Kalevi Kilkki

Towards a comprehensive framework for QOE and user behavior modelling

While the modeling of QoE has made significant advances over the last couple of years, currently existing models still lack an integration of user behavior aspects and user context factors along with the consideration of appropriate temporal scales. Therefore, the goal of this paper is to present a comprehensive QoE and user behavior model providing a framework which allows joining a multitude of existing modeling approaches under the perspectives of service provider benefit, user well-being and technical system performance. In addition, we discuss the role of a broad range of corresponding influence factors, with a specific emphasis on user and context issues, and illustrate our proposal through a series of related use cases.

Exploring Empirical Rank-Frequency Distributions Longitudinally through a Simple Stochastic Process

The frequent appearance of empirical rank-frequency laws, such as Zipf’s law, in a wide range of domains reinforces the importance of understanding and modeling these laws and rank-frequency distributions in general. In this spirit, we utilize a simple stochastic cascade process to simulate several empirical rank-frequency distributions longitudinally. We focus especially on limiting the process’s complexity to increase accessibility for non-experts in mathematics. The process provides a good fit for many empirical distributions because the stochastic multiplicative nature of the process leads to an often observed concave rank-frequency distribution (on a log-log scale) and the finiteness of the cascade replicates real-world finite size effects. Furthermore, we show that repeated trials of the process can roughly simulate the longitudinal variation of empirical ranks. However, we find that the empirical variation is often less that the average simulated process variation, likely due to longitudinal dependencies in the empirical datasets. Finally, we discuss the process limitations and practical applications.

Modeling the Value of End-to-End Multipath Protocols

Recently, adding multipath capability in the Internet protocol suite has attracted increasing interest. By letting end hosts discover several paths to communicate, end-to-end multipath protocols aim to improve utilization rate of Internet resources. Although many proposals for end-to-end multipath communication exist, they have not reached significant deployment. Since the multipath protocols are mainly designed for open multi-stakeholder environments, understanding their economic impact is important. This paper introduces a model for assessing the value of the end-to-end multipath protocols from the end user perspective. Without a net benefit of the end user, the end-to-end multipath communication only results in the reallocation of costs and benefits in the Internet connectivity market. The model indicates that wireless devices having access to multiple independent access operators via similar or dissimilar access technologies are crucial in achieving end user value out of multipath communication. Initially, the end user value seems higher when the radio interfaces to access operators are active one at a time but later on, along with higher-energy batteries and lower-energy protocols, full benefit of multipath communication can be achieved. The value of multipath protocols depends on the effective path diversity and available capacity on the Internet.

Does network quality matter?: A field study of mobile user satisfaction

Abstract Mobile quality of experience and user satisfaction are growing research topics. However, the relationship between a user’s satisfaction with network quality and the networks real performance in the field remains unexplored. This paper is the first to study both network and non-network predictors of user satisfaction in the wild. We report findings from a large sample (2224 users over 12 months) combining both questionnaires and network measurements. We found that minimum download goodput and device type predict satisfaction with network availability. Whereas for network speed, only download factors predicted satisfaction. We observe that users integrate over many measurements and exhibit a known peak-end effect in their ratings. These results can inform modeling efforts in quality of experience and user satisfaction.

Features as predictors of phone popularity

Applying Bayesian Networks, the study predicts mobile phone popularity with features.Annual predictivity elicits structural breaks which stem from changes in choice criteria.Most of discovered structural breaks relate to phone display, communication and camera.Structural break related to phone manufacturer brand is linked with market turning from hardware to software driven mode. This article analyzes dynamic changes in mobile phone popularity based on phone features. The time period, 2004-2013, selected for the study is interesting because many technical innovations took place molding the mobile phone market dramatically. The study utilizes comprehensive phone model and sales data collected in Finland, combined with temporally ordered probabilistic models to discover the time behavior of predictivity. More precisely, the Tree Augmented Naive Bayes - classification method is adapted to detect those phone characteristics that best predict the annual phone popularity measured as phone model unit sales. Linear regression and the Chow test are used to discover potential trends and structural breaks. The strength of the predictivity is measured as Kullback-Leibler Divergence. This kind of systematic longitudinal analysis highlights patterns, which are otherwise not possible to observe. The study discovered that the operating system is clearly the only feature with an increasing strength in predicting popularity over time. In contrast, sixteen features have structural breaks between 2004 and 2013. Most such breaks are related to the technical evolution of phones: their display, communication, and camera capabilities. Notably, the structural break in 2007-2008 related to the phone manufacturer brand is interpreted as the market turning from hardware to software driven mode, which contributed to Nokias failure with Symbian and Windows operating systems, and to Googles success with a hardware independent operating system Android.

Economic feasibility analysis of seamless multi-homing WAN solution

High-availability wide area network (WAN) connectivity has traditionally been implemented with a single service provider building links with high quality hardware, redundancy, and guarantees, or with an economical but technically limited approach. We are proposing a new approach, Redundant Array of Independent Internet Connections (RAIIC), which could enable same technical flexibility as traditional high-availability services but lower pricing. In this paper, we analyze the current market situation in WAN market and the impact of a new stakeholder, a virtual service provider (VSP), on the market. We consider the possibilities for an existing business transforming into a VSP.

Building social capital with mobile communication services

People may use different kinds of mobile communication services depending on if they are communicating with, for instance, friends, acquaintances or strangers. Thus, in this paper bonding and bridging social capital is studied in the context of two mobile communication services, short message services SMSs and voice calls. In Granovetters terms, bridging social capital refers to communication with weak or absent ties, while bonding social capital refers to communication with strong ties. We find that both SMSs and voice calls are used for bonding and bridging social capital, but SMSs are used more for bonding purposes than voice calls. Furthermore, media multiplexity is more associated with bonding than bridging social capital. We also present a method for studying social capital in the context of other, newer mobile communication services, and present results of a pilot study. The implications of the results are discussed from a number of perspectives including communication research, social network analysis SNA and mobile operators.

Mobile Application Usage Concentration in a Multidevice World

Mobile applications are a ubiquitous part of modern mobile devices. However the concentration of mobile n napplication usage has been primarily studied only in the smartphone context and only at an aggregate level. In n nthis work we examine the app usage concentration of a detailed multidevice panel of US users that includes n nsmartphones, tablets, and personal computers. Thus we study app usage concentration at both an aggregate n nand individual device level and we compare the app usage concentration of different device types. We detail n na variety of novel results. For example we show that the level of app usage concentration is not correlated n nbetween smartphones and tablets of the same user. Thus extrapolation between a userâx80x99s devices might be n ndifficult. Overall, the study results emphasize the importance of a multidevice and multilevel approach.