Personaalne andmeruum lingitud andmetele

Abstract

Personaalsete andmete ristkasutuse puudumine veebirakenduste vahel on viinud olukorrani, kus kasutajate identiteet ja andmed on hajutatud eri teenusepakkujate vahel. Sellest tulenevalt on suuremad teenusepakkujad, kel on rohkem teenuseid ja kasutajaid,\n\rväiksematega võrreldes eelisseisus kasutajate andmete pealt lisandväärtuse, sh analüütika, pakkumise seisukohast. Lisaks on sellisel andmete eraldamisel negatiivne mõju lõppkasutajatele, kellel on vaja sarnaseid andmeid korduvalt esitada või uuendada eri teenusepakkujate juures vaid selleks, et kasutada teenust maksimaalselt. Käesolevas töös kirjeldatakse personaalse andmeruumi disaini ja realisatsiooni, mis lihtsustab andmete jagamist rakenduste vahel. Lahenduses kasutatakse AppScale\n\rrakendusemootori identiteedi infrastruktuuri, millele lisatakse personaalse andmeruumi teenus, millele ligipääsu saab hallata kasutaja ise. Andmeruumi kasutatavus eri kasutuslugude jaoks tagatakse läbi linkandmete põhimõtete rakendamise.

Recent advances in cloud-based applications and services have led to the continuous replacement of traditional desktop applications with corresponding SaaS solutions. These cloud applications are provided by different service providers, and typically manage identity and personal data, such as user’s contact details, of its users by its own means.\n\rAs a result, the identities and personal data of users have been spread over different applications and servers, each capturing a partial snapshot of user data at certain time moment. This, however, has made maintenance of personal data for service providers difficult and resource-consuming. Furthermore, such kind of data segregation has the overall negative effect on the user experience of end-users who need to repeatedly re-enter and maintain in parallel the same data to gain the maximum benefit out of their applications. Finally, from an integration point of view – sealing of user data has led to the adoption of point-to-point integration models between service providers, which limits the evolution of application ecosystems compared to the models with content aggregators and brokers.\n\rIn this thesis, we will develop an application-agnostic personal storage, which allows sharing user data among applications. This will be achieved by extending AppScale app store identity infrastructure with a personal data storage, which can be easily accessed by any application in the cloud and it will be under the control of a user. Usability of data is leveraged via adoption of linked data principles.