The Journal of Supercomputing

, Volume 68, Issue 2, pp 509–556

Semantic-enabled CARE Resource Broker (SeCRB) for managing grid and cloud environment

  • Thamarai Selvi Somasundaram
  • Kannan Govindarajan
  • Usha Kiruthika
  • Rajkumar Buyya
Article

DOI: 10.1007/s11227-013-1047-z

Cite this article as:
Somasundaram, T.S., Govindarajan, K., Kiruthika, U. et al. J Supercomput (2014) 68: 509. doi:10.1007/s11227-013-1047-z
  • 384 Downloads

Abstract

Grid computing is mainly helpful for executing high-performance computing applications. However, conventional grid resources sometimes fail to offer a dynamic application execution environment and this increases the rate at which the job requests of users are rejected. Integrating emerging virtualization technologies in grid and cloud computing facilitates the provision of dynamic virtual resources in the required execution environment. Resource brokers play a significant role in managing grid and cloud resources as well as identifying potential resources that satisfy users’ application requests. This research paper proposes a semantic-enabled CARE Resource Broker (SeCRB) that provides a common framework to describe grid and cloud resources, and to discover them in an intelligent manner by considering software, hardware and quality of service (QoS) requirements. The proposed semantic resource discovery mechanism classifies the resources into three categories viz., exact, high-similarity subsume and high-similarity plug-in regions. To achieve the necessary user QoS requirements, we have included a service level agreement (SLA) negotiation mechanism that pairs users’ QoS requirements with matching resources to guarantee the execution of applications, and to achieve the desired QoS of users. Finally, we have implemented the QoS-based resource scheduling mechanism that selects the resources from the SLA negotiation accepted list in an optimal manner. The proposed work is simulated and evaluated by submitting real-world bio-informatics and image processing application for various test cases. The result of the experiment shows that for jobs submitted to the resource broker, job rejection rate is reduced while job success and scheduling rates are increased, thus making the resource management system more efficient.

Keywords

Cloud computing Grid computing High-performance computing (HPC) Resource broker Semantic description Semantic discovery Service level agreement (SLA) 

List of symbols

NS

Number of Schedule

\(S\,\leftarrow \,\left\{ {S_1 ,S_2 , \ldots ,S_p.S_\mathrm{NS}} \right\} \)

It represents set of schedules

\(S_j \)

It represents pth schedule

M

It represents the total number of resources

\(R\leftarrow \{{R_1 ,R_2,\ldots ,R_M} \}\)

It represents the list of resources

\(J_i\)

It represents ith job

\(R_j\)

It represents jth resource

CB

Cloud broker

\((\mathrm{ER})_{J_i}\)

Exact region contain resources that exactly satisfies the demands of job \(J_i\)

\((\mathrm{SR})_{J_i}\)

Subsume region contain resources that has more capability than demand of job \(J_i\)

\((\mathrm{PR})_{J_i}\)

Plug-in region contain resources that has less capability than demand of Job \(J_i\)

\(\mathrm {Sim\_mat}\)

Similarity matrix holds the cosine similar measure between jobs and resources in \((\mathrm{SR})_{J_i}\) and \((\mathrm{PR})_{J_i}\). Its dimension are \(({1 *\left| {(\mathrm{SR})_{J_i}} \right| })\) and \(({1*\left| {(\mathrm{PR})_{J_i}} \right| })\), respectively

\(({\mathrm{hs}\_\mathrm{SR}})_{J_i}\)

High-similarity resources that satisfy demand specified by job \(J_i\) from subsume region

\(({\mathrm{hs}\_\mathrm{PR}})_{J_i}\)

High-similarity resources that satisfy demand specified by job \(J_i \) from plug-in region

\(\mathrm{threshold}_\mathrm{SR}\)

To filter the resources from subsume region whose similarity is greater or equal to specified value

\(\mathrm{threshold}_\mathrm{PR}\)

To filter the resources from plug-in region whose similarity is greater or equal to specified value

\(\mathrm{Req}({J_i})\)

\(\{J_i^\mathrm{ID} ,\,\mathrm{no}\_\mathrm{Nodes}_{J_i} ,\,\mathrm{Ram}\_\mathrm{needed}_{J_i} ,\mathrm{Hd}\_\mathrm{needed}_{J_i }, \mathrm{Deadline}_{J_i} ,\,\mathrm{OS}_{J_i} ,\,\mathrm{Sw}_{J_i} \}\)

