The Storage System Research Group at BSC is developing a distributed storage platform to store and share data between applications. A distinguishing feature of this platform is that, from the point of view of the applications using it, data is stored in the form of objects, which include data, code (methods manipulating the objects) and behavior policies that are also stored together with the data. The main purpose for storing methods in the platform is to bring execution close to the data in order to avoid unnecessary data transfers from the data store to the application that is executed in the client.
By analyzing the code of the methods, one can obtain information about which objects are always accessed together, or which parts of an object are accessed together and which not. This information must be complemented with additional data gathered during execution and analyzed using machine-learning techniques.
Description
The Storage System Research Group at BSC is developing a distributed storage platform to store and share data between applications. A distinguishing feature of this platform is that, from the point of view of the applications using it, data is stored in the form of objects, which include data, code (methods manipulating the objects) and behavior policies that are also stored together with the data. The main purpose for storing methods in the platform is to bring execution close to the data in order to avoid unnecessary data transfers from the data store to the application that is executed in the client.
Taking advantage of the fact that the platform knows about the methods that directly manipulate the data, it can use this information to improve the performance of the applications using this data. In particular, by analyzing the code of the methods, one can obtain information about which objects are always accessed together, or which parts of an object are accessed together and which not. This information must be complemented with additional data gathered during execution and analyzed using machine-learning techniques. This is essential in order to discover additional usage patterns that cannot be obtained from the methods because they depend on the applications that access them.
The candidate should obtain both kinds of information and use it to improve data placement among the different back-ends, with the goal of minimizing communications and data transfers also within the platform for the sake of performance, scalability and energy efficiency.
To apply, please email a cover letter, CV, contact address of at least two professional references and copies of degree certificates to Dr. Anna Queralt. Incomplete applications will not be considered or answered.
Nr of positions available : 1
Research Fields
Computer science
Career Stage
Early stage researcher or 0-4 yrs (Post graduate)
Research Profiles
First Stage Researcher (R1)
Benefits
he candidate will be mainly hosted by the Barcelona Supercomputing Center. From month 30 to 36, the candidate will be in Universidad Politécnica de Madrid, where he will develop the collection and analysis of data generated at runtime using machine learning techniques.
Comment/web site for additional job details
bigstorage.oeg-upm.net/jobs.html
Requirements
| ENGLISH |
| Excellent |
| Computer science |
| At the time of recruitment, the applicant must not have lived in Spain for more than 12 month in the previous 36 month and he/she is in the first four years of his/her research career, starting at the date of obtaining the degree which would formally entitle to embark on a doctorate. Bachelor or Master degree on computer systems (or similar). |
| Master Degree or equivalent |
| Knowledge on • Object oriented programming in general • Compilers • Operating systems • File systems • Java and/or Python Other skills • English both spoken and written • Group working |