1. Large Events
Large events, with thousands of visitors and their mobile devices, such as festivals, football matches, Olympic Games and the like, nowadays create both offline and online communities. In fact these online communities give shape to a second event – the virtual counterpart – and represent one of the most significant motivating factors for subsequent activity on social media. The willingness to share, contribute and interact on both sides, though this can be restrained to such things as related to the event. Hitherto, most participants belong to only one of the communities but with increasing significance more and more belong to both. In the context of BASMATI, we propose the creation of a virtual storage and computation layer, called the Virtual Event, which realizes and makes available a shared storage, computation, and network platform to any on-site device as an ad-hoc micro cloud. This approach will enable the realization of contextualized collective applications, whose purposes include entertainment, business and social support. For example, users can read and contribute to a live event guide, create complex mosaics using mobile devices displays, have access to emergency support and localization services.
Challenges and Scenarios
In the context of a large event, every device and their associated user is handled as an isolated node and the enabling technical services depend heavily upon the available connectivity to the network, the service provider and their account. Currently, it is not possible to mirror real world phenomena with thousands of people performing and interacting together, singing, dancing and celebrating as a crowd and contributing to the social experience of all. Creating the virtual and online equivalent of this kind of ad-hoc community is currently very inefficient with conventional communication, computation and storage technologies. The quality of the connection between the two communities is not as extensive as it could be due to purely technical reasons. Each individual contribution has to pass through a centralized infrastructure and, if shared or requested, redelivered to the devices. This requires each device to have strong connectivity to the mobile network and a shared workspace is missing. This creates numerous secondary issues, latency for the individual, with people being excluded when losing mobile connectivity and consequently many services that rely upon shared data cannot be effectively provided. Moreover digital information exchange between infrastructure and visitors is difficult to establish in crowded areas.
How BASMATI would help
With the creation of the Virtual Event, mobile devices on-site the event / festival represent a shared decentralized and dynamic information space, by acting as virtual service provisioning platform. To that end, the brokerage platform of BASMATI that, by means of the BASMATI hybrid infrastructure, is able to seamlessly exploit both cloud resources and mobile devices, will determine the assignment of the necessary resources, in the context of a highly dynamic and ephemeral network. The BASMATI platform also takes care of QoS of the services provided (especially in terms of latency) by means of the advanced offloading and application reconfiguration techniques. The Virtual Event use case also benefits of the BASMATI advanced tools for user modelling and situational knowledge acquisition for deciding in a proactive way where and when allocate computational and storage resources.
2. Mobile Virtual Desktop
Recently, many companies and government organizations have begun to exploit smart work environments by exploiting Virtual Desktop Infrastructure (VDI) services in order to reduce the cost of the management and static hardware infrastructure. A VDI service may include company-specific business applications with high level security, software to handle various document formats, and even favourite personal applications. A VDI usually runs remote desktop operating systems and applications inside Virtual Machines (VM) that are hosted in the Cloud. VDI provides virtual desktops to users as a service in the cloud environment and enables the users to access their virtual desktop anywhere and anytime. A common VDI system is composed by three core entities: (i) the VDI client, such as a thin clients, a zero clients or a PC clients, (ii) the VDI server in which user’s virtual desktop are executed and stored, and (iii) the VDI manager that orchestrates the overall virtual desktop system.
Challenges and Scenarios
VDI services are often used by employees during business travel. Let us consider the case in which a Korean employee has a business trip to an European country, but still prefers to use their national VDI service to continuously perform business work, to access to specific data or to enjoy favourite personal applications.
In such conditions (i.e. cross bordering), the QoS enforcement of the VDI service is very challenging, due to the interactive nature of the application coupled with the degraded response time and performance caused by long distance data transfer. In this context, an advanced brokerage platform is needed to support the mobile virtual desktop on top of a federated Cloud environment in order to enhance the degraded QoS of VDI service when the user moves to another place, e.g. border-across movement. In this scenario, mobile VDI services can be freely transferred among federated Cloud , and adapting their behaviour in terms according to the users location, This approach allows the user to receive a more stable and effective VDI service anywhere regardless of their location.
How BASMATI would help
In this context, the federated Cloud environment of BASMATI helps to enhance the QoS associated with the VDI services. In particular, this use case exploits the cross-border brokerage ability of BASMATI to relocate the VDI services from Korea to EU, by exploring the tradeoff cost/performance. This approach provides users with a more stable and high-quality VDI service regardless of their location.
TripBuilder is a mobile service helping tourists to build their own personalized sightseeing tour of a city that they don’t know. Given a targeted touristic area, the time available for the visit, and the tourist’s profile, TripBuilder provides its users with a time-budgeted tour that maximizes a tourist’s interests and takes into account both the time needed to enjoy the attractions and the time required to move from one Point of Interest (PoI) to the next one. For the sightseeing tour generation TripBuilder exploits publicly available sources (Linked Open Data) to build a knowledge base and infer user interests. The sources currently exploited include Wikipedia, Flickr and Google Maps. The sightseeing tour generation relies on Clouds allowing to harness the required computing power whenever the service needs to update its knowledge bases, or when a large number of users from different countries congregate in a same area.
Challenges and Scenarios
TripBuilder leverages established technologies for parallel computing (Apache Spark, Storm). Providing service to end-users requires generating two logically separate knowledge bases (KB), the area-KB and the user-KB, as well as solving a rather heavy computational tasks in terms of the KBs and of the user query (optimal tours are obtained by an approximated algorithm for the actual NP-hard problem). The computation associated with the area-KB is network and memory intensive, and possibly computation intensive. It is a typical batch mode task for each target area. The user-KB can be trivial (manually specified preferences), or it can be pre-computed from user-provided information. From example, the user-KB can be obtained from static sources like user’s Flicker pictures, or from the context of the user, or by considering a combination of the two. The computation of user-KB may happen either on the mobile device itself or by offloading to the BASMATI Cloud infrastructure (in general independent from the Tripbuilder services), allowing for a flexible tradeoff between performance and privacy protection. The final tour generation is compute intensive and also has near real time constraints.
TripBuilder presents several challenges when supporting a large number of areas from one or more servers (compute and storage scalability), a very large number of users planning their trips in the same city (server scalability, server replication and dynamic load behaviour) and users moving to different cities, thus requiring service handover and load balance across different servers (monitoring and short-term performance prediction, situation-aware resource brokering). Different geo-location and legal constraints can also affect the brokering and application placement choices.
How BASMATI would help
BASMATI features improve TripBuilder’s back-end computations. TripBuilder represents a typical application for people travelling “across regional borders” and the BASMATI Cloud federation helps in relocating the services to different cloud providers. Since TripBuilder explicitly leverages the “wisdom of the crowd” models, it is highly relevant to react whenever a large number of people concentrates in a specific area. In this context the D2D capabilities of BASMATI can orchestrate the offloading process to the cloud infrastructure as well as merge the flow of information about POI coming from the multitude of local devices with the information already presented in the cloud backend. In addition, BASMATI helps in configuring touristic and business tours according to the changing of conditions in the context of the users, thanks to the flexible relocation model of services and the ability to scale the trajectories reconfiguration.