Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Prediction of Job Resource Requirements for Deadline Schedulers to Manage High-Level SLAs
 

Summary: Prediction of Job Resource Requirements for
Deadline Schedulers to Manage High-Level SLAs
on the Cloud
Gemma Reig, Javier Alonso, and Jordi Guitart
Universitat Polit`ecnica de Catalunya (UPC) and Barcelona Supercomputing Center (BSC)
Barcelona, Spain
Email: {greig, alonso, jguitart}@ac.upc.edu
Abstract--For a non IT expert to use services in the Cloud is
more natural to negotiate the QoS with the provider in terms
of service-level metrics ­e.g. job deadlines­ instead of resource-
level metrics ­e.g. CPU MHz. However, current infrastructures
only support resource-level metrics ­e.g. CPU share and memory
allocation­ and there is not a well-known mechanism to translate
from service-level metrics to resource-level metrics. Moreover,
the lack of precise information regarding the requirements of
the services leads to an inefficient resource allocation ­usually,
providers allocate whole resources to prevent SLA violations.
According to this, we propose a novel mechanism to overcome
this translation problem using an online prediction system which
includes a fast analytical predictor and an adaptive machine

  

Source: Alonso, Javier - Departament d'Arquitectura de Computadors, Universitat Politècnica de Catalunya

 

Collections: Computer Technologies and Information Sciences