Deutsche Telekom

About: Deutsche Telekom is a company organization based out in Welwyn Garden City, United Kingdom. It is known for research contribution in the topics: Telecommunications network & Signal. The organization has 3473 authors who have published 5208 publications receiving 65429 citations. The organization is also known as: DTAG & German Telecom.

Adaptive selection of colour transformations for reversible image compression

Abstract: This paper presents and investigates a new family of reversible low-complexity colour transformations. It shows that, for a reasonably large set of natural images, there is a colour transform which performs better in the context of lossless image compression than the reversible colour transform defined in the JPEG2000 standard, while having only slightly increased complexity. The optimal selection of a colour space for each single image can distinctly decrease the bitrate of the compressed image. A novel approach is proposed, which automatically selects a suitable colour space with negligible loss of performance compared to the optimal selection.

Relating L-resilience and wait-freedom via hitting sets

Abstract: The condition of t-resilience stipulates that an n-process program is only obliged to make progress when at least n - t processes are correct. Put another way, the live sets, the collection of process sets such that progress is required if all the processes in one of these sets are correct, are all sets with at least n - t processes. We show that the ability of arbitrary collection of live sets L to solve distributed tasks is tightly related to the minimum hitting set of L, a minimum cardinality subset of processes that has a non-empty intersection with every live set. Thus, finding the computing power of L is NP-complete. For the special case of colorless tasks that allow participating processes to adopt input or output values of each other, we use a simple simulation to show that a task can be solved L-resiliently if and only if it can be solved (h - 1)-resiliently, where h is the size of the minimum hitting set of L. For general tasks, we characterize L-resilient solvability of tasks with respect to a limited notion of weak solvability: in every execution where all processes in some set in L are correct, outputs must be produced for every process in some (possibly different) participating set in L. Given a task T, we construct another task TL such that T is solvable weakly L-resiliently if and only if TL is solvable weakly wait-free.

Enabling energy-aware mobile data offloading for smartphones through vertical collaboration

Abstract: Searching for mobile data offloading solutions has become topical in recent years. In the Metropolitan Advanced Delivery Network (MADNet) project, we explore the potential of multiple interfaces on smartphones through a collaborative design to enable energy-aware WiFi-based mobile data offloading. We advocate that the vertical collaboration across different wireless access technologies and network vendors is the key to aggregate the power of mobile operators, WiFi service providers, and end-users. Aiming at reducing the energy consumption on smartphones, we propose an energy-aware algorithm supported by our collaborative architecture to avoid offloading to low-throughput WiFi networks which may consume more energy. The evaluation of streaming applications on our prototyped Nokia N900 smartphones demonstrates that we are able to achieve more than 80% energy saving. Our experiment in the wild also shows that our design can tolerate the minor errors of localization, mobility prediction, and offload capacity estimation.

Detecting Vocal Irony

Abstract: We describe a data collection for vocal expression of ironic utterances and anger based on an Android app that was specifically developed for this study. The main aim of the investigation is to find evidence for a non-verbal expression of irony. A data set of 937 utterances was collected and labeled by six listeners for irony and anger. The automatically recognized textual content was labeled for sentiment. We report on experiments to classify ironic utterances based on sentiment and tone-of-voice. Baseline results show that an ironic voice can be detected automatically solely based on acoustic features in 69.3 UAR (unweighted average recall) and anger with 64.1 UAR. The performance drops by about 4% when it is calculated with a leave-one-speaker-out cross validation.

Containers vs Virtual Machines: Choosing the Right Virtualization Technology for Mobile Edge Cloud

Abstract: Mobile Edge Cloud (MEC) is paving the way for 5G networks with respect to latency and computation offloading. To support user mobility, a MEC application should be migrated from one MEC server to another MEC server, which is in close proximity to users. Providing low migration time and low downtime is one of the critical challenges in MEC. Prior studies mainly focused on the service migration in data centers where only the downtime is taken into account. Meanwhile, the total migration time in MEC significantly influences the service latency due to the user's mobility. In this paper, we present a comparative measurement study on service migration in MEC. The evaluation results, which are obtained from commercial edge servers, show the infeasibility of the investigated migration approaches. Even though containerization is undoubtedly a lightweight virtualization technology, it is not mature enough to support the service migration in MEC. Specifically, both VM migration and container migration introduce high migration time, and subsequently result in service degradation for latency-sensitive applications in MEC.