Skip to main content

Advances, Systems and Applications

Table 1 System and Workload Parameters for the Factor-at-a-Time Experiments

From: A resource management technique for processing deadline-constrained multi-stage workflows

Parameter

Values

Default Value

Workload

  

 Type

{CyberShake, LIGO, Genome}

–

 Job arrival rate (job/s)

λCS = {1/18, 1/22, 1/30}

λLG = {1/150, 1/180, 1/265}

λGN = {1/1800, 1/2290, 1/3205}

λCS = 1/22

λLG = 1/180

λGN = 1/2290

 Earliest start time of jobs, s j (sec)

\( {s}_j=\left\{\begin{array}{c}{at}_j,\kern7.75em x=0\\ {}{at}_j+ DU\left(1,{s}_{max}\right)\kern0.75em x=1\end{array}\right. \)

where at j is the arrival time of job j, x ~ Bernoulli(0.5), and s max  = {1, 5, 25} * 104

s max  = 50,000

 Job Deadline, d j (sec)

\( {d}_j=\left\lceil {s}_j+{SET_j^R}^{\ast } em\right\rceil \) where

em ~ U(1, em max ) and em max  = {2, 5, 10}

em max  = 5

System

  

 Number of Resources, m

m = {40, 50,60}

m = 50

 Resource Capacity

c r  = 2

–

Configuration of RM-DCWF

  

 Laxity Distribution Algorithm

{PD, ED}

–

 Approach to calculate the job laxity

{SL, TL}

–

 Task Scheduling Policy

{TSP1, TSP2}

–