Executive Summary. The number and density of wireless mobile devices today is significant and growing; in 2012, it was estimated that there were 6.8B mobile devices and that there were 128 devices per 100 persons in developed regions. Second, today’s mobile devices have more computational/storage power than the desktop/server processors of the last decade. Mobile devices have traditionally been viewed as “thin clients” or “edge devices” that serve primarily as user-input devices, but that are expected to offload the compute-intensive processing to (non-edge, back-end) servers. With technology advances, there exists the potential now to view mobile devices as “thick clients,” and going even further, to rethink them as “thin servers.” Given the proliferation and enhanced capabilities of mobile devices, it is now a real possibility for a “wireless cloud of nearby smartphones” to pose an interesting-enough collective computational/storage resource. The Hyrax project proposes a novel vision of a hyperlocal edge-cloud, i.e., a computational/storage cloud comprised solely of a collection of nearby wireless edge devices, with the purpose of pooling these devices’ data and processing power to support a new class of proximity-aware applications that benefit the owners of these devices. The premise behind these edge clouds is that all of the constituent nodes are edge (and not server-caliber) computers, and that any and all computation is performed completely within the edge cloud, i.e., there is no offloading/tethering of the computation/data to a non-edge, back-end, traditional-cloud infrastructure. Going a step further, there might even be situations where it is impractical/impossible to offload computation to a back-end cloud because of infrastructure unavailability or bandwidth limitations, e.g., in disaster scenarios, in high user-density settings such as stadiums/concerts. Hyrax has the potential to transform the mobile and cloud landscapes, and to enable applications that cannot exist today without the type of edge-cloud infrastructure that Hyrax will deliver. Intellectual Merit. The goals involve an inter-disciplinary combination of addressing systems, algorithms, empirical evaluation and usability. On the systems side, our goals are to develop the Hyrax hyperlocal edge-cloud infrastructure that is capable of (i) supporting a variety of heterogeneous edge devices, (ii) detecting and tolerating the churn (i.e., the dynamic movement) of edge devices that enter/leave the edge-cloud, (iii) migrating partial computations in the event of migrating edge devices, and (iv) enabling secure remote computations in the event of users dispatching code to be executed on an untrusted device belonging to a different owner, and (v) ensuring that the privacy of users’ data is not compromised when these devices are leveraged as the nodes of an edge-cloud. On the algorithms side, our goals are to develop (vi) a data-centric scheduling algorithm that decomposes a desired crowd-sourced computation onto a set of nearby devices, (vii) a data-centric crowd-sourcing algorithm that takes the partial/complete results from executing at the edge and synthesizes them into the desired final result. On the empirical evaluation side, our goals are to evaluate the Hyrax infrastructure under a variety of conditions, including (viii) conditions of poor/zero connectivity, (ix) different connection protocols, e.g., Wi-Fi and Bluetooth, (x) conditions of low/medium/high density of users, and (xi) different rates of churn (ingress/egress) of the edge nodes, as the device owners move around. On the usability side, our goals are to deploy the Hyrax infrastructure into our three candidate applications, and to evaluate the advantages and disadvantages of the resulting new computation model on the users’ experience. Broad impact. There are new applications that can result if users were able to pool the data and the processing power of their device with those of other devices/users in their proximity. It might enable a single mobile user to obtain alternative/detailed/richer sources of information from her surrounding environment, well beyond what she might discover through her lone device. It might even enable the user to combine her view of data with those of others nearby in order to obtain a more accurate, real-time picture of events that are unfolding. Specifically, through this proposal, we will aim to develop and demonstrate three new compelling applications: crowd-sourced image search on a distributed photo gallery, crowd-sourced user-generated replays inside a stadium, and crowd-sourced real-time road conditions. What makes these applications novel is that (i) we essentially drive all of the processing to the edge, (ii) the data to be processed is essentially created at the edge, (iii) we build on the idea of dynamic collections of wireless devices in proximity with each other, for a temporary period of time, (iv) we favor performing the computation at the edge over sending all of the data to a central server, and (v) we rely on untrusted devices sharing data with each other, including allowing other devices to query them and to dispatch computation remotely and wirelessly to them. These aspects (i)-(v) are fairly unique, and currently, there does not exist any infrastructure to support these proximity-aware crowd-sourced applications reliably, securely and effectively. In particular, for the in-stadium crowd-sourced user-generated replays, we have a commitment to resources from YinzCam, a U.S.