network capacities leading to a lack of performance of any Internet-based services.

Our analysis pinpoints that Cloud Computing services seem to be in a very early stage. It is shaped by extremely dynamic market forces, which means, that no potential developments can be precluded. In principle Cloud Computing encloses all kind of network based services and predetermines the road to a comprehensive industrialization of IT. The related market growth for Cloud Computing is regarded highly dynamic. Due to the estimations of Gardner the worldwide turnover will exceed 150 billion US$ in 2013.

It becomes obvious, that in the future any network based services and applications will organized and delivered over Cloud Computing platforms. Particularly datacenters will generate and terminate vast amounts of IP traffic. Therefore it will be hardly possible to discriminate between internet traffic originated by Cloud Computing and internet traffic generated by "normal" applications. Rather it becomes apparent that recent models of deployment and Cloud Computing services are shaped by economies of scale, which means, that any kind of IT-services including telecommunications services like telephony, fax or email will be offered by Cloud-based single-hand services.

Due to the interviewed experts it can be concluded, that Cloud Computing will, to a large amount, boost internet traffic and it will reinforce symmetric traffic induced by any kind of up- and downloads. The future behaviour of spontaneously using Cloud Computing services any time and any place by private and business customers will promote patterns which can be characterized as burst traffic.

Discussion Paper is available for download.

To top  |  Print