I am working on a topic i.e. Quality models in cloud computing. There are many existing quality model for SaaS, PaaS. so is it also possible to make quality model for resource scheduling by considering the QoS attributes of various existing resource scheduling techniques.
Why not consider a vector of requirements by each resource requesting service S(i).
The services S(i) compete for resources.
The requirements vector could have a summary (network, compute, store) for instance. Network requirement could itself be (speed, latency) or other relevant measures. Idem compute and store requirements can be defined appropriately. Then you have requirement vector R(S(i)).
Now for two competing services you may either have sequential order policy like priority to R(S(i0)) then R(S(i1)) etc and at each step service S(j) consumes the resource it requests from the remaining pool of resource.
You may alternatively consider a fair sharing policy where the total demand for a resource is computed, and the allocation fairness is defined (TBD).
Does this method help?
@ Renaud Di Francesco: Thanks for answer sir...
yes this method can be considered for resource scheduling.. but basically want i want is to firstly collect all important QoS parameters mainly used in all types of scheduling.like in scheduling of dependent or independent task. then used these QoS parameters to define some standard or model for cloud service providers.
i.e. according to QoS requirement of a user our model will tell which providers is satisfied these QoS requirements.
I am confused wether it is possible to make such a model or not..
Article Quality-of-service in cloud computing: modeling techniques a...
Article Quality-of-service in cloud computing: modeling techniques a...
- Badges
- Science method