\(\mathrm{Avail}({R_j})\)

\(\left\{ \!\!{\begin{array}{l} R_j^\mathrm{ID} ,\mathrm{no}\_\mathrm{Nodes}_{R_j}, \mathrm{Ram}\_\mathrm{Capacity}_{R_j}, \mathrm{Hd}\_\mathrm{Capacity}_{R_j},\\ \mathrm{Availability}_{R_j},\mathrm{OS}_{R_j} ,\mathrm{Sw}_{R_j}\!, \mathrm{Type}_{R_j} \\ \end{array}} \!\!\right\} \)

\(\left| {(\mathrm{ER})_{J_i}} \right| \)

Number of resources in exact region of job \(J_i \) that is represented as \(N\_\mathrm{ER}\)

\(\left| {\,({\mathrm{hs}\_\mathrm{SR}})_{J_i} \,} \right| \)

Number of resources in subsume region of job \(J_i \) that is represented as \(N\_\mathrm{SR}\)

\(\left| {\,({\mathrm{hs}\_\mathrm{PR}})_{J_i} \,} \right| \)

Number of resources in plug-in region that satisfies job \(J_i \) that is represented as \(N\_\mathrm{PR}\)

\(N_{(\mathrm{ER})_{J_i}}\)

List of resources to which negotiation is agreed in exact region by broker

\(N_{(\mathrm{SR})_{J_i}}\)

List of resources to which negotiation is agreed in subsume region by broker

\(N_{(\mathrm{PR})_{J_i}}\)

List of resources to which negotiation is agreed in plug-in region by broker

\(P_\mathrm{Init}\)

Initial proposal send by cloud broker

\(U_\mathrm{Init}\)

Utility value of the cloud broker

\(({P_\mathrm{c}})_{({R_j})}\)

Counter proposal given by resource \(R_j\) that receives \(P_\mathrm{Init}\)

\((\mathrm{No. proposal})_\mathrm{ER}\)

Number of resources in exact region that send counter proposals which is received by broker

\((\mathrm{No. proposal})_\mathrm{SR}\)

Number of resources in subsume region that send counter proposals which is received by broker

\((\mathrm{No. proposal})_\mathrm{PR}\)

Number of resources in plug-in region that send counter proposals which is received by broker

\(({U_\mathrm{c}})_{({R_j})}\)

Utility counter of the resource \(R_j\)

\(U_{(\mathrm{ER})_{J_i}}\)

Set of utilities received by broker from the resources in exact region of the job \(J_i \)

\(U_{(\mathrm{SR})_{J_i}}\)

Set of utilities received by broker from the resources in qualified subsume region of the job \(J_i \)

\(U_{(\mathrm{PR})_{J_i}}\)

Set of utilities received by broker from the resources in qualified plug-in region of the job \(J_i \)

\(P_\mathrm{Avg}\)

Average proposal received by the cloud broker

\(({R_{(\mathrm{PR})}})_{J_i}\)

Resources in plug-in region that is scheduled to Job \(J_i \)

\(\mathrm{JS}_{S_p}\)

Number of Jobs submitted for the Schedule \(S_p \)

\(\mathrm{JR}_{S_p}\)

Number of Jobs rejected for the schedule\(S_p \)

\(\mathrm{JRR}_{S_p}\)

Job rejection Rate for the schedule \(S_p \)

\(\mathrm{JSTS}_{S_p}\)

Number of Jobs submitted to the scheduler in the schedule \(S_p\)

\(\mathrm{JSR}_{S_p}\)

Job Success Rate for the schedule \(S_p \)

\(\mathrm{JSD}_{S_p}\)

Number of Jobs successfully met deadline in the schedule \(S_p \)

\(\mathrm{JFD}_{S_p}\)

Number of jobs failed to meet deadline in the schedule\(S_p \)

\(\mathrm{SSR}_{S_p}\)

Scheduling success rate for the schedule \(S_p \)

\(\hbox {TPDefected}_{\mathrm{ith}\_\mathrm{fruit}}\)

Total pixels defected in ith fruit

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  • Thamarai Selvi Somasundaram
    • 1
  • Kannan Govindarajan
    • 1
  • Usha Kiruthika
    • 1
  • Rajkumar Buyya
    • 2
  1. 1.Anna UniversityChennaiIndia
  2. 2.Cloud Computing and Distributed Systems LabThe University of MelbourneParkvilleAustralia

Personalised recommendations