Insights. 3.1 The insights that will be generated from the environmental data described above include insights relating to:
Insights. Numbrs uses one of the most advanced AI technologies for analysing financial data and helps you manage your finances more easily, quickly and smarter. The AI function analyses all of your transactions. The information is provided “as is”, without warranty of any kind, whether express or implied. Numbrs assumes no responsibility for errors or omissions and will not be liable for any direct, or indirect losses.
Insights. The Insights page describes the overall idea of the Beaconing Project and its most important facts. The page contains the following four sub-pages:
Insights. Where the brand should be The goals for the City may involve a number of elements: cohesive community identity and consistent marketing efforts, collective community conscience, business and resident recruitment/retention, gross receipts, overnight guests and bed-tax collected. Branding influences these goals by influencing expectations and affecting attitudes, thus affecting behavior and usage. The most successful brands establish an emotional – not simply an intellectual – connection. Our insights come from asking a number of thought-provoking questions. What brand “story” does the research tell? What emotional attachments can the brand hold? How does the brand fit into the consumer’s lifestyle? How can the brand be revised to elicit the community’s desired emotional/behavioral responses? It is from these insights that we determine the overall positioning of the brand. Situation Brief: Review of all research findings Blue Sky Meeting: Internal session for developing insights based on significant research patterns and findings .
Insights. The first experiment, on Mobile Journalism, revealed the potential value of media creation and collaboration tools, those tools proactively residing in the FLAME edge network, to assist the light equipped journalist in his fast-responsive creative work. One team did experiments with 360-degree media and virtual reality technologies to illustrate the potential value of, e.g.
Insights. As initially planned in deliverable D5.2, this PMM experiment has been designed to identify relevant technical feasibility aspects and business implications of distribution of media in smart city areas. The actual experimentation occurred in Barcelona has allowed us to derive useful insights for the main business actors of interest, i.e. Media/VoD service provider and Media/VoD technology provider but it allowed also to derive useful insights for the infrastructure owner/operator, as detailed in the following. From a technical perspective, the PMM experiment has emulated the behaviour of a group of people sharing a media streaming service where personal contents are stored and willing to consume these media while moving from a central place (home) into a specific area of the city. To emulate the localisation of the group of people in FLAME coverage, we configured our scale-out function to activate and connect all the placed Origin Server replicas in the three remaining edge cabinets. The implementation in a wider geographical area, e.g. in the city of Barcelona, would require using localisation functions to activate only the server replicas available in the district where users are identified and may require coverage along their walking paths in the area. The dimensioning of the experiment documented in the previous sections shows that the deployed media SF endpoint could not support the coexistence of 3 groups of 4-5 people each, with a high number of blocked streams, due to transcoding on the target media server (origin or replica). The analysis of results shows that the main cause for this behaviour can be the limited amount of computations resources made available for PMM SFEs in the Barcelona platform (2 virtual CPUs and 4GB RAM per media server). In fact, PMM media servers are generally dimensioned with 4-6 CPUs and 8-12 RAM to serve 4-10 users. Moreover, the use of different devices with different form factors (tablets, smartphones with and without FullHD, various form factors for the screens) caused a number of parallel transcoding activated by the media application to adapt the played content to the actual device capabilities. Despite the creation of pre-transcoded versions of the same contents which was aimed to avoid the transcoding event and make use of direct streaming, we experimented recurrent degradation of the overall QoE both in terms of objective parameters and if subjective user “acceptance” of the media service quality when run on FLAM...
Insights. Note: these are preliminary insights based on initial results. This section will be expanded upon when the full results are available. The FLAME network allows to spatially distribute media assets within a city, such that each asset is available near the location where it will be consumed. For a media service provider, this enables localized control of latency and bandwidth to reduce retrieval time. For the media consumer, the media content does not have to be pre-downloaded on to the device but can be streamed gradually as the story unfolds. The FLAME platform offers intelligent resource management capabilities. The scalable nature of the FLAME platform can cope with high user activity, by automatically starting new media content services throughout the city to distribute the load over more instances when latency is reaching critical levels at certain regions. The FLAME platform enables the location-based AR storytelling system to minimize platform utilization (e.g. by turning off or removing a media service) based on predicted local service requirements. The intelligent lifecycle management of localized media services results in more efficient resource utilization. The FLAME platform could be utilized to integrate additional low latency media services to enhance the immersion and interaction experienced with location-based AR stories. For example, high quality 3D renderings could be generated by powerful servers at edge locations and delivered directly to the client mobile device. Another example is to use low latency state synchronization of several devices connecting to the same edge location. This would enable a multiuser shared experience of an interactive story. VALIDATION GAMING EXPERIMENT: AUGMENTED REALITY LOCATION-BASED
Insights. Our evaluation indicates that the FLAME platform is capable of delivering 3D assets quickly and efficiently for video game applications. For larger number of players, the load on the server remained largely the same as for fewer players, which can be accredited to the Opportunistic Multicasting capabilities of the FLAME platform. To benefit from these capabilities, the underlying client code has to be tuned in a way that lets all clients request the same asset simultaneously. Players who were in the game since it started, never observed any delay in asset loading. Players who join mid-game, which may happen in multiplayer games, put an over-proportional high stress on our test system. These were also the only players who observed a short delay while the 3D models were continuously loaded into the game. The main reason for this is that these players have to download all currently used 3D assets as soon as they join the game. This process takes time and the servers cannot profit from the Opportunistic Multicasting, as only one player is requesting these assets at that moment in time. However, as the content was served from the edge of the network, the problem was mitigated. This drawback is not a limitation of the FLAME platform but of the game design and implementation. Finally, we conclude that the FLAME platform provides both high QoE and QoS as measured during our trials. Asset loading went smoothly and efficiently, while the gameplay remained uninhibited by the FLAME platform. Players were immersed in the virtual AR landscape and mostly unaware of 3D assets loading in the background. Opportunistic Multicasting helped at keeping the network load on our server at a minimum and request processing times were low, making the FLAME platform well-suited for urban multiplayer games. While this prototype of Gnome Trader targets the area of Millennium Square in Bristol, UK, and up to 10 players, future efforts could focus on exploring how similar games targeting larger areas, for instance, an entire city, perform on the FLAME platform. Our results indicate that such city-wide multiplayer games may be well-accommodated by the FLAME platform because of its scalability.
Insights d. Testimonials from the targeted audience as specified in section 2.7.1.