-based company that has agreed to working closely with Benfica and Wavecom to bring its commercially proven mobile-video streaming and mobile-replay technology to Portugal. YinzCam will commit its designer, developer and engineering resources (an estimated value of $200,000) to developing the official mobile application for Benfica, and to conducting the necessary site surveys and efforts necessarily for deploying this technology at Estadio da Luz as an initial first deployment. The impact of this is that, working with YinzCam through the Hyrax proposal, Wavecom and Benfica have an unprecedented opportunity to be the first mover and strategic innovator (not just in mobile video streaming, but also in sports technology) in all of Europe. The Hyrax project is designed to create measurable economic impact for Portuguese businesses and a transformative experience for Portuguese sports fans. Alignment with Goals of CMU-Portugal Program. The CMU-Portugal program is focused on the goals of furthering research, education and innovation. In this proposal, we will accomplish the research goals through an ambitious vision for developing a next-generation edge-cloud infrastructure, a brand-new concept that is well beyond the state-of-the-art and that advances the state-of-research in the areas of mobile computing, cloud computing, wireless systems and application-inspired embedded systems. We will accomplish the education goals through new courses that we are planning to teach that will involve students both at CMU and in Portugal, and that will allow students to leverage the Hyrax infrastructure to develop a new class of futuristic applications that will challenge their imaginations. We also intend to infuse these courses with an entrepreneurial spirit, by teaching these students the necessary business skills behind drafting a business plan, by emulating the necessary market surveys, pricing schemes, competitive analysis, product development and marketing, during the different phases of the course. We will encourage, nurture and inspire startup-oriented project ideas in these new courses, and incorporate guest lectures from the CMU Technology Transfer Office as well as successful entrepreneurs in the course. We will accomplish the innovation goals through the practical real-world deployment of the resulting technologies jointly with Portuguese and American industries. In fact, this proposal is unique in that it is bringing together a Portuguese company (Wavecom), along with an American company (YinzCam), to deploy technology into to benefit a Portuguese client (Benfica) as well as the Portuguese public (Benfica sports fans). YinzCam has proven commercial technology that will be brought to Portugal, through collaboration with Wavecom and Benfica, with YinzCam committing resources and waiving costs towards a successful deployment. This proposal is unique in that Estadio da Luz will serve as the pilot testbed this research, with Wavecom deploying the in-stadium Wi-Fi network, and the Hyrax team generating the research, and the YinzCam team launching these capabilities into Portuguese fans’ hands in the form of the official Benfica Mobile app. This proposal has the potential to make Benfica and Wavecom the first movers in the space of delivering a unique in-stadium fan experience, and will also generate new revenue streams that these Portuguese companies do not possess. PI Narasimhan and Co-PI Gandhi have a track-record of establishing university-based spin-off companies, and will bring their experience and lessons learned to the project and to the students involved in the project. In summary, the Hyrax proposal (i) is futuristic and bold in its research vision, (ii) nurtures an entrepreneurial-style of education and teaching with skill-building in both business and technology, and (iii) encompasses application-inspired research that is designed to be demonstrated through a real-world pilot in Portugal with accompanying tangible economic impact for Portuguese companies. Expected Outcomes. The expected outcomes are four-fold: (i) a body of research results that is well beyond the state-of-the-art and the state of research today, along with undergraduate/Master’s/doctoral theses that will embody these new research results, (ii) a new infrastructure, Hyrax, that will serve as the first instance of a next-generation edge-cloud infrastructure, and that is of significant relevance to both military and civilian applications, (iii) new joint entrepreneurially-oriented project-based courses that build on the new Hyrax infrastructure will expose students to the problems and opportunities presented by mobile computing, cloud computing, embedded systems, and (iv) real-world pilots conducted jointly between Portuguese and American companies committed to the proposal, with the Hyrax team’s research being deployed to provide tangible economic impact for the Portuguese companies along with the invaluable hands-on exposure to the involved students of how to take a concept from inception to implementation to prototype to pilot to commercial impact. Team Formation and Roles. The lead Portugal PI (Silva) and the lead Carnegie Mellon PI (Narasimhan) have a successful track-record of prior collaboration. Their prior collaboration was organically established, effectively resulting in a co-advised Ph.D. student, Rolando Martins, who is now a post-doctoral researcher at Carnegie Mellon University in Pittsburgh. PI Silva, PI Narasimhan and Co-PI Lopes have a successful history of working together, even outside the previous phase of the CMU-Portugal program, inspired by their mutual interest in collaborating in the areas of computer science. This proposal formally strengthens their collaboration, and also serves as a vehicle for them to now bring together a Portuguese company (Wavecom) to collaborate with a U.S. company (YinzCam) to benefit the Portuguese economy and the Portuguese public in a tangible, demonstrable way. The prior track record of collaboration, the ambition/vision of the proposed research, along with the built-in commitment to a real-world pilot and technology transfer, make this proposal unique. An inter-disciplinary vision requires an inter-disciplinary team. The Hyrax team is inter-disciplinary by design and is organized into three sub-groups: (i) systems research, to be conducted by PI Narasimhan, PI Silva, Co-PI Lopes, Co-PI Gandhi, (ii) algorithms research, to be conducted by Co-PI Coimbra, Co-PI Preguiça and Co-PI Lourenço, and (iii) applications research, to be conducted by all of the PIs jointly with the students, the Wavecom team and the YinzCam team, using Estadio da Luz as a pilot and the Benfica Mobile app as the building-block for transferring the research results to the real world. The proposal is written, and its tasks expected to be executed, with the three sub-groups closely aligned and closely collaborating with each other to realize the vision of Hyrax.
135826
CMUP-ERI/FIA/0048/2013
FCT - Fundação para a Ciência e a Tecnologia, I.P.
Portugal
5665-PICT
350,000.00 €
2014-04-21
2018-07-31
No documents found for this project/scholarship.
<script type="text/javascript">
<!--
document.write('<div id="rcaap-widget"></div>');
document.write('<script type="text/javascript" src="https://www.rcaap.pt/snippet?resource=documents&project=FCT%2F5665-PICT%2F135826&fields=id,titles,creators,issueDate,link,descriptions"></script>');
-->
</script>