Seadmelt-seadmele mobiilsed mängud

Abstract

Tänu läbimurretele mobiilsete seadmete ja sotsiaalvõrgustikes vastastikuse mobiilse suhtlemise valdkondades on seadmelt-seadmele (ingl. k. device-to-device) mobiilsed mängud muutunud aktuaalseks trendiks. Selleks, et säästa rakenduste poolt nõutavat energiat ja kiirendada nende reaktsiooniaega, on võimalikuks vahendiks kasutada koodi mahalaadimist pilve vahendusel või seadmelt-seadmele. Teatavasti on andmevahetuses eelistatud madal latentsusaeg, mille tõttu on seadmelt-seadmele mahalaadimine sobilikum. Sellegipoolest ei ole lähedal asuvale seadmele mahalaadimine praktikas otstarbekas, sest kasutaja ei pruugi olla nõus teise seadme poolt edastatud ülesande lahendamises, kuna sellega kaasneb lisanduv energia kadu. Antud töös läheneme probleemile uuest küljest: selle asemel, et lasta teisel seadmel töö ära teha on võimalik kasutada juba lahendatud ülesannete tulemusi. Püstitatud eesmärgi saavutamiseks arendati välja raamistik ja teostati juhtumiuuring. Valideerimise tulemusele põhinedes leidsime, et lähedal asuvate, omavahel ühendatud seadmete puhul on võimalik vähendada rakenduse koormust.

Device-to-Device(D2D) mobile gaming is a new trend which is emerging as a result of the increasing advances in mobile devices and social network interaction with mobile peers. As these games are played between players in proximity, it is possible to take advantage of computational offloading to balance the load of these applications. Smartphone games can be instrumentalized with computational offloading mechanisms in order to save energy and increase response time of the applications. In this context, remote cloud and D2D offloading has been proposed. It is well known that low latency is preferable to high latency in the communication when offloading, and as a result, D2D offloading is more suitable than remote cloud. However, the idea of offloading to a nearby device is not feasible in practice, because a user may not be willing to process the task from another device. This can be clearly seen as processing a task from another device does not represent a gain but rather a loss in resources for the device that executes the task. In this thesis, we investigate a new perspective, in which a device is not requested to process a task, but it is alleviated from processing one task that another device has already processed. To achieve this purpose, we develop a framework and a case study. Based on the result of the validation, we found out that it is possible to balance the execution load of an application between nearby interconnected devices.