Atxutegi Narbona, Eneko

ABSTRACT
Many years of research have been dedicated toward the definition of sophisticated quality of service (QoS) measurement tools and methods. Nonetheless, the results given to end users by most commonly used online speed measurement tools are still far from being accurate. Among other factors, the reliability of the measurement methods is affected by different static and dynamic constraints of involved network nodes and TCP/IP implementations. The clarification of such uncertainties will help establish the baseline for the definition of any comparable and technically feasible measurement methodology, in terms of measurement periods, number of concurrent connections, and convergence time. This article presents a comprehensive description of the research work and proposes mechanisms, aiming to obtain a full understanding of cross-layer effects during a speed test targeting end users.

ABSTRACT
This paper aims at increasing our understanding of the behaviour of some of the major TCP variants in LTE networks. A simulation study on the behaviour of seven TCP variants is performed. The evolution of the throughput, congestion window and queuing delay are studied for four scenarios with different network loads and flow types. Our measurements show that, in average radio conditions, most variants are able to quickly reach full link utilisation. However, to achieve the same throughput, they create different amounts of queuing delay. On the one hand, loss-based algorithms tend to completely fill the queue, creating huge queuing delays and inducing packet losses. On the other hand, delay-based variants manage to limit the queue size and decrease the amount of packets dropped by the eNodeB, but struggle to reach the maximum throughput in some circumstances.

ABSTRACT
The proper execution of performance tests is of utmost importance for the analysis and evaluation of Internet-related technologies, protocols and deployment strategies. Nowadays, there are plenty of tools available to experimenters ranging from simulation tools, emulation equipment, to small and large scale experimentation testbeds. Each of these performance evaluation frameworks introduces a series of capabilities and drawbacks. Additionally, there is a need for using a common methodology to perform different performance tests and to create comparable outcomes. Seeking for the standardization of reliable and comparable Internet speed measurements and trying to provide end-users with trustworthy measurement tools, different SDOs have standardized both TCP-based and IP-based measurement methodologies. This paper focuses on the applicability of the different measurement frameworks and methodologies in the field of mobile Internet, with special focus on the impact of using TCP over mobile broadband connections. The paper describes a series of experiments over different frameworks and, based on the obtained results, identifies the implications of the different types of performance tests into the evaluation outcomes and states the need for large scale measurements.

ABSTRACT
Mobile internet has been widely adopted and it is expected to rise to almost 4 billion users by 2020. Despite the research effort dedicated to the enhancement of its performance, there still exists a gap in the understanding of how TCP and its many variants work over LTE. To this end, this paper evaluates the extent to which five common TCP variants, CUBIC, NewReno, Westwood+, Illinois, and CAIA Delay Gradient (CDG), are able to utilise available radio resources under hard conditions, such as during start-up and in mobile scenarios at different speeds. The paper suggests that CUBIC, due to its Hybrid Slow-Start mechanism, enters congestion avoidance prematurely, and thus experiences a prolonged start-up phase. As a result, it is unable to efficiently utilise radio resources during shorter transmission sessions. Besides, CUBIC, Illinois and NewReno, i.e., the loss-based TCP implementations, offer better throughput, and are able to better utilise available resources during mobility than Westwood+ and CDG – the delay-based variants do.

ABSTRACT
The evolution of Internet access technologies, together with the wide diversity of customer devices, has led to a complex scenario where measuring basic metrics with accuracy has become a rather complicated task. Although nowadays there are a lot of tools to assess the rate of Internet speed, most of them share neither the methodology nor the infrastructure to produce comparable results.

In this regard, the development of a unified approach to measure the Internet speed would be beneficial for all ICT players. The establishment of such proposal would inspire better confidence in consumers through the provision of precise comparisons, and it would also be very useful to operators, regulators and providers. Towards this aim, the ITU-T has been working on the definition of a unified methodology and measurement framework to assess the rate of Internet speed.

This paper presents a detailed description of the work that is being done at present in the definition of the aforementioned framework.

ABSTRACT
After several years of research towards sophisticated QoS measurement tools and methods, the results given to end-users by most commonly used on-line speed measurement tools are still far from being precise. In order to define a reliable Internet speed measurement methodology for end- users, the impact that the static and dynamic constraints of network nodes and TCP/IP implementations could impose must be first carefully analyzed. Such constraints will deter- mine the measurement methodology to be defined in terms of measurement periods, number of concurrent connections and convergence time by deployment of controlled simula- tion/emulation environments and real world comparisons. This paper presents a detailed description of the works and leaves hints to be followed, aiming to get a full understanding of cross-layer effects during a speed test targeting end-user.